Professional Documents
Culture Documents
Risk Analysis and Management of Petroleum Exploration Ventures
Risk Analysis and Management of Petroleum Exploration Ventures
by Peter R. Rose
Managing Partner .
Rose & Associates, LLP
Austin, Texas, U.S.A.
Published by
--
Copyright © 2001 by
AAPG grants permission for a single photocopy of an item from this publication for personal use.
Authorization for additional copies of items from this publication for personal or internal use is
granted by AAPG provided that the base fee of $3.50 per copy and $.50 per page is paid directly . I
to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, Massachusetts 01923 (phone: .
978/750-8400). Fees are subject to change. Any form of electronic or digital scanning or other digital
transformation of portions of this publication into computer-readable and/ or transmittable form for
personal or corporate use requires special permission from, and is subject to fee charges by, the AAPG.
ii
"
~""-
I
AAPG
wishes to thank
toJNOC
...
Iii
Preface
I
/
I
~
In the summer of 1996, I was indted by Japan ll1e following volume is an overview of this important
National Oil Company (JNOC) to prepare a report aspect of petroleum exploration and development. Many
summarizing the-current status of exploration risk of the concepts and techniques were originated by profes
analysis as practiced by modern international oil and sional colleagues in other companies and shared through
gas companies-the main concepts, procedures, and many conferences and publications (most sponsored by (
problems as I had observed them during my deep the AAPG) as well as consulting assignments. The fol
involvement in that area during the late 1980s and lowing people deserve special mention as significant con
1990s. Following the presentation and acceptance of tributors to risk analysis as practiced in the late 1990s:
that report in March 1997, JNOC asked me to prepare
a companion report covering risk analysis of explo Jeff Brown, David Cook, Eric Dion, Mike Effler (Mobil)
ration plays, which was presented in Tokyo in r-.'Iarch Laurens Garenstroom, Rien Nederlof, the late Dieter
1998. Later that year, encouraged by the American Schluijk, Hany Thomsen, Daniel Zweidler (Shell)
Association of Petroleum Geologists' L-\A.PG) Publi Bob Baker, the late Harry Gehman, Bill James, Bob Megill,
iv
How~ver, most of the present publication draws on own. My profound thanks go to Ms. Elizabeth Ethridge
concepts and procedures that emerged and evolved for her usual excellence and cheerfulness in preparing
during the AAPG school, "Managing and Assessing the original reports for JNOC, as well as the merged
Exploration Risk," team-taught from 1984 until 1998 final manuscript for publication. Ms. Anne Th~mas
by Ed Capen, Bob Clapp, Bob Megill, and myself. I and Ms. Rowena Mills provided expert editing assis
acknowledge their substantial contributions to this tance through the AAPG's Publications Department.
volume with admiration, gratitude, and many pleas Finally, I thank my wife, Alice, and our family for their
ant recollections. encouragement and patience during the production of
Roger Holeywell, Daniel Zweidler, and Chap Cron this book.
quist reviewed the manuscript, to its great benefit. My
colleagues at R&A, Jeff Brown, Gary Citron, and Mark Peter R. Rose /
McLane, improved the book through their encourage Rose & Associates, LLP
ment, critical review, and suggestions. Any errors of Austin, Texas
commission or omission, however, are entirely my December 5, 2000
-C" . •
v
List of Abbreviations
vi
Dedication -------------...,.....---_
.... ,"
~
• .,;>
vii
J
/
AAPG
wishes to thank the following
to
AEC International
Pluspetrol E&P SA
Samson
Texaco Inc.
. _.
---
._~
viii
Table of Contents -_...;.,.-----:-.:;..:;..-.,------
Chapter 1: Introduction 1
Lognormality 9
Industry Experience , 28
Introduction 49
ix
,
Chapter 5: Exploration Plays-Risk Analysis and Economic Assessment 57
Introduction 57
Field-size Distributions 67
Primary Parameters 84
Introduction 91
Conditions of Acquisition : 94
Oral Auctions 98
Performance Contracts 98
Private Treaties 98
Corporate Acquisitions - 99
Conclusion 99
Requirements - " , 99
Benefits 99
xl
32. Integrated methods produce more efficient exploration results \ 66
33. Creaming curves 68
34. Global field-size distribution 72
35. (a, b, c) Use of analog FSDs in forecasting actual FSDs 73
36. (a, b) Use of analog FSDs in forecasting actual FSDs: second stage 75
37. (a, b, c, d) Downward shifts in FSDs in area Y between first cycle of exploration and second cycle " 76-77
38. (a, b) Offshore FSDs are truncated by platform cost 79
39. Effects of operative geologic chance factors must coincide in the prospective area 81
40. Staged exploration 9-!
41. Sealed bidding for uncertain reserves leads to the winner's curse 95
42. Net purchases by company: Gulf of Mexico sales 96
- 43. Gulf of Mexico bidding efficiency :96
44. High bids and second bids: area-wide GOM lease overbids averaged 50% 97
45. High bids and second bids: area-wide GOM lease overbids averaged 75% 97
46. Spinner for simulating chance of success and reserves discovered 103
47. Predictive accuracy of portfolio performance improves with the number of wells in the portfolio 104
48. The efficient frontier 105
49. a) Post-1900 BBOE J;scovered by yc~:; b) Discovery data, 1990-1999 108
50. Six generalized planning steps 112
51. Lognormality-two modes of portrayal 118
52. Exploration reality 119
TABLES
1. Estimated geotechnical parameters 5
2. Biases affecting judgments under uncertainty 8
3. Ways to improve accuracy and build confidence in estimating , ' 13
4. Reality checks (1): Characteristics of the endpoints of the reserves distribution 20
5. Getting a sense of scale about HC-recovery factor 24
6. Reality checks (2): Characteristic ranges associated with oil and gas ventures ,
haYing different magnitudes of uncertainity 26
7. Expected value examples (coin toss) 31
8. Example calculation of simplistic probability of geologic
success and failure using three geologic chance factors 34
9. Generalized success rates of various well classes drilled in the 1980s ., 45
10. Present value factors '1. ••••••••••••••••••••••••••• 51
11. Factor comparison in the four levels of petroleum investigation 60
12. Exploration risk analysis-prospects versus plays 61
13. System of graduated size classes used currently by the U.s. Geological Survey 70
14. Comparison of analog, predicted, and actual FSDs 74
15. Geologic chance factors required for play and prospect success ~ 81
16. Geotechnical parameters required for play analysis 84
17. Form: flow sheet for play risk analysis ',' _' 85
18. Biases affecting risk decisions 92
19. EV implies "risk-neutral" . 92
20. A model prospect portfolio : 101
21. Simulation of results for a prospect portfolio 102
22. Uncertainty leads to common underperformance of exploration portfolios 106
EQUATIONS
1. Generic expected value 1
2. Expected net present value of an exploratory prospect 5
3. Expected net present value of mean commercial reserves case 54
4. Investment efficiency 54
5. Risked investment effiCiency 54
6. Risk-adjusted value 55
7. Optimum working interest 55
8. Discovery process model 67
9. Calculating minimum economic field size -.::.. " 78
..... r::.
> ~$ • ..>
10. Calculating estimated chance of c~'vnorr";rylaysuccess 83
11. Prospect grading procedure 88
12. Application of OWl to a prospect : 92
xii
About -th~ A!.Jthor
xlii
Chapter
-
Introduction
.Duringth,~ 1~9.9~'II:\any intemationaIpetroleum colJ,l game are w~ll known to the owner, and they are set to
pani ·:bf;(. . ' _.~ . -'signiltr be slightly in his favor. He is playing a repeated-trials
-()~tioli~ game in which the expected value (EV) of each trial, for
':" ,. _~M-~l~ ;~ him, is positive. Please note here the definition of
gieS~ While exploration risk cannot be eliminated, it can expected value: The chance of success times the value of
certainly be reduced substantially, on a portfolio scale. success, minus the chance of failure times the cost of fail
And the widespread adoption of standardized risk ure (Equation 1). When EV is positive, you're investing;
analysis methods during the 1990s brought badly when it's negative, you're gambling. If the casino opera
needed discipline to petroleum exploration. tor knows the number of tables, the number of players in
By the mid-1980s, most well:'informed major inter an evening, and the house rules, he can predict with con
national petroleum firms that were engaged in explo 'siderable precision what his profit will be. He is not a
ration recognized that, globally, the average size of gambler, any more than a life insurance company is. He
new discoveries was diminishing (Figure 1). Not coin is an investor. And he knows what the odds are,
cidentally, the class of exploratory prospects catego
rized as "high-risk/high-potential" was showing EXPECTED VALUE: (1)
marked signs of underperformance. For major compa ,(Chance of Success x Value of Success)
nies such as Shell, Mobil, i},nd Amoco, when all such (Chance of Failure x COl't of Failure)
1600
"0
...
Q)
Q)
1400
>
0
0 1200
(I) • > 1BBOE
c 1000
(I) [II500 MMBOE-1 BBOE
"0
Q) 800 [j 200-500 MMBOE
-...
u:
0
(1)
600
Q 100-200 MMBOE
III 50-100 MMBOE
"
.0
E 400
o 25-50 MMBOE
::J
Z 200
0
1960s 1970s 1980s 1990-99
Figure 1 Cloballarge field discoveries, 7960-7999. While the rate of global scale (>7 BBOE, or billion barrels of oil
equivalent) discoveries has decreased since 7980, opportunities for smaller, but significant, fields have remained high.
MMBOE = million barrels of oil equivalent. Source: Petroconsultants; reprinted with permission.
and 18th centuries (Bernstein, 1996) through published Today, the concepts and methodologies of explo
and unpublished works by Cardono (1545), Pascal and ration risk analysis used by most large oil companies
Fermat (1654, ~"':pub~lt~d !D'. Bemstein, 1996), Graunt and state petroleum agencies ha\'e converged to the
and Petty (1662), Halley (1693), Jacob Bernoulli (1713, status of"a generally accepted technology" (MacKay,
posthumously), De Moivre (1733), Daniel Bemoulli 1996), and many smaller firms are now adopting this
(1738), and Bayes (1764, posthumously). The impetus technology.
and applications for the early studies concerned games
and gambling; subsequent applications had to do with
actuarial and insurance matters. Risk Analysis and /
Petroleum exploration-like many serial businESs Petroleum Exploration
ventures~is clearly a process of repeated tr:ials under
conditions of uncertainty, each trial requiring a sub The most critical decision in petroleum exploration is
stantial commitment of investment capital. As such, not which prospect to drill-it is which basin or trend
the casino analog is apt, but because we do not know ("play") to explore (Rose, 1996a, 1996b; Br~wn and
certainly what the odds are, or the size of the prize, we Rose, 2000). A play is a family of geologically similar
have learned to employ modem science and technol fields, discO\'eries, prospects, and leads (McCrossan,
ogy to refine our "bets." For example: 1973; Roy, '1975; Baker et aI., 1986; Miller, 1986; White,
1992). Because of the similarity of reservoir type and
• Stratigraphy may help us predict the presence and trap geometry, commonality of hydrocarbon charge,
quality of reservoir rocks; and consistent exploration and development methods,
• Geochemistry may give us a better idea whether or it is possible to carry out economic evaluations·of plays
not oil or natural gas has migrated into the area of as full-cycle economic ventures. Such evaluations uti
our prospect, thus allowing better estimates of the lize regional geologic, geochemical, and geophysical
chance of hydrocarbon presence; studies (see Appendix E and Appendix G), and they are
• Geophysics may help refine our estimates as to conducted by applying basic principles of prospect risk
how large an accumulation may be contained in anah'sis, as outlined in the first part of this volume.
the trap, thus the "size of the prize," as well as the
Ho\\:ever, in order to understand the principles of play
likelihood that the trap indeed contains reser
voired oil or gas; analYsis, we must first understand the risk analysis of
prospects, simply because plays are aggregates of geo
• DrilliJlg techJlology may let us reduce our invest
logically similar prospects.
ment by testing the prospect at lower cost; and
Accordingly, we begin with the assumption that
• Reservoir techllology may allow us to de\'elop and
produce the successful prospect (= discovery) the company has already committed to explore in
mOre efficiently, thus incre~sing our profit. several different basins and trends (i.e., the process
of play selection has already occurred), and drill
T~e success of exploration science and technology ing prospects are now being searched for, identi
d unng the 1945-1995 period helped advance the fied, and evaluated. Later, in Chapter 5, we return to
myth that petroleum exploration was "driven by sci play analysis.
ence" and obscure its analogy to the casino example, Step one in petroleum exploration, once the explo
which was distasteful to many businessmen and sci ration trend has been selected, is idelltification of tlte
entists. Pioneering publications by Grayson (1960) on drillillg prospect by geoscientists. This is the basic
decision analysis in petroleum exploration, and by value-creating act. It requires geotechnical skill and
Kaufman (1962) on lognormality, provided the basis creative imagination. .
for practical applications in valuing exploration ven After the exploration prospect has been identified,
tures. However, the systematic employment of statis there are key tasks involved throughout the life cycle
tics and probability theory in exploration did not of petroleum exploration and development. The first
begin until the mid-1960s, led by Exxon, Shell, Arco, three tasks deal with measuring the prospect's value.
and Cities Service (Newendorp, 1975). By the late Step two in petroleum explora tion is lIlenSll ring
1970s, companies like Chevron, BP, and ELF, as well vallie:
as some governmental agencies (G.S.c., I.F.P., and
U.s.G.5.) were also begim1ing to employ risk analysis 1. Estimatillg llOw large tlle producible reserves are
routinely in their exploration evaluations and ven likely to be (assuming that a hydrocarbon accu
tur~s. Bu t in the late 1980s and early 1990s, a techno mulation is indeed present);
loglcal explosion in risk analysis took place because 2. Estimating tile chance that a producible hydrocar
most modern oil and gas companies saw the need for bon accumulation is present; and
s'ystematic management of their exploration portEo 3. Estimating t!le profitability of tife en(:~ _ prl:="~t,
lJOs on a worldwide basis. given that producible hydrocarbons are present.
4 Introduction
Step three consists of illlplcl/lcntl1tiolll1l1d I11l1tll1ge acknowledges that companies usually explore on the
mellt of exploration projects I1S business ventures, and basis of regional trends and seismic programs, hoping
includes three additional tasks: to make multiple discoveries-given successful explo
riltion-even though individual prospects constitute
4. Acquisition strategies-determining the terms the economic "building blocks" in the overall evalua
under which the company would commit to tions of the trend.
explore and develop the prospect, whether Chapter 6 concerns effective practices by which suc
acquisition is by sealed bonus-bidding, oml auc cessful compilnies make and implement sound busi
tion, performance contract, serial contract negoti ness decisions based on proper risk anillY6is of
ations, private treaty, contract renegotiations, or prospects .ilS weHas plays. It deals with topics stich as
the like. Note: acquisition strategies often apply risk-reduction techniques, acquisition strategies,
to properties much larger than individual inventory and portfolio management, and manage
prospects, such ilS plays and contruct areas. ment of exploration plays. Chapter 7 reviews current
5. Jnvclltory alid portfolio ml1nl1gelllellt-choosing organizational practices commonly llsed by industry
which prospects :;!'.auld be included in the in carrying out exploration risk analysis.
annual drilling program to maximize economic Seven appendices cover vari(;us aspects of risk
return, consistent with the compilny's risk toler analysis concepts and procedures, which are too
ance. This topiC milY also include consideril tions detailed to be included in the text.
of risk aversion or utility theory.
6. Operntio/ls-carrying out the \'arious business
operations that result in leasing, data acquisition, Differing Definitions of Reserves
drilling, completion, and delineation of new dis
coveries. The cost-effectiveness of data acquisition Much confusion attends the use of the term
for emerging prospects may be ilddressed operil "reserves." The term is commonly used in both formal
tionally, using principles of Va/I/ntioll of lllfil/'llll1thl/l. and casual ways. It is important to understand that
there are two primary definitions of reserves in
Tasks (1) through (5) are ordinarily included in petroleum exploration and production.
the overall subject of exploration risk analysis, which The term "reserves" has measured, fiduciary impli
may also include considerations such as the \'alue of cations to the engineer, chief financiill officer, banker, or
information. financial analyst. Terms such as "proved," "probable,"
"possible," "developed," and "undeveloped" connote
varying levels bf'confidence in the existence of the dif
Purposes and Organization
ferent categories. A fundamental problem with such
of this Volume
terms is that no consistent definition for such categories
can exist, beyond nebulous guidelines like "reasonable
The purposes of this book are to rt'\'iew the main certainty." The engineering profession appears to be
principles and procedures of exploration risk analysis, evolving toward the more preferable probabilistic
and to discuss industry patterns of implementation, expression of such confidence (Cronquist, 1997).
experience, and general results. The book follows the The term "reserves" us~d'by the explorationist usu
general sequence outlined abon.>. Followii1g this intro ally means the projected ultimate recovery of hydrocar
duction, Chapter 2 addresses theory, procedures, and bons from a given field or drilling prospect; genernl
problems in making geotechnical estimates under synonyms are "geologic reserves," "volumetric
uncertainty, and techniques fM impw\'ing staff per reserves," "potential reserves," or "estimilted ultimate
formance in estimating. recovery "(EUR). Thus, "field size," as used in a field
Chapter 3 reviews the key aspects of risk analysis of size distribution, represents the projected final cumula
the prospect as the economic unit of exploration. The tive production from a given field-cumulative
key aspects include (1) estimation of the range of production, plus proved reserves, plus projected future
potential recoverable reserves, (2) chances of geologic reserves additions (i.e., anticipated reserves apprecia
and commercial success, and (3) combination of tion). This same concept is used when explorationists
reserves and chance leading to ex~"'\ectl'd monetary talk about "prospect reserves" or "potential reserves."
value (EMV). Chapter 4 cO\'ers the topic of project This usage naturally lends itself to the probabilistic
profitability, given success, and re\'iews considera expression of confidence in the existence (and e\'entual
tions of both resen'es as well as chance of success. recovery) of such hydrocarbons under the economic
With the principles of prospect risk analysis set and technologic conditions currently prevailing.
out, we proceed to Chapter 5, which covers the key In ni~t plc~s in ·,....;s book, the intended meaning
--C". •
.....
aspects of analysis of the exploration play, which is the should be clear. Where any doubt may exist, I have
operational unit of exploriltion. This distinction attempted to specify the meaning intended.
~h~pter ..
. .,.- Geotechnical Estimates
under Uncertainty
/
There are several important observations to be and old exploration theaters. Accordingly, explo
made about this equation. First, in an American-style rationists will always have to deal with uncomfortably
tax and royalty license, the owners of the producing large uncertainties. Figure 2 shows, by cross-plot, four
property usually pay 100% of the costs but receive a actual experiences of modern oil companies in
reduced proportion-ordinarily from about 70% to attempting to estimate reserves contained in prospects
87.5%-of the revenues from production. This that turned out to be discoveries (Capen, 1992). Figure
reduced proportion is the net revenue interest (NRI); 15 shows similarly wide uncertainties by a large num
the remainder goes to the royalty owner(s)-generally ber of companies exploring in the Norwegian North
the landowner. In production-sharing contracts the Sea over an extended period of time. _
formula is different, although the general principle The cross-plots in Figure 2 express two important
still prevails that the operator provides most or all of attributes of prospect reserves forecasting:
the capital, especially the exploration investment, but
receives only a part of the production revenues. Sec 1. Prevalent optil/listic bias of estimates vs. actual
ond, the equation expresses the profit or (loss) as if it outcomes-in cross-plots B, C, and 0, most of
were a "lump-sum" payment, whereas it is actually the outCOP1PS are overestimates (cross-plot A
received over a long period of time: a complex net is an unbiased data set). We will address bias
cash-flow stream combining investments, production later.
decline, price fluctuations, expenses (including laxes-)~ 2. Sl/bsfalltia1-zmcerfnillty,expressed as a characteris
and inflation. Third, in order to consider the time tic, wide "scatter" of estimates. Given a continu
value of money, the net cash flows are expressed as a ing and substantial number of exploration
discol/llted cash-flow stream so the entire venture can ventures (repeated trials), statistics offers a prac
be compared with current alternative investments. tical way to deal with the prevailing large uncer
Wherever a dollar value is expressed as Present Value tair,lies that characterize petroleum exploration.
(PV), it means that the value has been discounted to
reflect the time value of money. This characteristic, wide "scatter" conveys a pro
Uncertainty attends every item in this ENPV equa found (if disagreeable) message that petroleum
tion except the net revenue interest. These uncertain prospectors would do well to absorb: Exploratiollists
ties are diverse, relating to geology, engineering, law, cnImot predict very well how ml/ell oil or gns their sllccess
politics, economics, and acts of God. It is the special fill prospects (= discoveries) will COlltnill. We calt generally
professional responsibility of geotechnical staff to esti idcl/tify t!lOse -closures tllnt nre too smnll to cOlltnill large
mate the magnitude of reserves, production rates, and i'oillmes, but we call1lot forecast lIow much oil or gas the
costs; to reduce the level of uncertainty as much as large closllres lila}) cOlltailt, often because of lI/wnticipated
possible through sound scientific and technological l'ariatiolls ill trap fill-up, reljervoir qualify, alld dip rate.
judgment (and additional investigation, \vhere war Stated more pragmatically, we can usually distin
ranted); and to accurately and consistently convey guish between a 5-million-barrel (MM bbl) closure and
estimates-as well as uncertainty levels-to manage a 50-million-barrel closure, but we cannot distingUish
ment. Otherwise management's investment decisions between 3-million and 5-million-barrel closures, or 30
may be misguided and imprudent. Thus the financial million and 50-million-barrel closures. Moreover, we
consequences of their geotechnical predictions and often cannot tell how much ~il may be present in a clo
estimates con!':~iiute L1 weighty professional responsi sure having 50-miIlion-barrel capacity-l million,S
bility of the geotechnical staff. .. million, 20 million, or even 50 million. This basic "fact
of life" has profotmd (and often ignored) implications
for explorationists.
Magnitude of Geotechnical
Uncertainty Ranges and Probabilities
The earth is a coarse filter. Even though petroleum The problem is how to express our tedmical uncer
explorationists employ increasingly sophisticated and tainties realistically, and in a form by which they can be
discriminating tedmology, our precision in measuring utilized in economic equations and formulae and
most of the important geotechnical factors and para subjected to subsequent evaluation. The most conunon
meter~ bearing on prospect value is much more lim convention in use today by modern petroleum cvrpora
ited than many of us care to admit. Technology can lions involves the formulation of a range of anticipated
reduce, but not eliminate, uncertainty. Moreover, values for a given parameter, with probabilities
exploration always operates at the cutting edge-the ordin~ly.90%, 50%, and 10%-assigned to the values
~
thlt.~,old of resolution-so that the potential effective tha t consti tu~ the ·.iallge. For example, the geologist
ness of new tools and concepts is constantly and may believe there is a 90% chance that the anticipated
aggressively being evaluated by applications in new pay-zone will be more than 10 feet thick, and she may
Chapter 2
A B
Company Z CompanyY
100 .....--,---.,.--...,...---,...,.
w
g 10
:s
::e
I 1 ar---+--.....~~--.I...--~
~
t=
~ 0.1 Ir-~~--f----+-~f--f
G)
a:
§ 0.01 ~----i---l---=-t-----i ~.
..... 0.01 ~-+--+---::j---t
~ Y u
< 0
0.001 ~.........wu..--'-a...&.UILL-..l-L..u.LJIl"--l..,LL.lJ.U.I 0.001 L-L.J..LIJlJLiL..L..LJ.LIIJ,lL.-..l...L.LLJllUl-.........".~
0.01 0.1 1 10 100 0.Q1 0.1 1 10 100
Reserves Estimate - MMBOE Reserves Estimate - MMBOE
c D
Figure 2 Geotechnical estimates of prospect reserves,' modified after Capen, 1992.
be 50% confident that it will be more than 20 feet thick, productioIl~~te, decline-rate, wellhead price, drilling
but she is only 10% sure that it could be more than 40 costs, and the likE-;:" -
feet thick. The same procedure may be applied to any However, such estimates cannot be pulled out of
parameter-productive area, reservoir-yield, initial the air! They must rely on objective consider·ations of
11
8 Geotechnical Estimates under Uncertainty
Table 2 Bioses affecting judgments under uncertainty (modified after Rose, 1987).
Overconfidence Predictive ranges are too narrow, indicating that estimators are much less accurate than
they think they are. Symptom: frequent surprises about exploration results.
Representativeness Analog based on small sample size may not be statistically significant; chosen analog may
'not be analogous.
Availability Recent or spectacular.. f'_xamples are more prone to be cited, regardless of their real fre
quency in natL!rt:;
f
limited imagination limits number of possible interpretations.
Anchoring In estimating, the desired iterative-reiterative process is attenuated, so a low starting point
leads to a lower final estimate, and a high starting point leads to a higher final estimate. Con
servative estimators find difficulty in accepting the possibility of a large outcome.
Unrecognized !bits Geologists forecasting future discoveries may disregard nongeologic factors that affect such
discoveries. ".
Overoptimism Prospectors exaggerate magnitude of reserves or chance of success in order to sell the deal.
Conservatism Technical staff may think that overestimating a project is worse than underestimating it, and
therefore err on the safe side.
all relevant data, especially maps, cross sections, geo . . significantly less tJ.1an the ranges they think they are
physical data, borehole log interpretations, analogous· setting. A common result is that, for prospect reserves
production records, and the like. Moreo\'er, geotechni forecasts, the anticipated "low-side" reserves predic
cal professionals must arrive at a final distribution for tion is too large and the projected "high-side" predic
each parameter by repeated iterations-making trial tion is too small. The common operational symptom of
estimates, e"xamining the implications of various val the problem is that prospectors experience freguent
ues in the distribution, determining through credibil surprises on reserve sizes of discoveries (Capen, 1976),
_ ity checks and reality checks that the distribution as well as many other geotechnical parameters, Figure 2
makes sense, comparing it with analog data, consider indicates that t~e real ranges of new-field wildcat
ing the independently derived opinions of other pro prospect reserves uncertainties are commonly about
fessionals, and adjusting it repeatedly ~ntil finally two orders of magnitude (powers of 10) at about the
becoming comfortable with all the estimates in the dis 90% confidence level.
tribution, as constituting a "best fit" to the facts.
Conservatism
This bias commonly leads to underestima tes
Biases in Estimating under Uncertainty because professionals, fearing criticism if results are
Bias is a more serious problem in geotechnical fore disappointing, may think it is worse to overestimate a
casting than is the characteristic wide unc~rtainty, If project than to unclerestimaie it. The psyrhology of ,u1
the exploration company's decisionmakers consis unexpected upward revision in project profitability is
tently receive biased estimates concerning prospect much more pleasant than a disappointing downward
value, their investment decisions will be correspond revision. In fact, however, either error may result in a
ingly flawed, leading to suboptimal economic perfor loss to the investor (Rose, 1987). Overestimates result
mance of the company's exploration portfolio. The in overinvesting in projects, whereas underestimating
stockholder will suffer. Table 2 lists the most signifi may cause the firm to invest too little, or even decline
cant biases observed in modern exploration compa to invest. Either result is a loss.
nies (Rose, 1987).
Overoptimism
For the exploration company, three of these biases
are especially dangerous. This form of motivational bias leads to overesti
mates because of perceived career or economic self
Overconfidence interest on the part of the professional. The most
This bias typically leads to excessi vely narrow common example in exploration is prospeCtors inflat
ranges: Technical specialists think they know more~,;. . in~ estimates of prospect ~,;e~ves or probability of
success'ill order to "sell the deal ell":u. get~we prospect
j
than they reafly do, so they tend naturally to set '
Lognormality
Basis
Statistics is routinely taught to students by employ
ing the "normal" distribution-the \vell-known sym
metrical "bell-shaped" curve. Even though students of
mathematical statistics have long knmvn the signifi
cance of the Central Limit Theorem, the lognormal dis
a -variable (arilhmclic scale)
,
tribution in petroleum science has only gained wide
i
acceptance-and more imp6l:tantly, 'routine analytical
application-during the past decade.
The Central Limit Theorem states that distributions
resulting from the natural addiJiol/ of independent ran
dom variables will be "normal"-that is, afreqlleucy dis
tribUliOIl will tend to take the form of the familiar
"bell-shaped" curve, in which the vertical axis is ordi 1lXJ'; -i::~-+-~-1-+-1I-1--t--t-
narily expressed as a percent of the total, and the hori b -vari:lble (urilhmetic scale)
zontal axis is an arithmetic scale expressing some
variable such as dimension or value (Figure 3a).
Another convention for presenti.ng the same data is the
ell III Illative pro/;nbilitN distributioll, in which the vertical
axis is 0-100% and the horizontal Clxis displays a dimen
sional variable, using an arithmetic scale (Figure 3b).
111e power of the cumulative probability distribution is
that, conceptually at least, it represents the full universe
of all possible outcomes-100%-and probability is
expressed as a cumulative percent of some outcome c -variable (log sl:ale)
"equal to or less than" or "equal to or more than" a par
ticular value.] Thus the cumulative probability distribu
tion is especially useful as a predictive tool.
A special type of cumulative probability graph paper
has been developed on which the vertical probability
axis is symmetrical around the 50% probability, with graph is that a cumulative probability distribution that is
complementary probability intervals (40-50% and perfectly normal will plot as a straight, sloping line, .~ .
50-60%; 30-40% and 60-70°,{" etc.) that are equal but of .,Ib~~l:!I'\tmlLJn:lHJh~<:>r~!TI also provides that dis
increasing spans upward and downward (Figure 4a). . 'f.:' e.·Ai}.JYl!;l~.!!!!,I.n!p/iclltiOIl o£
Maximum and minimum probabilities are 1% and 99%, I U!~"I .. I"
rather than 0% and 100%. The horizontal axis is arith . tha t is;"tnepeqiiencyd;stribl~t~~"~ill'~~\lt~t:m"41
metic. TIle special property of the cumulative probability symrrtetrlcal'~'beU~haped"1:,(ffvew nere the horizontal
I I .1 I P1
\..
I
i I 'j' P2
i
j. I.. .1
I
.. i. .~ .
,.
I
P5
I I I
-
> I i I P10
i I· I
; .\ I P20
.0
co
.0
I i· .
I
i I ... ,. i
I
!
P30
~
0
a.
I I'"
i ··1
... !.
i I . [
P50
I i I
I. ,.
i !I .,! .
.'.
I· .,"j
Q)
-> '1
:.=
co
1
.. -- ..i
··1-- .1.l .... i ...
I
I
i-
i_. .,
"1
-- P70
::J
·r··- I
I' .
i
P80
E I
I
P90
::J
U I·
;
1
....
;
+, .. ! .i
P95
i
! ",1-, :
I.
I P98
P99
0 20 40 60 80 100
Figure 4a Cumulative probability graph; note vertical axis is nonlinear probability scale, whereas horizontal axis is
arithmetic scale.
i i
... I ,, P1
.i
! P2
I - .-1! ..-!._. .-1 L_
i ; --' • L
P5
1\
A
A
i i .. I
I I
I I
P10
>
.~
! ._ __ I
;
.~. _"
I- _:
i!
_. ~
i
_.• )
:
I L.
P20
:.aco I ~ .....j
P30
I' .~ . .
-j L_
.0 I
o...
1
P50
a.
Q)
-i.. .. 1- .
I I.
> ··.. 1 -
P70
'';:;
~ .i. .. / I. ---.I i_.
P80
:J .
E
:J
I P90
o I
; I
.1
-. .- t· -~ ... i . P95
.. "'j .... - - .., . .- _. 1-' : P98
P99
o 1 10 100
......-:'
Figure 4b Cumulative log probability graph; note vertical axisis nail/mear probability scale, whereas horizontal axis
is logarithmic scale.
C OJ.-A.O ~~
e2 Q~ \,~ d \s+"·ib,,c(ou. d-e ,\ os 'Oc:'>'~C:'",,;,(r~-ros ~,'e~.ectQv<.,i,
~
J,_
5..., ,_ OS ".)~~ ',,,, <1:- _---~ Co";,,,
\J - ter ",--11
t'00--1 --+c-
I ,; , " \ \ '/" [ ,", I _
0 Q.A/\:-!---o '--:'e. '":"",
Cr-.J-: C v"I 0 _l'1'-. \ (,
' i...A~ 1C>9_" \ C \ Q',A 0 -, <:'
0 Q C .\.,'
..,;, v
\ ( <
.J C'..
"0 '.
- ,(
-
'-/0' \AA ! ,\,A
axis is logarithmic (Figure 3~. When a lognormal dis should be constrained by the expected form of the
tribution is plotted as a frequency curve on a regular distribution.
coordinate graph (Le., arithmetic scale), it takes the
form of a severely right-skewed frequency curve (Fig
ure 3d). Another special type of graph paper has also Field-size Distributions
been developed for plotting "cumlllative log probabil Distributions of reserve sizes (projected ultimate
ity," in which the vertical axis is the cumulative prob reco\"eries) in fields in a given trend, play, or basin
-n
ability scale, as described in the preceding paragraph, show a pronounced tendency to follow a conventional
whereas the horizontal axis is a logarithmic scale:A* lo,gnormal pattern: . /
cUJf\1:!Jptive Pt()p~£ility (U~ti!u!!i9n that isp!i!_~fect~
lognormal will appear aa straight'-Sloping line (Fig • just a few very small fields,
ure 3e). Figure 4b shows a cumulative log probability • a great many small fields, ,J ,,-.,
graph; note that maximum and minimum values on • a handful of medium-size fields, and I I "
the vertical probability scale are 1% and 99%, and the • a very few very large fields
horizontal axis is a log scale. ,
~" - tip1t!;,~tipn of independent, randorrf The reason, of course, why field-size distributions
ognormal distributions. Milst impor (FSDs) are lognormal is that the parameters control
tant geotechnical parameters involved with oil and ling field size are multiplicative: Field Area x Average
gas occurrence are lognormal (Megill, 1984; Capen, Net Pay Thickness x Hydrocarbon Recovery Factor =
1984, 1992). Geoscientists who are aware of the Field Reserves. Several attributes of FSDs are'notewor
prevalence of lognormality (and who constrain their thy. First, they typiCally shift toward smaller sizes as
estimates in the expectation of lognormality) will exploration progresses (Figure Sa). Second, where
tend to make better predictions of most parameters manv small fields (1,000 to 100,000 DOE [barrels of oil
having to do with oil and gas reserves (Rose, 1996c). equi~"alent]) are included, the FSD may depart from a
A few petroleum parameters are exponential; fewer straight line on a cumulative log probability graph,
still are normal. Predictions of all such parameters taking a conca\"e form at the lower end (Figure Sa),
t--
- - - -
Mean, MMBOE
l-tl~ - ~1 H - -
- -
!
-r
-,
-
-
- -
- -
-:
- ~
3
~~ -
~-
Wt-
:t-
-
• -
Early 5.1 • ~
.Ji..
I P10
~.
Middle 2.9 • M
~ ra·
~
r-- - - - - ,~- - - -
:-~
Late ~l 0.9 -,'
,..;/'~"jj f-
~
r- - - - 1- ,-I - - -
r- - - - - - r- - - - t- r-I- - - - _ ." l'fI~
(.~~v.
f- - - r-
r'" P50
- - - l ~-
i
t-- fill II"
t- - - - r- H- -,fT··'·...
._ ',/1
0 , - - - f- - - r- -
r- - - - - - -
'
- - -
~D1i~ ~
- - -
1--- r- l-
i: f- r-
- - - p't'
r- - -
t- - - - r- f- ,. -rJ- - - - - - t- -
lq , ," r~
~ ~iI ~
t-
';l;! ~" I
P90
r- - - - - - - "I~ - - I-f- - - - t- - - - - I- - - r-
-T- li~
lJI
I~ - - - - - - -
- f- t- - - r- - - - t- - - r--
P99
0.001 0.01 0.1 1.0 10 100
MMnOE
Figure Sa Field-size distributions change as play matures.
12 Geotechnical Estimates under Uncertainty , ~.
PI
P99 = 0.1 - - f- - - - - - _ I -
P90 =0.2 I
P50 =1.3
- - f- - - :..... - -.1- -
~t
Mean =2.8 P10
PI0 =7.3 ~~
1/
.. I- - - f- - f- - - - -
PI = 30.5
I-
f-f - - - - -
'- - - - l- I- - I-
-.J-I ! - - - - - - - -
~~,.
P50
f- - - - l- f- -
,f - -
I-- - - - - - - -
- - - - ,~ - - - - - - - - -
-,
f- f- - -
t
- - - 111 1f- - - - r- - - - - - - - - -
I P90
-I - f- f- - - - - - - - - - - - - -
.'
,-- - f- f- - - - - - - - - - - - - -
P99
0.1 0.2 0.5 1.0 2.0 5.0 10 20
MMBOE
Figure Sb Economic truncation causes convex curve at lower end of field-size distributions.
because of incomplete sampling of smaller fields. Such severely truncated at the lower end by such minimum
smaller accumulations may be incompletely repre economic requirements. Parent distributions in such
sented in the population of discovered fields because areas contain very many un<;:ompleted accumulalions,
of economic and technological censoring: ordinarily reported as "shows," many of which were
not even testpd. If such a trend were located onshore,
1. anomalies recognized to be small may therefore howe\'er, many such "shows" would have been com
not be drilled; pleted as small fields. This point is elaborated further
2. discoveries recognized to be very small by testing on page 80 and in Appendix F.
may not be completed for production; and Construction of FSDs sheds great light on explo
3. small anomalies may not be visible geotechni ration of most trends and basins: They are recognized
cally, and therefore never drilled. as an indispensable tool by most modern companies,
serving as "reality checks" and giving essential per
When FSDs are truncated at the low end to elimi spective on proposed exploration ventures.
nate fields that are noncommercial, the resulting dis
tribution typically fits a straight, sloping !ine, but in
the lower part (the P99o/o-P80% sector), the FSD has a Calculating the Mean of
characteristic convex shape at the lower end as a con Lognormal Distributions
sequence of the arbitrary elimination of the small part
of the sample (Figure 5b). FSDs of trends in ecol"!~m!- Statistically, the best single representation of a log
cally demanding regions, such as the North Sea or ---~iH;lL... ' disl;,,4;ution is the mean, or average. Because
deep Gulf of Mexico, where only larger discO\·eries events in the low-probability end of the distribution
qualify for platform installation, have already been have disproportionately much gJ:eater "weight" than
Chapter 2 13
in the high-probability part, the arithmetic mean of a routinely set up models of reservoir behavior. Economic
lognormal distribution typically increases as sample models predict economic trends, given certain techno
size (n) increases. The statistical mean assumes a logic and market developments. Such models effectively
continuous distribution; that is, that It ::: and charac
00 widen our conceptual and predictive ranges by provid
teristically represents the largest possible mean value, ing flexible templates and characteristic associations that
On the other hand, if we simply calculate the arith would never ha\'e been available to someone using only
metic average of a lognormal distribution composed of traditional geologic or economic principles. However,
a small number of values (say 11 ::: 6, or n ::: 10), that experienced geoscientists and engineers have learned
mean will be smaller than the statistical mean. that utilizing models too literally can l~.ad to predictive /
A practical problem with use of the statistical mean errors; the lesSon is to maintain flexibility in interpreling~
in exploration forecasting is that extremely low new geology based on analog rAodels.
probability events (less than P1%), which have
extremely large values, contribute to the mean. But
such events are sufficiently unlikely that we are justi Multiple Working
fied in treating them as "geologically impossible." By Hypotheses and Maps
truncating such distributions above Pl%, the resulting T.e. ChC\mberlin's (1931) classic paper emphasizes
mean values are more realistic. the importance in scientific investigations of the con
A widely used alternative is Swanson's Mean scious identification and evaluation of independent,
(Megill, 1984), which works well for (1) 11 values con multiple working hypotheses. To the exploration
sistent with exploration experience (i.e., most trends mind, it offers a disciplined method to widen pre
do not contain an infinite number of fields), (2) distri dictive ranges because it forces the investigator to
butions truncated at the upper end, beyond P1%, and systematically construct and evaluate alternative
(3) distributions of low to moderate variance, includ interpretations of incomplete data sets. In its simplest
ing distributions truncated toward the low end by eco practical form, it requires the prospector to make sev
nomic threshold requirements. Appendix A illustrates eral possible maps of various prospect parameters,
and reviews various techniques· for calculating the showing optimistic, intermediate, and pessimistic
mean of a lognormal distribution. possible cases, or various possible structural or depo
sitional interpretations of the geotechnical data.
Techniques for Improving
Geotechnical Estimates
Table 3 Ways to improve accuracy and build
confi&nce in estimating.
EX~lora tion staffs can learn ~o i!1'prove their geo 1. USE OF <;;EOTECHNICAL MODELS AS ANALOGS
technIcal estimating performance by using at least 2. USE OF M~LTIPLE WORKING HYPOTHESES
seven techniques (Table 3). AND MAPS
3. INDEPENDENT MULTIPLE ESTIMATES
Geotechnical Analog Models
"Delphi Ro~n,ds"
Since about 1950, geoscientists have increasingly
Team Exploration
developed and used "al1alog models"-exceptionally
well-documented and'well-wlderstood examples of var Peer and Committee Reviews
i?us "type" geologic situations-to anticipate dimen Technical Subcommittees in Joint Ventures
SIOns, patterns, and associations of newly encountered 4. "NATURE'S ENVELOPES"
(an? therefore poorly documented) counterpart geo
Lognormality
logiC phenomena. The first such models were strati
graphie. One example is the very well-known carbonate Known Ranges of Parameters
faCies-con:plex of the middle Permian Guadalupean Plausibility Checks
shelf-margm of west Texas and New Mexico (Newell et 5. "REALIlY CHECKS"
aI., 1953; Pray, 1988). Another is the modern delta of the
Field-size Distributions
Mississippi River feeding into the Gulf of Mexico (Fisk,
19.54; Coleman and Prior, 1982). Stratigraphers familiar Historical Record
~I~h such models can often make far-reaching and Comparisons with Worldwide Databases
I~Slg~tful forecasts about newly encountered geologic Iteration and Tests for Reasonableness
sl~atlOns, even though very little prospect-specific data
6. PROPER STATISTICAL PROCEDURES
eXISt. Now we also have structural models, such as the
ba.lan:ed.structural models used to resolve and interpret 7. PRACTICE AND COMPARISON OF PRIOR
seIsmIC lInes in complex thrusted terranes. Engineers PREDICTIONS WITH OUTCOMES
'll
14 Geotechnical Estimates under Uncertainty ,
Independent Multiple Estimates to ensure reasonability and obtain a best-fit. FSDs pro
When we are estimating under uncertainty, the con vide such a "reality check" against which prospect ..
sideration and reconciliation of independent multiple reserves estimates can be compared. The historical
estimates of the parameter yields forecasts that are drilling record can provide a basis for evaluating esti
generally less biased and closer to reality than the mates of discovery probability. Comparison of pre
more orthodox procedure of devoting more lime, dicted prospect parameters against parameters
money, and technology to additional study by a single measured in fields of similar type in the trend or basin,
investigator. Modern exploration firms accomplish or against worldwide databases, can help eyaluate
those predictions. Comparison of the prospect's
jects, formal prospect review by a centralized gous prospect types also provides useful reality checks
2Allhough a few international companies han adapted Discussion, justification, and refinement of geotech
P5%-PSO%-P95% as consistently used parameters, most utilize nical estimates among professional staff prOVides an
PlO% and P90%, primarily because oi early published work by excellent way to clarify, standardize, and improve their
Megill involving applications of Swanson's Rule (see Appendix A), ability to make sound and consistent estimates of
and because estimators seem to be more comfortable with a gener prospect parameters. In addition, disciplined compari
ally used convention such as a 10% confidence than wit·~~,;, srna!~!:.
confidence such as 5%. Experience suggests tha t subjectil'e pwba .;,un of predictions with actual outcomes provides objec
bility estimates become increasingly tenuous at extreme probahlity tive feedback as to individual, team, and organizational
levels (Boccia, 1996). performance in predicting prospect parameters. This
Chapter 2 15
-c . .
,I
Chapter
Ri~k Ana-'ysis
of
Exploration Prospects
/
.~~
discovers a producible hydrocarbon accumulation, and a straight, sloping line is "best-fitted" to the data,
\vhich may become an oil or gas field. As discussed forcing the distribution to be lognorrilaI. The PSO%
•
;
r
earlier, this value does not equate to "proved," "prob case is consequential, derived from the intersection of
I able," or "possible" reserves, as formally defined engi the P50% horizontal line and the sloping line. Next, the
" I I
neering parameters (Capen, 1996; Cronquist, 1997). provisional distribution is projected out to the low
Those engineering definitions arose out of fiduciarY (P99%) and high (PI %) extremes, which are assessed
needs and are subject to continual revision throughol;t for their plausibility as being highly unlikely but pos
the life of the field, They involve considerations of sible outcomes. The PI % value should be so large that
reservoir volume as well as detailed reservoir parame -it is barely possible, honoring thE' data; similarly, the
ters, flow rates and decline curves, and multiple eco P99% "alue should be just large enough to be consis
nomic assumptions, Obviously, such details are tent with a very small detectable reservoired accumu
usually not available for exploration ventures. lation. A "final" distribution should be developed,
--- !
,I Accordingly, many companies'eniploy a simpler set following several iterations and adjustments, utilizing
of parameters (Figure 6) that are mOre consistent with given plausibility checks and reality checks such as
, r the high degree of uncertainty that attends exploratory those shown on Table 4. There is no "formula" for
•I prospects: Prospect Reserves = Productive Area (in deriving the [>90% area. It must be prospect-specific
! acres, hectares, or kilometers 2 ) x Average Net Pay and consistent with anticipated structural configura
\ I
Thickness (in feet or meters) x Hydrocarbon-Recovery tion of the closure and stratigraphy of the reservoir
I Factor (in bbl or md [thousand cubic feet] per net acre
I section.3 In general, however, the P90'Yo area should be
,
!, foot, bbl per net m 3 jhectare-meter, or m 3 per net
2
small, corresponding very roughly to an onshore, mar
km -m). The parnmeters shown in figure 6 are deter ginally economic area of drainage (assuming pay
i
i
ministic; that is, single-value estimates for each para
meter, all of which, because of substantial geotechnical
thickness and hydrocarbon [HC] recovery are suffi
cient to warrant completion)-a one- or t\\'o-wellfie\d.
uncertainty, are much better forecast as a probabilistic Based on theoretical grounds, empirical observations
range of possible outcomes. Deterministic predictions (figure 8), plots of actual fields in real basins (Figure
I are generally unreliable; fortunately, theiruse in the 9), and detailed analyses (Squire, 1996), the distribu
modern exploration industry is diminishing. tion of area forecasts is lognormal in form.
Productive Area In practice, geoscientists seem to be able to arrive
I at high-side estimates for prospect area (Pl"!o and
I Utilizing geologic and geophysical data in the PlO°!.,) fairly readily. However, settling on appropriate
prospect area, professional staff are asked to make
maps shOWing a reasonable high-side productive area,
3ElongatE'd trap areas nnd resl:'Tvoirs having low nE't-ta-gTl15S ralias
given optimistic geologic conditions such as seismic
will tend to be associ~ted, of n~'Cessil}', with larger 1"90% areas than
velocities, ;liip nltes, contour configurations, fault equidimensianilJ closures and high net-to-gross ralias, as described
extents, trap fill-up, and the like. They also construct in Appendix C.
11
17
.:> Risk Analysis of Exploration Prospects
/
Area
(1200 ac)
x
Average Net Pay
r
----
(75 ft)
(200 bbl/af)
= Prospect Reserves
(18,000,000 bbl)
Figure 6 Reserves parameters for exploration prospects (deterministic).
low-side values seems to be more problematic. The derived values are implausible, the line is
most common error is that estimates of P90 0 0 and adjusted until a credible best fit is achieved. This
P99% areas are too large, leading to frequent o\'eresti approach is preferred, especially for less experi
mates of mean area and thus mean reserves. But the enced prospectors.
determination of the area distribution cannot be "for 2. Assuming all plausible conditions maximizing
mulaic"-it must be consistent with the data bearing productive area are operative (velocities as they
on prospect geometry, and it must be prospect affect dip rates, contouring, trap fill-up), the
specific. This is discussed in detail in Appendix C. resulting maximum outline of that area is
Two general e~ti.mating approaches seem to dominate: mapped and measured; this "max area" is provi
sionally assigned to Pl%. Next, the smallest pos
1. Estimates are made of optimistic (PIO%) and pes sible area is assumed (consistent with the
simistic (P90%) cases, which are then plotted on a geologic attributes of the prospect) that would be
log probability graph, and a straight, sloping line consistent with a reservoired small He-accumu
-C'. • :lra\,~~l)et\\'eenthe h\'o points. ;;;ds lin~-1.; then
:... lation just large enough to sustain flow. This area
extended out to the Pl% and P99% extremes, is assigned a provisional P99% value. Both \'alues
which are evaluated for plaL!sibility. If any of the are plotted on a log probability graph, From the
Chapter 3 19
-
I
.,
I
~1: !. !: .~ -- ...: ,:. !-I" .i~ ..-AI::], .;"i P30 ~
'i I" .-' :'- - -': - --: "15.1"
1---........:.._-..;...-4-+.. .:. .:- :-.----:-~\O~ 0 .
ir.. '- "'2BO" . rr-'- - - , '-il-!-·:-,-r.;-:_---;.._ _~-'--'_;_:___j :;~
30: I '.
;
0
""1 ... ~~~«) .. "';~'1~/. i'-' '~i I'" ~/"- ...... !:: pao
I! i; if?
1----...--~'*-----_o20 -Iff-.
:0.6 . of 40/ I: ~-. i i:'i i , P90
i 811150-+--'-::'-.'-.-'------~....,."""j.'""i
V
-'
I .. i : / l;~.
. ~ I . ,!~I I! . .. .,{!.•.
• I
. f ,'" , : '-"'",!
.! i. '·
I i . P95
'://1 i, '~;
I I :
; I !, ;1';:
L£~P9~8~·}7",:~:,,:~~1i2=---.lI..-L.I·J.·i . .:.·i. A:·~:
-_:_._.__._.:....;._._'_"_LI·...:,·i_·_·.-•
...J1..l.·I_-_·._. ·_-_.....L_ _ . .i. !. .:.i. .:.~·-"- \.:. i '_._---'-.:..;'.'-l! : : :
Area (acres) }. .
Average Net Pay (feet) = Variable In:reasm g
He-Recovery (Bbl/acre-foot) '.
Figure 7 Area, average net pay, and He-recovery factor are lognormal.
-',
distribution line, the consequent P90%, P50%, estimates should be derived in an analogolls manner to
and PIO% values are derived and checked for the area estimates previously described and employing
credibility. Again, the line is adjusted as neces similar reality checks (Table 4). This parameter also typi
sary to produce a plausible best-fit distribution. cally follows a lognormal distribution (Figure 8), but one
ordinarily having less variance than the area distribution.
It shouid be recognized that, following these meth In order to consider porous and tight intervals, reservoir
ods, any ou tcome larger than the PI % forecast is net-ta-gross ratio must be considered. Also, a geometric
treated as impossible, which is not strictly true but adjustment will need to be made to take into account the
pragmatically can be dismissed as such. Thus no value geometry of the oil/water contact in relation to the reser
larger than the PI% value can contribute to the mean voir geometry (Figure 10). For most new-field wildcat
of the distribution. (NFW) prospects, a distribution of estimated average net
Average Net Pay Thickness pay thickness (determined as described earlier), through
out the productive area will suffice. It is essential to
Using lithofacies, isopach, porosity and net-to-gross understand that the "average net pay thickness" estimate
maps, geologic studies of pertinent depositional models, integrates internal reservoir distribution, porosity cut
and analog field studies, and considering the interactive offs, trap fill-up, net-io-gross ratio, and the geometry of
effects of the oil/water contact, geologic structure, and the "top-reservoirloil-water contact" couplet.
reservoir distribution,' the exploration team generates a
probabilistic distribution of estimated average net pay Gross Rock Volume
_...~J,r,01.!,g-hout!hf- area of the accumulation based on Plo;;" A useful alternative to the Area x Average Net Pay
PIO%, P50'Yo, P90%, and P99% confidence levels approach outlined previously is to estimate the proba
plotted on log probability gr~phs (Figure 7). Such bilistic range in Cross Rock Volume (CRV). There are
1 ~
N
o
Table 4 Reality checks (7): characteristics of the endpoints of the reserves distribution.'"
"Low End" "High End" ;:I:l
;;;
;I:"
Absolute Reasonable Onshore Absolute Known Max Values :>
for N American :::::l
Minimum (P99%) Minimum (P90%) Development Maximum (Pl %) DI
Well Fields -<
II'
;;;
AREA Area too small to be Approximately the area of Onshore Honoring the data, the 500,000 acres (oil) ....
0
i m
t': economic anywhere; an onshore field that is production maximum area possible )(
"C
roughly the area of a economically marginal; spacing units: if all relevant geological 5,000,000 acres (gas) .,DI0
t very small one-well field.
Must be consistent with
typically about 1-3 onshore
production spacing units. May
10-60 acres (oil) and geophysical factors
are most favorable; area
...0
'i.
:::::l
expee:ted trap geometry. vilry according to structural 160-640 so large as to be barely "'til
,configuration. Could'be larger acres (gas) possible. a
II'
in a frontier play area. "C
AVERAGE Pay zone just thick Roughly the minimum More than 3 feet· Maximum average net pay'" 1,000' oil
...
III
n
1ft
NET PAY enough to sustain flow thickness you would (= 1 meter) thickness possible based
of mobile HCs consider completing in on regional isopach and
sufficient to sense in an onshore field. Could be net/gross maps, and
mudstream or logs, thicker in a frontier play area. considering geometric
and sustain on a D5T. factor. Thickness so large
as to be barely possible.
HC-RECOVERY Ordinarily less than Minimum porosity and More than Compare critiGlly with 1,200 bbl/af (oil)
FACTOR 100 bbllaf; too tight for permeability to sustain flow, 100 bbl/af (oil) highest projected;
an economic reservoir, or to be commercial onshore, HC-recovery factor 2,000 md/af (gas)
even onshore; sufficient consistent with pertinent .~ More than known (or analogous) for
permeability to barely data in trend. Pay attention 125 mcf/af (gas) trend. Also justify against
flow. Must be consistent to values in analog fields. high-side estimates of
with known reservoir 0, So, % Rec., and FVF.
properties in trend.
-
ULTIMATE A small, noneconomic A modest onshore field More than Compare critically with the 16,000 BOE
RECOVtERY BOE accumulation with only of about 1-3 wells; could - 40,000 bbl oil largest existing field in the (Prudhoe)
:' ~nough reserves to be larger in a frontier trend, or reasonable analog; 6,000 MMBOE (E. Texas)
barely flow; comparable play area. More than check against trend F5D;
( with a mediocre onshore 250,000 MCFG must be prospect-specific. 70,000 BCFG
one-well field. (Hugoton)
*origir,ally developed by the writer, and vetted by geot ~chnical staff from Conoco, Marathon, and Mobil.
j
....
Chapter 3 21
30
30 U 30 30 U.
N
L N L
U
T U T
.",....
I I
M M M
B
E 20
'T"
A B
E 20
T
A
R 20 E 20 E
R
~.,
R R
0
E 0 E
f
C f C
0 0
f
V v /
. -. I 10
F
10 E - 1 10 10 E
R
. R
E
Y E Y
1.
L
",1 L loll
0
0
B
B
S
B S
B
L 0 o L
S S
0 100 200 300 400 500 >500
30 30 U 30 30 U
N L N L
U T U T
M I I
M M M
B A B A
E T T
20 20 E E 20 20 E
R R
R R
o E o E
F C C
o F o
F V V
F .
..... /10 10 E I 10 -10 ~
R
E Y Y
E
L loll L loll
o o B
i S
B
B S B
~ L L
f o '00 200 ~OO 400 ~ 600 700 800 900 1000 1100 S a 2~ 75 100 IZ!:I 150 17~ 200 225 250 275 >-275 S
I
I (c) He-recovery factor, in Bbl/acre-foot (d) Recovery, in thousands of barrels per acre
~
both advantages and disadvantages to this alternative. support s;,s,tained flow, provides a reality check
The advantages are: consistent with P99% for CRY; and
5. Plotting these upper and lower values as PI %
1. It eliminates minor problems caused by the com and P99% on log probability graph paper allows
mon partial dependency of average net pay derivation of PIO'}';" PSO%, and P90% values for
thickness on productive area; CRY.
2. Calculation of the geometry factor required for
correct expressions of average net pay can be dis The disadvantages, however, are more compelling:
pensed with; CRY is a complex number combining area, thickness,
3. Integration of maximum (PI %) area with associ and reservoir Iwater-level geometry considerations.
ated vertical column height and gross reservoir Accordingly, geoscientists find it quite difficult to relate
thickness ("area vs. depth plots") will generate a CRY intuitively to maps or cross sections and, there
maximum gross rock volume that represents a real fore, often don't recognize whether their probabilistic
ity check on the upper limit for CRY (Figure 11); estimates of CRY may be implausible. Moreover, when
4. Consideration of minimum pay zone thickness CRY estimates are checked in post-audit drilling
(taking into account expected netl gross reservoir reviews, it is not immediately apparent whether errors
ratios and required HC column heights), together were related to area, thickness, or geometric misjudg
with the resulting associated area of accumula O10nts. The fundilmental problem here is t;~.~ eX}:7!'0'
--(:". 1
tion that would be required to provide a reser rationists conceptualize prospects in relation to maps
voired accumulation of sufficient volume to (area) and logs and cross sections (thickness), and
"
1,000,000
4
10,000 I
t/)
...
Q)
(,)
- '--- I~'
<!
1000
.. +•• P I
100
-' ~~ _., ._._.•..... ~
I
_ _....•..\_......
:
. .. ....- _.. _............ _....• -_._... . - __ .
10
1 2 5 10 20 30 40 50 60 70 80 90 95 98 99
Probability, % Less Than
Figure 9 Productive field areas, East Texas, Capen.
~ (
(011
TYPE 1
DOME
TYPE 2
ANTICLINE
11
0 TYPE 3
FLAT -TOP
11
TYPE 4
FLAT • TOP
/
II" .
TYPE 5
BLOCK (or
(or CONE. (or PRISM, DOME ANTICLINE VERTICAL
PYRAMID) CYLINDER) CYLINDER)
1..JW 1 2 I >10 1 2 I >10 N/A
Modified aller Gehman et al. (1970)
Result
0.1 1--;------;
0.0 ......--;---+-..;--l--+--I---r----+--'"
I Geometric Correction Factor 0.94
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Nole: LengthIWidth ratios of faulted, compartmentalized
.r'.• _._-==-""' Geomelric
...........~.....:...:; Correction Faclor
. ~-.:.::.:..=.:...::..:.:::.:...... ---J
structures are taken as if the missing segments were
restored.
~ -7, 600 1
~
-8,000 T-~-.'~----+----------,"""---""
=
~
~
~
-8400
,
_-L.
;
I
I
.. . , _
Cl -8,800 -- - -- i
~.-._.
I
_.. -
~
I
~
I
rJJ -9,200' .. __._._.----._-_._----_. j. .
§ i
rJJ -9,600 --_ .. , ..••.• - . - - - - _ .• ._~. __ •• __ . . . . . . _1.
therefore they have a much better subjective apprecia gas per acre-foot 9f reservoir. It is readily adaptable
tion of the credibility of such estimates than of their to t~e metric system, or to expressing oil in metric
combination as GRY, which is a parameter having three tons. 'Some companies choose to break this parameter
variables (area, thickness, and geometry), and thus'hav dovm into its four constituent components: porosity,
ing a wide range of possible combinations capable of HC-saturation, percent recovery, and formation vol
giving the same product. ume factor.'This is not wrong, but for most
In fact, utilization of both approaches often prf) exploratory ventures, it represents false precision
vides valuable cross-checks for consistency, and usu and inefficient use of geotechnical effort: He-recov
ally leads to better estimates of prospect reserves. ery factor is eqtirely adequate given the high level of
That's why use of both procedures is recommended. uncertainty that attends most exploratory prospects.
In any case, however, GRY, combined probabilisti However, many companies choose to estimate the
cally with estimates of reservoir net/gross and He four constituent components as reasonable PIO%
recovery factor, yields a lognormal probabilistic -P50%-P90% ranges, and combine them, via Monte
reserves distribution that should be compatible with Carlo simulation, as a reality check to ensure that
the reserves distribution derived via the "Area x Aver HC-recovery factor estimates are indeed credibl¢ val
age Net Pay x He-recovery Factor" method. ues consistent with known or postulated reservoir
Estimation of both the PI % and P99% cases will parameters. Also, expressing yield as bbls orind per
require construction of an "Area vs. Depth" plot, as acre-foot is compatible with the widespread use of
shown in Figure 11. Such plots are readily constructed analog field models in modern ~xploration. Again, a
by measuring the areas enclosed by successive con range of rational probabilistic estimates is generated
tours. "Depth" is understood to indicate the vertical by the exploration team, analogous to the process
d~fference from the crest of the feature to the projected described previously (Figure 7), and plotted on a
oil/water contact (or gas/water or gas/oil contact cumulative log probability graph, then iterated and
where appropriate). adjusted until a best-fit is reached. This parameter
Hydrocarbon-recovery Factor also takes a lognormal form (Figure 8). The compo
nents of HC-recovery factor and metric equivalencies
This parameter expresses "reservoir yield" as bar are shown on Table 5. All estimates should be real
rels (bbl~ of oil or mef (thousand cubic feet) of natural ity-checked (Table 4).
II
24 Risk Analysis of Exploration Prospects
FVF %REC
·Sw=water saturation; FVF=formation volume factor; % REC=percent recoverable; BAF=barrels per acre-foot
!--+---rlI----j 10%
/----t-'-'f---j-----j50%
/
PH}%
I
=600 ac PSO% =200 ac
'----\o;--'--~'-----'99%
\0 1000 10.000
(P99'" 25) Area (acres)
I O' Ill'!,
.t. ;.A..r'
P90% = 10' SO'k
40'
PSO% = 20' 90'7c
Ill'k
BAF =ban'els per acre-foot (also haj) (P99 = 60) L::==:±::~~L-_--.I 9'Yk
10.· 100 IllOO
HC·Ree (bal)
HC-REC =
01'7<
'---~~--'-"""'"':. 01 'Ie
~
10'k /
I()'i;
/ I-----j----jr'"llr--, 10'lr
/
1/
5ll'h /
50'i;
/ 50'k /-----j--~r__-"---jSOCk
J
1/
If
90% 90'i! 90'k
L
10
')9'i!
I
99~f
I 99'h
/---Hr--r---"---j 90'k
00:
100 IO()OIO.O()O.1 I III 100 \0 100 1000 '47 99<;f.
=
M IJbl III ilf iOlI b a r r e l s .LO-'M ---'-.-If\-1--'I-.l-)~-1~ IOM
Al = millioll ill this figure
.3 (0.2) + .4 (0.8) + .3 (3.3) =1.37 iVI M = Mz
Figure 12 Graphical method for analytim/solution for combining three lognornlal distributions by multiplication.
and the chance of eco/lolllic success, given that flowing and ecoTi0illil. s:'" :ess,~....,.?ectively, represents the chance
hydrocarbons are discovered. The proportioll, as a per of finding those volumes of reserves (or more), giveJl that
ce~t: of the reserves distribution that is larger than the reservoired mobile Ilydrocnrbo/ls are foulld at n,1l. This con
mInimum reserve amount \leeded to justify cOll1111ercial cept is discussed in further detail on pages 39-41.
26 Risk Analysis of Exploration Prospects
t
Table 6 Reality checks (2): characteristic ranges associated with oil and gas ventures having different magnitudes of
uncertainty. * 1
Wildcat Rank Rank
in Known Wildcat Wildcat in
Statistical Development Step-Out! Productive in Proven New Play or
Parameters Well Extension Trend Trend New Basin
P10%/P90% 2.2 --) 7.0 5 --) 25 10 --) 120 55 --) 220 120 --) 650
. )
Field-size distributions (FSDs) Eor the basin or play continuous efforts to improve such forecasts because
(see p. 11) provide effective reality checks against of their profound impact on project profitability.
which the prospect-reserves distribution should be Two methods are currently in use. The first (Rose,
compared. Also, the variance of the prospect-reserves 1987) employs simple log-log cross-plots (Figure 13) of
distribution should be compatible with the explo prospect parameters such as:
ration well class, as shown in Table 6. But the variance
of the prospect-reserves distribution is also a function • prospect area
of the quantity and quality of pertinent information, as • average net pay thickness
well as of the geoteclmical skill of the professional staff • HC-recovery factor (or its constituents-porosity,
evaluating it, and all such factors must be considered. oil saturation, percent recovery, and formation
For all these reasons, the prospect-reserves distribu volume factor)
tion need not have the same variance as the parent • initial produttion rate
. FSD. In fact, prospect-reserves distribu tions com • percentage decline
monly (but not always) show variances sliInller than • prospect reserves (::: EUR)
their parent FSD. Because FSDs tend to get smaller as
exploration progresses (Figure Sa), prospect-reserves By plotting the median of predicted parameter distri
distributions should be compared with fields that butions against the actual outcomes, we can, with a rel
ha ve been recen tl y discovered, using com pa tible atively small number of tria,ls, gain insight into the
exploration concepts and technologies. variance (scatter) as well as any existing bias. If staff are
unbiased, we should have about the same number of
overestimates as underestimates. About 10% of out
Monitoring and Improving
10'
u
.:!.
z
°ii 10' 10'
5
:IE
::> ~
..
u
u
Z
~
0
..
~IO'
<5
~
;10'
~
'"S ..
::>
::l
~
'"
"z "~
;;;
;;; 101
0 ~ 10'
~
2
2
!
l5 /
/ /1:/ !
w
oIII
<'l
o
I
e(
t>
~ 10' --- - - .. --.------t-----..r-----.7.. . ------- ~ IO'I--...l-----+-----.l~---"7!''---.----7'9
:i'
:>
i /
...l
~
I / Z
o
z ? . "'a:«
2
cr
IO)~--+- glO'
gcr a:
,.
a
J:
o en
>
:z: ~
<Ii :>
z If)
w
o
i= ~ 10'
u z
Ci :::;
w ...l
cr ii:
c..
Cl
a
t
Z V>
:J
..J ~
a: +-----~._-._ .
~o
c.. • 10 1 }Ol IO( 10" ~ ~ ~ 10' ~
PREDRILLING PREDICTIONS. HYDROCARBON·RECOVERY FACTOR (BOE/AC.FT) PREDRILLING PREDICTIONS. HYDROCARBQN VOLUMES (103 BOE)
(c) (d)
reserves-added against chance-weighted mean suspect that there are serious problems with reserves
reserves of all prospects in the program. A cumulative or chance predictions by geotechnical staff;~ and cor
"actual reserves-added" curve is plotted in relation to rective measures must be instituted. This second
an "envelope" bounded by the cumulative PlO% and method requires Monte Carlo (or Latin Hypercube)
P90% predictions as well as the "expected" (== mean) simulation to add (rather than 11I11ltiply) prospect
curves. A predictive envelope for numbers of new dis reserves parameters of serial exploration ventures. The
coveries can similarly be constructed, which can reveal
bias in "chance of success" forecasts. Any time the
~A s'e~jes of outcomes that collectively filII below the P90% bound·
actual performance goes outside the forecast ary would have less than a 10% chance of being the result of ran
PlO%-P900/0 envelope, management legitimately can dom chance alone.
"
'0'
.0
/, ..D
~.ot'~~
--
Do '0 P.rcent
§ to •••••..•••••.•••••••••••••;:.:.::,:."' .
". 10 Percenl ' . 10 P*,cenl
~ 10 .a •.•..•..•..........•.....
(a) (b)
25 .
to I'D- to p.rc.n~
..
P.rcent
o_
-G
0-
-0- E~pechtd
Go!
. 20
-0- EJlpDcted
z-
Q
E u
:J ..
... Acluel E u
:J
z- Q
.. "
... Actu.r
to 20 30 40 SO SO 70
Hum"' 0' Welle
(c) (d)
Figure 14 (a, b, c, d) Continuous tracking of organizational predictive performance (Copyright SPE #22038. From
Clapp & Stibolt, 7997, reprinted with permission.)
only limitation to this method is that additional analy of the new reserves predicted. Moreover, this bias is not
sis, such as cross-plots as described earlier, are neces improved by modern technology: as an example, BP
sary to isolate and identify the specific causes of Amoco (Harper, 1999) reports persistent overoptimistic
estimation bias because the method reveals only bias, bias in their deep-water exploration program, which
not its root causes. employs state-of-the-art seismic technology, since the
Such routine monitoring of actual perfbrmance can early 1990s (Figure 16)'. BP-Amoco's successful deep
be instituted in exploration organizations if (1) man water ventures found only 45% of the reserves pre
agement insists on such activity being carried out, (2) dicted. Before addressing the real causes of reserves
results are circulated openly within the organization, overestimation, it is important to emphasize that the
(3) geotechnical staff accept unbiased predictive per problem is /lot caused by comparison of "before-discov
formance as part of their professional obligations, and ery EURs" with "after-discovery booked (or fiduciary)
(4) results are used COllstructively, to improve staff per reserves." Exploration estimates of reserves before
formance, rather than pUllitively. drilling and after discovery relate to the same parameter:
EUR. At least six real causes are responsible for this bias.
Industry Experience Motivational Bias
Persistent Overestimation Imli·,' iJ. ........: prvspectol s, as 'iidi II::; their ma/lagers,
~
s... ] ,000
~~
50
~~
.- ~ 100
e~
/
o
.- ~
N
CF.1
~
~
,.j
!
\;
°'1
0
~
~
~
'-'
"'0
~
~
~
;>
- - Actual Volumes in All Discoveries
r----I/
I
~
.,j
0
~
v
~.
~
~
.~
I
~
~\
<
-,~
."
0 I _......
.,j en
<' "\
~
Q)
E I ; -'
"
~-<
-::l
-----"'
L';
_.,{ Z ,
'",
--.} ~;
.......001
------------" ,
'r .{,
j 0 25 50 75 100 125
<~
reserves from those prospects that are not. But they The few oil companies that can report unbiased
usually cannot identify large prospects th.at are signifi reserves estimations (McMaster, 1998) are organiza
cantly underfilled, thus containing small-volume tions whose staff professionals know that manage
reserves. Furthermore, other parameters such as aver ment desires their objective geotechnical estimates,
age net pay and HC-recovery have such a large impact that management is monitoring staff predictive perfor
on reserves and are so variable that our ability to pre mance, and that unbiased predictive performance will
dict them with precision is limited. be recognized and rewarded. Such staffs have
received and utilized training in estimating proposed
Deterministic Estimate! R!1ther parameters, and they know the patterns leading to
than Probab1l1stic Ranges bias. They typically accept professional accountability
Single-value estimates of uncertain parameters pre for objective estimating. Nevertheless, it must be
dict an outcome that is possible, usually optimistic, acknowledged that overestimation of prospect
and nearly always wrong. reserves is a widespread industry bias that has proved
difficult to eliminate Qohns et al., 1998; Alexander and
Using Triangular Distributions Lohr, 1998; Harpe.. , loQ9) The single most effective
Rather than Lognormal Distributions remedy is to ensure that the P99% values reason"cly
Triangular distributions are a very poor proxy for represent a very small reservoired accumulation that
the prevailing lognormal distribution and usually lead would be tlowable.
to sub>tantial overestimation.
Reducing Uncertainty-What Is Possible?
Nonrepresentative Analog Field-size Distributions With regard to reducing uncertainty, companies
FSDs are very useful as reality checks, indicating that rigorously employ methods like those described
the ch;l;acteristic sizes of fields that exist in a basin or earlier report that they can consistently reduce predic
trend. However, it is important that the FSD employed tive ranges of reserves from 2 to 3 orders of magni
as analog is in fact i1 valid eX<lmple: it must comprise tude at 90% confidence, to 1.0 to 1.5 orders of
fields that have been discovered in cun:ent or recent magnitude at80% confidence. Obviously, data quality
Chapter 3 31
°
+$10,000= EV
$10,000 Correct call +$20,000 - $10,000 = +$10,000 x 0.5 = +$5,000 /
trial Incorrect call 0- $10,000 = -$10,000 x 0.5 =.. - -$5,000
)
i O=EV
$4,000 Correct call +$20,000 - $4,000 = +$16,000 x 0.5 = +$8,000
trial Incorrect call 0- $4,000 = -$4,000 x 0.5 = -$2,000
+$6,000= EV
and quantity have substantial influence here. For the emphasize that for most companies involved in oil and
components of prospect reserves (area, average net gas exploration there are many ventures, each with an
pay, and HC-recovery), capable geotechnical profes uncertain outcome. Furthermore, the employment of
sionals should expect consistent predictions within . EV as a decision criterion encourages repeated trials,
about 1 order of magnitude at 80% confidence, and so that EV is the average profit per decisioll, assuming
they should strive for accuracy of 0.5X to 2X at 80 0 0 repeated trials are made. The power of the EV concept
confidence. However, consistently achieving such is tha tit allows (1) a high-risk prospect with large
levels of predictive performance is, realistically, resen'es potential to be compared with a low-risk
unlikely, using conventional exploration methods, prospect ha\'ing small reserves potential; and (2) an
However, where geologic conditions render applica excessively risky prospect (one with a negative EV) to
tion of 3-D seismic data feasible and economic, be identified and avoided.
. reserve forecasting at such levels may be achieved, Faced with choosing al1"\Ong several options, the
especially if amplitUde anomalies and other DHIs decision rule is to select the option having the highest
reduce the likelihood of very large and very small EV. Obviously, when operators choose to participate
reserves outcomes. in \''entures hadng negative expected values, they are
gambling-"in effect, betting against the house."
In exploratory ventures, the cost of failure usually
..........
Chance of Prospect Success includes dry-hole cost, cost for lease bonuses of the
condenmed leases, and some geological and geophysi
The Expected Value Concept cal (G&G) costs. For development ventures, there may
."" Imagine that you have the opportunity to participate also be sUbstant.ial additional capital investments plus
in a simple game in which you are asked to correctly expense items that will have to be written off
call the fair toss of a coin. If your call is correct, you will expenditures that were needed in order to determine
win $20,000; if it is incorrect, you will win nothing. the Yiability of the project, such as several confirma
If you were able to play such a game free of charge, tion or delineation wells, equipment, materials, and
the Expected Value (EV) of each trial would be supplies. Ne\-',;endorp (1975) presents the subject of
(+)$10,000 (Table 7). If you had to pay $10,000 each expected value very thoroughly for the reader who
time you played, the EV would be zero, so that, statis wishes additional background.
tically, you then would be "trading dollars." If you
were willing to invest $4,000 in one trial of this game,
the EV would be (+)$6,000. In this example, there are Requirements for a Corporate System
only two possible outcomes, and you are restricted to to Estimate Geologic Chance
one trial. The chances of either outcome, as provided In order to calculate a prospect's expected value, we
by our knowledge about coin-tossing, are essential to must have a basis for estimating the chance that nature
calculating the EV of the venture. has provided a detectable HC-accumulation in the
In order to calculate the EV for an exploratory well, objective section under the drilling location. Geology
we will use Our knowledge about petroleum geology anq geophysics provide that basis. Moreover, for most
to estimate the chance (= our confiderlCe) that a reser companies, many such prospects will be proposed
v~ired petroleum accumulation is in fact present and annually, from many different basins, each competing
wIll be encountered by the drill bit. It is important to for precious corporate capital. In order to construct a
Ii
/
..
. ..... , .
32
I
Figure 17 Exploration failure and exploration success (economic, commercial, and geologic).
consistent system for evalua ting all prospects equi reserves and flow rates-that is,· the system
tably, 10 requirements must be met: must work just as well in a mature onshore
province, such as the Permian Basin of west
1. The system must be geologically sound, so that Texas, as it does in an economically demanding
all geologic aspects of oil genera tion, migra tion, province such as the North Sea; and
reservoir emplacement, containment, preserva 10. All prospectors utilizing the system must
tion, and geotechnical detection are considered; be trained so they understand its basis and
2. The system must be readily usable by many dif application.
ferent geotechnical professionals;
3. The system must apply equally well in all geo
logic provinces;
Recorded Success Rates versus
4. The system must apply equally well to all types Geologic Success Estimates
of petroleum traps-structural, stratigraphic, Geologic success (Pg) is not necessarily the same as
combination, basin-centered (generational), commercial success (Pc) or even economic success
hydrodynamic, etc.; (pe). The well-known phrase "geologic success but
5. The system must <lpply just as well to explo economic failure" refers to this issue-there me differ
ration plays as it does to the individual ent definitions of success. According to official explo
prospects that constitute the play; ra tio~!Jlli~~~~ics, s~chjls thos.~.J~p_Q.r-ttd
6. Chance, expressed probabilistically as geologic a®uallY-.EY.~ta~~c!-~~!!2.12~petroleumagencies, the
confidence, must be expressed numerically, not conventional definition of "success" means simply
subjectively; that the subject well~~_~omp'!eted an:!- did produce
7. The system must relate to reality (i.e., calibra some hydrocarbons. This does not mean that the ven
tion) by periodic comparisons of portfolio out turemadea-proll1TIn fact, such standard definitions of
comes (actual success-ratio) against forecasts "success" contain five possible "successful" outcomes
(predicted success ratio); (Figure 17):
8. The geologic components of chance must be
independent of one another, or, if dependency • The \vell was completed as the discovery well for a
is suspected, its influence must be understood field in which average wells will generate sufficient
anti 6t;·.;,; . . ted: __ production revenues to recover the cost to drill,
9. The system must be independent of economic complete, and operate them (as well as the sunk
requirements such as minimull0- required costs to find the field), plus a reasonable profi t.
Chapter 3 33
This is an ~~ss,on a full-cycle basis, rep e'conomic requirements, such as deep overpres
resented as PV > 0 at the finn's discount rate. sured tests, offshore prospects, remote bontier
There are three possible outcomes comprising ventures, or international contract areas with
commercial s/lccess or comp/etio1l success, wherein severe financial obligations; and
the exploratory well was completed but was not 3. It ensures approximate compatibility between
profitable on a full-cycle economic basis: onshore completion success and minimum
• The well was completed because anticipated detectable reserves volumes, assuming that
future production revenues will return a profit on minimum but finite dimensions are required for
the cost of completing and operating it, but not on area of accumulation, average I1et pay, and HC-/
the costsof exploratory drilling, leasing, and seis recovery factor. The concept here is that som~
mic, which are thus viewed as sunk and not small but finite volume and minimum reservoir
recoverable (Capen, 1991). Such a well is an incre quality must exist for an accumulation even to
mental S/lccess. Ordinarily, no more wells would be detected by an operator. In other words, the
be drilled on such a prospect by the investor, lower limit of an accumulation, thus defined,
assuming that subsequent events do not provide is substantially larger than one barrel of oil!
new encouragement to drill again. Two independent lines of evidence suggest that
• The well was completed as either an incremental appropriate minimum reserves values are
success or an apparent economic success, but sub 'genera-By quite small (consistent with trap
sequent performance was inadequate even to geometry): Drilling experience indicates that
recover completion and operating costs, resulting reserves volumes in the range of perhaps 1,000
in early abandonment; completion of-such a well to 20,000 BOE are generally capable of being
was dearly a mistake. detected and sustaining flow into the borehole,
• The well was completed only for business reasons; at least for a few hours. Also, P99% values in
that is, to hold a lease position or to satisfy a con FSDs in mature onshore trends and basins are
tractual or regulatory obligation. Some production commonly around 1,000 to 10,000 BOE (see p. 30
revenues will be recovered, but perhaps not even and AppendiX C).
enough to cover completion and operating costs.
• The fifth outcome \'I.'e call ge%gicsllcces5. meaning Chance of Success Equals "Flowability"
!b.at a reservoired accumulation was found that As previously stated, it is important to recognize
was at least large ~nough to support a flowing that the conventional reporting standard for
test. F2!..l!lQ~L9~.~_~!.~ncesany w.elTThat exploratory success in most petroleum-producing
f1.?y.~~nLsJH5g1y.,.~0.,b:'_~?E~!et~9-L bu_~n ysucn natiorts is not whether the exploratory well discovered
small reservoirs encountered offshore are often a commercial new field-rather, it is simply whether
rep~rI~~,o.ijiY'assl1ow~~-, - _..-_ - ..
-. The geologic chance factors (Rose, 1992a, 1995) are
the exploratory well was completed for production.
The beauty of such· a reporting standard is that it is
unequivocal, and results of all wells ordinarily become
defined (pp. 34-36,38) so as to exclude an onshore well matters of public record. In the case of expendable
hat discovers a petroleum accumulation too small to exploratory weJIs, most governmental agencies
;;warrant the expense of completing and operating it require the operator, within a reasonable time, to
(Figure 17). Such very small accumulations are com declare whether or not the well encountered commer
monly recognized or,ly as shows. Practically speaking, cial hydrocarbons.
we h~ve eliminated this class by introducing concepts Moreover, different companies have different crite
of minimum dimensionality or volumetrics into the ria for commercial success, thus rendering profitability
definitions of the geologic chance factors, using the to be an inconsistent criterion among exploratory
U.s. onshore as an effective minimum standard. This wells drilled by many different operators. So, in oider
accomplishes three important purposes: to compare geologic chance of success estimates with
actual outcomes by all companies, we have set up the
1. It allows the geologic chance factors to yield a geologic chance system to be consistent with the
product that corresponds effectively with the chance of discovering enough oil or gas to complete
world's most liberal definition of chance of com the well for production in the case of an onshore,
pletion Success, thus permitting the comparison mature petroleum province (Rose, 1992a). A practical
of conventional reporting standards of success proxy for this case is "encountering enough reser
with independently derived geologic estimates voired oil or gas to sustain flow." This criterion is
of probability of success; widely used by most exploration firms as the basis for
2. It p':"c"irle~}l genf'~1 and basic standard against exploration success because it is independent of vari
which subseque;t adjustments can be made for able economic requirements of different trends or
exploration projects having more demanding basins.
Ii
34 Risk Analysis of Exploration Prospects
ing the HCs (Landes, 1951; Dolt and Reynolds, 1969). . multipTIcation of the chance fnctors then yields th~ geo
All three of these general requirements must be met logic chance of success for the prospect.
for a HC accumulation to form; if anyone of these A few companies have used geotechnical chnnce fac
requirements fail, no accumulntion will be present. tors to focus on the chance of failurl', belieVing that bet
This pnrndigm becomes the fundamental basis for ter results may be obtained by focu~ing on the specific
employing geologic chance filctnrs in estimating prob geologic conditions that could cause project failure.
ability of geologic success (Table 8). In its simplest Expressions of geologic confidence are subjective
form, each of the three geologic chance factors is probability estimates, and they depend on three factors:
treated as an independent variable having a probabil
ity ranging from 0 to 1.0. 1. reliability of evidence (direct, intermediate, indirect);
Serial multiplication of all three factors produces a 2. judgment about adequacy relative to the P99'Yo
decimal fraction equivalent to the probability that a reserves case, and to the P90% cases for area,
Hc-accumulation is present, which is the probability of average net pay thickness, and HC-recovery fac
geologic success, or Pg, 0.28 in this example. By sub tor (more than adequate, adequate, less than ade
tracting Pg from 1.0, we get its derived counterpart, quate, inadequate); and
the probability of geologic failure, or pf. These expres 3. professional experience in estimating chance factors.
sions, along with estimates of prospect \"alue and fail
ure cost, are needed to calculate the expected value of Geologic Chance Factors
the venture. Recommended System
Explorationists commonly ask whether one geo For all exploratory prospects, including extensions
logic chance factor is more important than another, (= step-outs), deeper-pool or shallower-pool tests, and
thereby deserving more weight. The answer is defi NFWs, explorationists should independently express
nitely 110. The chance factors should be thought of as their confidence, a's probability, in five critical geologic
links in a chain: if any link breaks, the chain fails attributes of the prospect. Each of the five geologic
(p. 62). By analogy, if anyone of the geologic factors is chance factors has several subcomponents that must
zero, the prospect will be dry! In other words, all of the be considered in arriVing at a confidence estimate for
essential geologic chance factors must coincide the chance factor. Because most of the subcomponents
in space and time if one or more reservoired HC are partially dependent, we recommend use of the
accumulations can occur. Coincidence is discussed "weak link" approach, in whiSh the lowest probability
further on pages 48 and 80. assigned to the several subcomponents within one
Many different schemes and combinations of geo chance factor is us",d as the probability for the parent
logic chance factors have been proposed and utilized by chance factor (Rose, 1996a).
Table 8 Example calculation of simplistic probability of geologic success and failure using three geologic chance
factors.
Hydrocarbon Source Rocks prospect area in sufficient volume, porosity, and deliv"
First, we assess the probability (= confidence) that erability to support one or more flowing wells.•Reser
thermally mature HC source rocks are present in ade voir thickness and quality in the exploratory well b.:,-:.2
quate thickness, extent, organic richness, and type to must be consistent with the P90% forecast for prospect
provide at least a modicum of HC-charge to the average net pay and HC-recovery factor. Essential
are:
• \'olume (thickness, extent),
• quantity (thickness, extent, organic richness), • porosity, and . /
• HC type (oil, natural gas, mixed), and .... reservoir performance (permeability, drive mecll J
Discussion In many frontier basins, confidence in Discussion Note that the resen'oir chance compo
HC generation may be relatively low, and therefore it nents are set up such that some minimum threshold
is often one of the most important requirements that standards of volume, porosity, and deliverability
must be met if exploration is to proceed. In productive must be met or exceeded-specifically these reservoir
basins and established trends, however, confidence in componeJ,.lts must be adequate, on an independent
this chance factor tends to be considerably higher. HC basis, to allow detectable, sustained flow of reser
volumes generated must be at least large enough to voired HCs into the borehole. Under this approach,
satisfy the P99% reserves value of the prospect. For encountering a water-bearing, reservoir-quality sand
confirmation wells, step-outs, and development wells, stone would not be a failure in the reservoir category,
this requirement has ordinarily been met. but rather, a failure in one of the other four categories,
such as an unexpected structural low, an absence of
Migration
HC-charge, or a leaky trap. However, the presence of
Second, we evaluate the probability (= confidence) a 1-ft tight siltstone where a lO-ft porous sandstone
that HCs have migrated, utilizing conduits, carrier beds, objective was forecast would constitute a failure in
~ractures and/or faults, folloWing migration pathways resen'oir prediction. Reservoir thickness must be
mto the location of the existing c!osure(s), in volumes of compatible with at least the P90% average net pay
oil and / or natural gas sufficient to charge such clo forecast for the prospect, and porosity and reservoir
sure(s). Components to consider are: performance with the P90% HC-recovery factor fore
cast (pp. 24-25).
• conduits (carrier beds or zones, fractures or \
faults), Closure
• migration routes from "kitchen" to prospect area, Fourth, we assess the probability (= confidence)
• efficiency (cdncentration during transmission vs. that a structural and/or stratigraphic closure involv
dispersal), and ing the reservoir objective, and of minimally adequate
• timing (sealed closures existed when migration area (consistent with the P90% area forecast) and verti
occurred). cal relief, is present in the prospect area, and can be
detected using' current geotechnical means. Compo
. Discussion Again, this chance factor may be an nents to be considered are:
Important uncertainty in frontier exploration, but it is of
much less concern in known productive basins and • Closure exists and is of adequate area and vertical
trends, and it may be entirely satisfied in cases of confir relief to contain a volume of reservoired HCs suf
mation wells, step-outs, and development wells. For ficient to support flow (given that reservoir rock is
tion was dispersed, or concentrated by "structural • We have confidence in our ability to detect and
focus," whether it was primarily vE:l'tical or lateral, and delineate them using available geotechnological
confidence must be compatible with at least the P90% quantified one. A second reason is that many profes
area estimate (pp. 24-25). sionals have never been trained in techniques of subjec
The closure chance factor is formulated to apply tive probability estimating, nor have they been
equally well to stratigraphic as well as to structural traps. encouraged to examine the accuracy of their prior pre
In combination with the reservoir chance factor, it dictions. Third, a significant chance always exists that
focuses on the geometry of the envisioned oil or gas such predictions mny turn out to be wrong, and our
accumulation and on the volumes of fluids necessary to cultural and corpornte values often associn te scientific
sustain a production test or prudent drillstem test. Stated error with mediocrity or even moral turpitude, thus
in this way, the scheme can apply to all types of traps. generating criticism, guilt, and loss of status. Yet a ;
fourth reason is that geoscientists may not wish to
Containment acknowledge the high degree of uncertainty involved
The fifth geologic chance factor is the probability (== in petroleum exploration. still preferring to believe that
confidence) that containment has occurred-that the secret to explorntion success lies nlmost entirely in
effective sealing rocks are present adjacent to the geotechnical skill nnd effort. Finally, we hear the very
reservoir, rind that emplaced hydrocarbons have been common excuse: "We don't hnve enough datn to make
preserved. Essential considerations are: an estimate." Unfortunately, explorationists lIever have
"enough" information-this is inherent in the business!
• seal effectiveness (differential permeahility, seal Moreover, the more uncertainty that attends a given
thickness, absence of open fractures), prospect, the more a systematic expression of subjective
• preservation from subsequent spillage (fault leak probability is needed-not less! All that can be reason
age, later fraeluring, breaching, tilt-and-spill, etc.), ably requested is the best objective estimates possible,
and given the time, skill, and budget available. This is
• preservation from hydrocarbon degradation (bio indeed the professional explorationist's responsibility.
logic degradation, oxidation, thermal destruction).
Discussion Three issues are addressed here-first Practical Aspects of Implementation
the question of sealing capability between reservoir No substitute exists for actual experience in assessing
and top seals, seat seals, and lateral seals, whether and estimating confidence (= probability) in the various
formed by stratigraphic contrasts, diagenesis, or fault geologic chance factors, and comparing forecasts with
gouge, and compatible with at least the P99°;, reserves outcomes. Figure 18 relates various subjective phrases
forecast (pp. 24-25). The second issue concerns later used by the writer in relation to a complete probability
geologic events that may have resulted in leakage scale, which may help the novice get started. One point
and/ or flushing of HCs from the trap. The third issue should be emphasi'zed here: the use of probabilities of
deals with degrada tion of reservoired HCs by biologic, 1.0 for any geologic chance factor involving a NFW
chemical, or thermal agents. For most (but not all) con prospect in new trends or basins should be reserved for
firmation wells and step-outs, most of the "contain those cases where the positive evidence is overwhelm
ment" requirements have been met; for de\'elopment ing. Often the focused question-"What could hurt me
wells, all have been met-otherwise there would not regarding this chance factor?"-can illuminate prob
be a field to develop! lems that otherwise might be oyerlooked.
Figure 19 is a matrix widely used among many
Subji;~tiveProbability
companies, whose origin is not known to the writer.
Assuming that a geological model, or concept, is recog
Estimates in Exploration
nized, it compares (1) quantity and quality of informa
Expert judgments about the probability of disco\' tion against (2) what the information is signifying with
ery of any drilling prospect are classic examples of respect to at least minimal adequacy about the partiCll
subjective probability estimates, \Nhich some geoscien lar geologic chance fador. A partial consequence of Fig
tists resist or even reject (Rose, 1992a); they claim it is ure 19 is that in order for us to render judgments of
just guessing. The record argues otherwise. Given a high confidence, either encouraging or discouraging,
logical procedure, knowledgeable explorationists can we require considerable data of good quality. Con
generate such estimates with surprising consistency, versely, sparse or poor-quality data frequently allow
agreeing not only on discovery probability but also on only intermediate confidence statements. Many explo
the relative certainty or uncertainty of the se\"eral geo ration teams find this matrix to be quite helpful in con
logic chance factors in a given prospect. sistently assigning confidence values to various
Geoscientists have many reasons for their reluctance geologic components of chance pertaining to their
to estima te the chance of success. One has to do prospects. Above all, geoscientists should recognize
• with the traditions of geology as an observational that the absence of information does not, by itself,
and descriptive natural science, not a predictive and imply a negntive outcome-only that there are T~O data.
Chapter 3 37
....··WiII Occur
.. . ..
'
1.0
"VIRTUAL CERTAINTY"
.9
I
.8
MORE LIKELY TO BE PRESENT THAN ABSENT )
.7 (="REASONABLE GEOLOGIC CONFIDENCE")
.6
.s
.4
.::::::::::::~~~~~~~'::}"SIG N.IFICANT UNCERTAINTY"
3 LESS LIKELY TO BE PRESENTTHAN ABSENT
.2
.1 "HIGH-RISK GEOLOGIC FACTOR"
o --.
-'"" - Won't Occur
For development wells and most extension and in the high-risk factors ("critical risks") did indeed cor
field explora tory ventures, the HC source rock, respond to'the geologic reasons why the hole was dry
petroleum migration, ar.d trap/seal factors have ordi (Rose, 1987). Although dry holes are an inevitable
narily been met. Only rese'rvoir adequacy and the aspect of petroleum exploration, the' capable profes
structural aspect remain as serious unknowns. sional will usually-but not always-find that he or
. Rigorous analysis and discussion of real prospects she has correctly anticipated the real geologic reasons
~vlth a pe~r group are an effective method for acquir for failure. It not, this indicates that the geology of
Ing confIdence in estimating geologic chance prospects is not adequately understood or identified.
faLtors and probability of discovery. This also helps ,Again, the writer emphasizes that assessment of geo
explorationists standardize their definitions and technical performance regarding chance requires a
procedures. completed program of wells, not a single well.
. Naturally, accuracy of predictions on probability of Explorationists can also gain experience by making
dIscovery cannot be judged on a single prospect or estimates on wells drilled by competitors operating in
even on two or three. Only the outcome of a program of the same trend or basin. Commonly, the geoscientist
exploratory ventures can provide a fair indication as to has considerable knowledge about the subsurface con
whether the assignment of discovery probability has ditions attending such wells, which thus provides
been optimistic, objective, or pessimistic. Of course, if expanded opportunities for developing a useful expe
the program has involved many high-risk wells, a rience base for predictive confidence.
larger sample may be required. Finally, it is difficult to overemphasize the power of
However, results of even a relatively small pro independent, multiple judgments in assessing the geo
gram can be instructive with regard to correct identifi logic chance factors and the probability of discovery.
cation of several geologic chance factprs. You should Firms are well advised to obtain several different opin
review predrilJing projections of d'ry holes to see ions and to combine them into a final estimate of the
whether those geologic chance factors identified as probability of project success.
II
r
38 Risk Analysis of Exploration Prospects
RISK SEVERITY
t
!
<JI-----------
HIGH I MODERATE I LOW GOOD LOTS
t
Direct Evidence
HIGH
-
.1 - .3 .8 - 1.0
/
I
MODERATE .2 - .4 .4 - .7 .7 - .9
LOW .3 - .5 .5 - .6 .6 - .7
ICONCLUSIONS INTERPRETED FROM DATA I POOR SPARSE
Virtues of Geologic Chance Factors differing definitions of success was shown on Figure
Separation of the various components of geologic 17. Remember that the geologic chance factors (espe
chance allows them to be analyzed more thoroughly cially reservoir and containment) are purposefully
and objectively, and leads us to better geologic under defined to be consistent with the presence of a HC
standing of the prospect. Also, the identification of the accumula tion of a t least minim um reserve size (or
geologic chance factor having the lowest probability (= greater)~"at least enough reservoired mobile oil to
"critical risk") helps exploration management and staff sustain flow." Such a definition is equally operative in
focus on those items of greatest uncertainty. For exam a mature, inexpensive onshore province such as west
ple/ if closure is the critical risk element of a prospect, Texas, and in an expensive' offshore province such as
additional seismic or reprocessing may be warranted. the North Sea.
Conversely, when geoscientists see that additional data In a mature, inexpensiv p onsholt~ 'p~'ovince, .:ll1Y
acquisition can be curtailed because it is not likely to exploratory well that flows will probably be com
materially increase the chance associated with a given pleted for production, either as an eco/lomic slIccess,
geologic factor, exploration becomes more cost effec recovering at least all investments on a full-cycle
tive and usually more timely. But most important, basis, at interest (or more), or as a cOlllmercial slIccess,
explorationists cannot really analyze, and thus which may only return a profitable investment in tub
improve, their performance in predicting the probabil ing/ tank battery, and completion. In such theaters,
ity of discovery unless they systematically identify, geologic success (Pg) and commercial slIccess (Pc) may
forecast, and inspect the predictive results that attend well coincide. But the same onshore commercial
the several geologic chance factors for a portfolio of subsurface accumulation encountered offshore
exploratory ventures (Rose, 1987).
North Sea would be d~scribed only as a show. A very
much larger volume of reservoired HCs would be
Geologic, Commercial, and required there to interest the opera tor in installing
Economic Chances of Success an expensive platform and developing the field dis
The relationship between the prospect-reserves covered by the-;rd!, -e.::oug1;.~servoired, producible
distribution, the geologic chance of success, and oil to cover all exploratory costs, plus platform
Chapter 3 39
construction and full development and operation of distribution is represented by the solid blac~ sloping
the new field, as well as a satisfactory return on the line, which .b~s bE'~ truncated above PI %. Reserve
total investment. That amount of discovered oil parameters are: . ""
would qualify as an economic Sl/ccess (Pe). If a lesser
volume of recoverable oil is discovered, enough to P99% = O.05MM bbl
warrant platform installation and development P90% = OAOMM bbl
Mean Reserves" =
drilling but not enough to recover all exploration P50% = 4.7MM bbl 20MMbbl
costs, then that smaller discovery would qualify as a PlO% = 5-1MM bbl } Swanson's Mean =
/
commercial success (Pc). Any exploratory well that is PI % = 400MM bbl I8MMbbl
completed for production therefore qualifies as a 'di,trilmtion truncated above Ploo; refer to Figure 20 and page 13
commercial success, "~ommerce" being the produc
tion and sale of oil and/or natural gas. Only some of As the manager, you have asked the staff to esti
these are economic successes, on a full-scale basis. mate the chance that Prospect Alpha will discover
enough oil (or more) to justify completing it for pro
Estimating Chance of Commercial
duction-the chance of a commercial sllccess.
and/or Economic Success
Your angineering and economic departments esti
figure 20 represents the concept and process by mate that, in this location, at least 0.95 million barrels
\"'hich geoscientists may progress from chance of geo (M~l bbl) must be present to warrant a completion;
logic Success (Pg), to chance of commercial success this is the minimum commercial field size (MCFS).
(Pc), to chance of economic success (Pe). On the reserves distribution, 0.95 MM bbl occurs at
Let's assume that your geotechnical staff estimate P80 o o, so 20% of the distribution is smaller than the
that Prospect Alpha has a 3Q!'ochance of geologic SllC minimum reserves required, and 80% is larger.
cess (i.e., that the well will encounter reservoired flo\\" Therefore the chance of commercial success (Pc) is:
ing HCs). The Prospect Alpha cumulative resen"e Pc =0.8 x 0.3 = 0.24.
P1
.- !... .- ;·1·'I - -. _. - -- -
I;
P2
iI
- --.:. -T t: .! r. .
I PS
,
' I .I
I
P10
II
II
I.
i
i
-l-' -.' ..
,
P20
II
II
II
II P30
>
P4D
.n
ro
.n
PSO
Q
o
.... P60
OJ
>
P70
.i _
rc
::J
. . ..
PSO
P90
P95
P98
will be a dry hole. Furthermore, the surviving distrib Geoscientists and engineers seem to struggle with
ution is no longer lognormal because the lower part, the "Truncation Problem." The procedure described
containing the mode and smaller outcomes, has been previously is useful because it models the actual deci
removed. What survives has a form that resembles an sion-behavior involved, after an exploratory well has )
exponential distribution if plotted in the frequency made a discovery, and management selects the logical ,
i
domain. The mean of the commcrcial reserves distribu next steps.
tion must now be recalculated. The new distribution, Implicit in this process is the "sunk-cost" concept
from P80% to PI %, now has a new set of reserves that completion decisions must be made on a "point
parameters; note that in moving from geologic chance forward" basis because funds already invested in
to commercial chance, chance of success has decreased seismic, leasing, and exploratory drilling are by then
(30% ~ 24%), whereas the mean of the commercial gone-sunk. "Can we make a reasonable profit on our
part oHhe r:ese-r-ves distribution has iI/creased (20MM investments in pipe, stimulation, tank-battery, and
~24MM): lateral lines in order to bring the well into produc
tion?" For offshore projects, such investments may
Commercial P99%= 0.9SMM b~l } Mean Commercial also include platform cost as well, which is usually so
Commercial P90%= l.SMM bbl Reserves" = large tha t the" commercial" threshold may well
Commercial P50%= 7.4MM bb! 24MM bbl approach the "economic" threshold. However, in
Commercial PIO%= 62.0MM bbl Swanson's Mean = mature producing provinces, like the Permian Basin
Commercial PI % = 400MM bbl 22MM bbl of west Texas, or the Canadian plains, most wells that
'origin~1 distribution trunc~tL'd ~b,,\'e 1'1.... and belo\\' 1'80"0; refer 10 Figure 20 flow are likely .to be completed, so the" commercial"
threshold is likely to be at or just above the "geologic
Next, the Exploration vice president asks you what chance of success," and may lie substantially below
the chance is that Prospect Alpha will be an ecol/omic the "economic" threshold. Obviously no rational
discovery-that is, that the project will find enough oil investor will continue to drill a series of development
to earn at least the corporate minimum return on wells that are consistently commercial (rather than
investment on a full-cycle basis and pay for all explo economic) successes.
ration, development, and operating costs. Again your Some people are tempted to truncate all projects at
engineers and economists study the problem and the economic threshold, rather than the commercial
advise that an economic discovery will require at least threshold. This conservative position tends to prevent
7.SMM bbl (= MEFS). On the original prospect, participation in many projects that could well be prof
reserves distribution, 7.SMM bbl occurs at P40%, so itable. The criterion should be that the average (= mean)
60% of the distribution is uneconomic, and 40% is eco of the entire reserves distribution above the commer
nomic. Therefore, the chance of an ecol/olllic success is: cial cutoff is at least economic. For such a prospect,
Pe =0.4 x 0.3 = 0.12. then, we accept that the outcome may turn out to be a
Again, you nlliSt7ecalculate the mean size of the small discovery that is commercial but not economic,
economic-reserves distribution because you have but the average of all outcomes is clearly economic; that
deleted the smaller, uneconomic accumulations from is, profitable on a full-cycle economic basis. The pro
the parent-reserves distribution. The surviving distrib ject's NPV is greater than 0 at the firm's mandated dis
ution, from P40% to PI %, now has a new set of count rate.
reserves parameters: Each firm should determine its own consistent
guidelines to be used in predicting the "commercial"
P99% = 7.5MM bbl Mean Economic
cutoff for an exploratory venture. The "commercial"
P90% = 9.IMM bbl Reserves" =
cutoff is not a precise value, and there is no doubt that
PSO% = 22.4MM bbl 4SMMbbi
} producing rate must be considered as well as mini
PIO% = IIO.OMM bbl Swanson's Mean =
PI % = 400.0MM bbl 4SMMbbi mum reserves. Many experienced operators correctly
believe that the risk associated with the decision to
'original distribution lruncated above PI".. and belo\\' I'.j{)"o; rdt'r to Figure 20
complete marginal commercial wells is the single most
So, as the chance of success goes doWI/ to reflect underestimilted, potentially dangerous risk involved
decreasing chances of finding larger accumulations of after exploration has made a discovery. Much care is
a commercial or eeol/olllic scale, the commercial and/or warranted here.
Chapter 3 41
One-step versus Two-step Method for there is no false precision. Setting reserves thre~holds
Estimating a Prospect's Chance of Success _very low maximizes the likelihood of correctly captur
Two methods have commonly been used by modem ing l~d1ity. It allows a company-eonsistent process to be
oil companies to estimate a prospect's chance of success. applied to widely varying economic thresholds in differ
ent operating theaters, thus promoting valid ranking
The One-step Method is favored by White (1993) and capital allocations. Finally, this method allows real
and assumes that: ity checks by comparing a consistent geologic chance
· --- estimate-usable in all geologic theateJ,'s-to a compati-/
1. Modern explorationists can reliably distinguish ble universal public reporting standard. (Most modem
large reserves prospects from small reserves companies have now endorsed the two-step method.) ~
prospects, so the geotechnical exploration effort
.......
can indeed identify and drill only those prospects
capable of containing some designated "signifi Independent versus
-", 2. By setting "significant" limits of closure areas, Many exploratory prospects seek to evaluate a single
average net pay, and He-recovery factor, explor main reseryoirobjective on a projected closure. For most
ers can screen out unp:ospective parts of such ventures, all the geologic chance factors may be
'mappedtrends; and treated as if they are mutually independent. Although
3. Prospectors can reliably link "adequacy" of there are a few cases where some dependency (either
-" source rocks, migration, closure volumes, and posith'e or negative) does appear to exist, especially
seals to the chance of success. Thus, for chance between resen'oir presence and structural closure or
, factors such as source rocks or migration to qual
ify, the geologist must judge whether theywere
between seal effectiveness and structural closure, most
knowledgeable risk analysts agree that the geologic
adequate to provide enough oil to a prospect for chance factors, as defined, do not present any serious
it to meet or exceed that reserves quantity problems with regard to dependency. However, depen
--,
deemed to be significant. dency does become a potential problem in dealing with
multiple-objective exploratory prospects or with several
Advantages of this method are that explorers don't prospects contained within one exploration play.
have to worry about the low-reserves ends of FSDs. Dependency in play analysis is discussed in Chapter 5.
Estimating chance of success is a one-step process,
keyed to finding at least a field of significant (= eco Multirle-objective Prospects
-', nomic? commercial?) reserves size or larger. Prospects having multiple objectives tend to be
more attractive than one-objective prospects because
The Two-step Method is favored by Rose (1995) and the combined chance for at least one of the objectives
....., i~ necessitated because all three of White's key ~ssump to be productive is higher (see Appendix D). However,
hons require geologic resolution beyond modern tech a prospect having two objectives will not be twice as
nical capability (i.e., false precision). Instead, in the likely to be successful as a one-objective prospect-nor
two-step method Pg is keyed to a very small ("flow does its probability of discovery equal the sun1 of the
, able") reserves discovery (= P99%) consistent with doc
umented industry performance. Then, in a second step,
probabilities of discovery of the two multiple-objective
targets. If all the geologic chance factors for the two
the part of the prospect-reserves distribution that is con obj~ctives are independent of one another, then P(dis
sidered to be commercial (the percent of the distribution covery) and P(failure) can be derived by binomial
that is greater than the commercial reserves threshold) expansion as shown in Appendix D.
is determined and then multiplied by the Pg. As a result, In most such cases, however, some geologic chance
~l estimation of Pc (or Pe) is a two-step process: factors are independent and some are dependent (i.e.,
-, t they are common to both objectives of the prospect). In
-~ I
1. Use geologic chance factors to find the chance of
a small ("flowable") reservoired accumulation
such cases, Pg and Pf are derived via another two-step
process, as shown in Appendix D. Discovery probabil
(or larger); then ity (and often EV) for a multiple-objective prospect
2. Estimate what percent of the prospect-reserves having independent geologic chance factors is higher
, distribution is of commercial or economic size than for a multiple-objective prospect having one or
and multiply the geologic chance by this com more dependent geologic chance factors. Thus, the
mercial or economic chance. effect of dependent geologic chance factors is to reduce
prospect discovery probability. Dependent geologic
Advantages of this method are that the two-step method chance factors (i.e., common to both objectives of the
reflects real industry performance capabiJities multiple-objective prospect) most often include the
, i
--- t
!l
42 Risk Analysis of Exploration Prospects
structural aspect and the He-charge and migration Also, mechanical chance factors should be consid
aspects. The reservoir and containment aspects tend to ered, such as the chance of not gelling the well down
be independent. However, exceptions to this pattern to the objective, the chance of incorrectly locating the
are not uncommon. well, and the chance of geotechnical errors in map
An additional complication sometimes occurs when ping, logging, or testing. However, if you anticipate
a given geologic chance factor contains several subfac that the well will be redrilled if such difficulties occur,
tors, some independent and others dependent. The task you should make probabilistic provision for such trou
here is to assign relative fractional weights to the subfac ble costs in the cash-flow sched ule for the project
tors so thanheir product equals the probability of the rather than as a chance-of-success factor. Gen~rally,
p~irent chance factor. Appendix D shows such a case. In such considerations do not make a substantial differ
general, dependency among the subfactors is common; ence except in economically marginal prospects.
that is the primary justification for the "we"k-link" Some authorities suggest including a chance factor
approach in establishing confidence levels (see p. 34). that deals with the probability that the ex~""loratory
Although having legitimate multiple objectives well has been located and evaluated properly (Baker,
usually makes a prospect somewhat more attractive 1988). It is certainly true that significant fields <lIt:: u.:ca
economically, aggressive explorationists often are sionally discovered (or recognized) only after several
tempted to include secondary objectives that are only apparently unsuccessful penetrations. However, the
of marginal value. The improvement to the prospect's writer's experience is that, for most prospects, this
value is illusory, however, for several reasons: aspect can be covered within the closure or reservoir
rock ca tegories; tha t is, if the well turns ou t to be
1. For most multiple-objective prospects, several of improperly located, it is commonly perceived as a faii
the geologic chance factors are in fact dependent, ure to adequately assess structural or reservoir risk.
.thus reducing the apparent chance of success sig However, when <;lealing with frontier basins and
nificantly from the independent case. plays, it may be advisable to include a separate chance
2. The secondary objective(s) commonly represents factor assessing the confidence that the well will be
only a marginal or incremental completion that located properly and any productive reservoir zones
may pay for pipe, maintenance, and some part of will be identified and evaluated adequately. Careful
the drilling cost while not adding substantial new review of the stratigraphic column and consideration
resen·es. Such ventures do not add much "alue of active petroleum systems will reduce the chance of
and therefore should not be actively sought; overlooking a productive zone while drilling. In other
3. Dual-objective discoveries are often completed words, "serend~pity" may be a euphemism for less
and produced first in the deeper zone until than-thorough consideration of all stratigraphic possi
depletion, then completed in the upper zone. bilities and thus for incomplete risk analysis.
Because of the time value of money, this delayed
production often reduces the PV of the upper Expressing Business and Political Risks
productive zone significantly; and In assessing major projects that require large front
4. Dual-objective completions are relatively more end investments or long elapsed time between expen
expensive and require higher maintenance than diture and payout, the firm may wish to appraise the
conventional completions. likelihood of a severe and extended drop in wellhead
prices (or ri5~ in open~ting costs or taxes). Basically,
For all of these reasons, a multiple-objecth'e prospect the procedure here is to identify what sustained low
having a high chance of success should be \'iewed price levc1s--or elevated costs-would cause termina
with caution. The most attractive multizone prospects tion of the project, then try to obtain estimates from
are those in which both or most zones clearly stand knowledgeable petroleum economists about the prob
alone economically, and production from both zones ability and timing of such occurrences. The chance of
may be commingled. commercial success (Pc) is then multiplied by (1- the
chance of such economic failure). Less severe price
Nongeologic Aspects and cost fluctuations should simply be considered as
of Success and Failure variant cases within the project cash-flow model.
Political uncertainty can be expressed similarly.
Technical and Mechanical Effects Again, knowledgeable, objective political experts
Many "ariables other than geologic chance factors should express their opinions about the likelihood of a
affect exploration success. For example, firms that use change of regulation, law, or regime severe enough to
state-of-the-art technology seem to ha\'c rates~J.sl!C cause a project's termination or change its economic sta
cess much greater than firms that drill without benefit.... ttT;:"'Commonly, this probability should be directed to
of such advanced geotechnical guidance; this should the anticipated time of greatest vulnerability-after
be taken into account. large capital investment.s for development of discovered
Chapter 3 43
fields but before recovery of those investments via pro management is justified in suspecting professional bias
duction revenues-which can then be related to the cash affecting the estimates of chance, and in caning for
flow model. The chance of commercial success (P0cail "" immediate corrective measures. Although performance
then be mUltiplied by (1- the chance of political failure). tracking is ail. excellent tool for monitoring geotechnical
performance, it does not indicate wily predictions may
Monitoring and Improving
be biased. For that, we must hun to dry-hole analysis.
Predictive Performance
Dry-hole Analysis
If the exploration organiza Hon is serious about
Professional staff charged witl). monitoring and
impi"uving staff performance in estimating the chance
improving geotechnical productive performances
of success, management themselves must undertake
should collect and analyze all unsuccessful exploratory
equally serious procedural changes.
efforts. "\Thy were the dry holes dry? Was critical risk
(Le., the geologic chance factor having lowest proba
Universal Prospect Risking Scheme
bility) correctly identified for most dry holes? That is,
Management must insist on adoption and utiliza
\yere most dry holes caused by failure for that chance
tion of a consistent geologic risking system that meets
factor to be satisfied? One U.S. company found that
the requirements outlined .earlier. Geoteclmical staff
they were correctly identifying reservoir risk but were
must be trained in its use; periodic management audits
not recognizing structural risk (Rose, 1987), even
and (where necessary) retraining will ensure its con
though structural errors caused 43% of the company's
sistent, universal application.
dry holes-more than.a!lY other geologic chance factor
(Figure 21). Incorrect predictions of reservoir rock
Keeping Records
presence were responsible for 40% of the company's
Management should set up procedures whereby
exploratory dry holes, but resen'oir rock was correctly
forecasts of geologic chance of success (and individual
anticipated as the critical risk in 80% of those dry
chance factors) are preserved and routinely compared
holes. The same company also found that they were
against actual outcomes of exploratory wells. Indi\"id
consistenth- consen'ative in their chance-of-success
ual geoscien tists should be encouraged to make geo estimates (iOO~ predicted vs. 31 % actual, over two con
.logic predictions on competitor wells and to compare secutive years) mostly because their chance estimates
them with announced results, as a way to expand the of He-charge were pessimistic (60% predicted vs. 95%
sample size of their predictive experience and reduce actual). In retrospect such consen'atism was especially
the time necessary to begin monitoring, measuring, diff\cult to justify considering that almost all of the
and improving predictions of chance. company's exploration efforts took place in estab
lished petroleum-producing U.s. basins! If this com
Predicted versus Actual Success Rates
pany had properly assessed the chance of HC-charge,
Compare the average predicted chance of success for its predicted success rates would have matched actual
last year's portfolio with the actual success rate (Rose, results very closely,
1987). Is there bias? Are some exploration teams biased Previously (p. 38) it was emphasized that better
more than others? Why? Separate all prospects into three estimates of apt.:ospect's chance of success result when
groups-high, medium, and low risk. Then see what the the geologic components of chance are analyzed sepa
average success rate was of each group-are staff distin rately. A sf'rnnd advCl'1tage of this method is that tech
guishing high-risk prospectS from low-risk ones? nology can then be focused on the critical risk-for
example, if closure is the greatest risk, additional seis
Performance Tracking
mic data mav be a cost-effective way to reduce risk.
" The methodology ofCJapp and StiboIt (1991) (p. 26 Based on Ro;e's (1987) experience, a third advantage is
and Figure 14) for continuously monitoring reserves apparent: this approach then allows geotechnical staff
additions throughout the annual exploration program to improve their predictive performance by identify
may also be applied to chance-of-success forecasts. A ing and correcting prevalent patterns of error. Amoco
PlO%-P90°!., "envelope" of expectations is created, (McMaster, 1998), Santos Uohns et aL, 1998), and Uno
employing Monte Carlo simulation, for the range of cal (Alexander and Lohr, 1998) report analogous learn
numbers of discoveries (Figure 14c and d). Both the ing from their results.
mean and PSO% trends m::y be projected within the
P10%-P90% envelope. Then as the exploratory wells of Industry Experience in Estimating
the annual portfolio are drilled, the numbers of actual
discoveries are plotted as a line rising from the start of
Prospect Chance of Success
the program toward the completion of the portfoIT~; anc"¥ A;';"la' Industry Performance
are compared with expectations. Any time the actual In international theaters, NFW success rates since
performance moves out of the P10%-P90% envelope, about 1960 have been remarkably consistent at about
II
44 Risk Analysis of Exploration Prospects
INCORRECT STRUCTURAL
{
INTERPRETATIONS WERE CHIEF CAUSE OF
43% OF DRY HOLES
STRUCTURE RISK CORRECTLY
ANTICII'ATED IN 23% OF
'- _ _---' CASES
INCORRECT PREDICTIONS OF i
RESERVOIR ROCK WERE CHIEF CAUSE
OF 40% OF DRY HOLES
RESERVOIR ROCK { RISK CORRECTLY ANTICIPATED
IN 79% OF CASES
INCORRECT PREDICTIONS OF
TRAPPING
CONDITIONS { CHIEF CAUSE OF 13% OF DRY
f------,---------' HOLES
{
CHARGE WERE CHIEF CAUSE OF 3% OF DRY
HYDROCARBON HOLES
CHARGE RISK CORRECTLY ANTICIPATED IN 50%
OF CASES
Figure 21 Relative frequency of four geologic chance factors causing dry holes during a company's 7977-7978
exploration programs, plus performance by company geologists in correctly anticipating which geologic chance factor
indeed represented greatest technical risk (from Rose, 7987).
25% overall (Figure 22). U.s. exploration success rates thus improve their exploration success rates. ll1is
during the 1980s (before 3-D seismic) for all is accomplished through clearer resolution of geo
exploratory wells were commensurate (20-30%), but logic structure (which leads to improved location
showed considerable variation among different of exploratory test wells) and better discrimination
classes (Table 9). In particular, annual success rates for of reservoir rock and sealing rock distributions.
U.S. NFWs (onshore and offshore) ranged from 13% to Abo, positive DHI indicators allow enhanced con
18% during the 1980s. fidence regarding the presence of reservoired HCs
in the traps. During the 1990s, provinces such as
Impact of 3-D Seismic Data the North Sea and deep-water Gulf of Mexico
In those geologic provinces in which 3-D seismic report prevailing NFW exploration success fa tes of
da ta collection is feasible, discrimina ting, and cost 30% or more.
effective, it improves exploration performance in three 2. Second, estimates of prospect reserves can be
different ways, especially where DHI technology is improved through improved resolution of geo
also incorporated. logic structure and reservoir thickness and
extent. Also, DHI signals can reduce uncertainty
1. First, by improving their prospect chance-of-suc (variance) on trap volume by indicating approxi
cess estimates, explorers can be more selective and mate position of oil/water or gas/water contacts.
.......
.;.l~!j..7
.-."
l;;" '-"- ~--
2- 5,.5
IOJ'- ,
)
..............(;l........
,
~
<1:;1;;1 '0>, '\
!...l..i':':'" Cf ..~'\/( __.~'-~'
' I J
...~~ 1500C ./
Chapter 3 45
25,000 ~
-,
~.-
- . .,
26%
20,000
27% 24%
15,000
-.
I
--.
10,000
---. 5000
-,
"". 0
1960s 1970s' 1980s 1990-1999
•o DISCOVERIES
WILDCATS 12,250
2955
13,864 19,297 15,842
3734 5117 3794
Source: Pelroconsultunls
-.
"WILDCAT CHANCE" (RATIO OF DISCOVERIES TO WILDCATS) HAS
REMAINED REMARKABLY CONSTANT THROUGH TIME.
Figure 22' Global discovery percentages through time (excludes the US. and Canada).
,
3. Third, prospect profitability may be improved actuai experience and should call for the presence of
through optimum location of exploration and multiple attributes if they are to represent a discrimi
development wells, whicn results in higher nating basis for elevated Pg estimates, rather than just
initial production rates, larger iI1dividual routine amplitude anomalies.
-.
well ultimate recoveries, and fewer required
development wells. Nevertheless, 3-D seismic Characteristic Patterns of Predictive
may not be feasible or cost-effective in manv Bias in Estimat,ing Chance of Success
. , onshore provinces, especially ill tlIe early stag;s Several international exploration organizations
of exploration. that have adopted companywide geotechnical risk
analysis of all prospects report that, during the first
However, characterizing seismic anomalies as year or two, a prevalent performance pattern
"DHls" should be rigorous and calibrated against emerged: for high-risk NFW prospects, geotechnical
staff were overly optimistic and overestimated' true
chances of success (Otis and Schneidermann, 1997;
Table 9 Generalized success rates of various well
Alexander awl Lohr, 1998). This will be addressed
classes drilled in the U.S. onshore and offshore during
further on page 47. For intermediate-risk NFWs, those
the 7980s as reported by the CSD (Rose, 79920).
"
in the 20-35% range, actual success rates were gener
Well Class Percentage Successful ally about right and matched predicted success rates
. --, fairly closely. For low-risk exploratory ventures, those
.----. Development wells 75-80 in the 35-60% range, actual success rates were conser
All exploratory wells vative-more of these ventures were successful thtin
20-30 predictec:;l.. But for high-confidence ventures-those
Extensions (outposts) 40--45 in the 60-90% range-actual results v,:ere notably
In-field Wildcats 25-35 lower. Apparently, prospectors tend to be overopti
New-field wildcats 13-18 misticwhen identifying an exploration venture as a
sure thing!
Ii
46 Risk Analysis of Exploration Pro!pects
However, by maintaining (and circulating) records 2. Was the quality of their datil commensurate with
of predictions vs. outcomes, by making geotechnical yours? (This particularly applies to vintage and
professionals aware of the causes and consequences of acquisition parameters of seismic data.)
predictive bias, and by constructively addressing the 3. Remember that all wells drilled through the sub
reasons for error and bias in specific prospects, we can ject zone are counted as valid exploration tests of
improve forecasts of a prospect's chance of success, that horizon, even if they simply passed through
and reduce predictive bias if not eliminate it. the subjeCt zone on the way down to a deeper
exploration objective. Therefore, such wells may
Using Trend or Basin Success Rates not have been legitimate exploratory tests of the J
In some cases, where geotechnical staff or manage subject zone.
ment mistrust geologically derived prospect success 4. Why were the completed wells completed? Were
estimates (or lack the geologic skill to reliably derive the economic parameters of those operators simi
them), observed success rates (number of discoveries lar to yours? Remember that not all completions
-;- total number of exploratory wells) have been used as are legitimate economic attempts, and a small
a proxy. In some geologic settings, this procedure is company may complete a weil that a jarge opera
reservoirs are lenticular and beyond predictive geot 5. Trend success rates tend to decrease with time
echnical resolution. But in most trends, such obsen'ed (Figure 23); ignoring this pattern may lend to
success rates are a poor substitute for prospect-specific overly optimistic expectations from trend success
chance-of-success determination and tend to give mis rates, especially if your venture is in a fairly
leading results that are either overly pessimistic or mature exploration theater.
optimistic.
There are at least five important unknown aspects Accordingly, cau.tion is recommended in substitut
of dry-hole ratio trends: ing observed trend success rates for geologically
derived prospect estimates of the chance of success.
1. Were the concepts and geotechnical skills of Certainly they should not be ignored, but their best use
those prior operators commensurate with yours? is as a reality check after the responsible geoscientist
100% r - - - - - - . - - - - - - - . - - - - - - - . , - - - - - - . , - - - - ,
t 80% I - - - - - - - + - - - - - - + - - - - - - + - - - - - r - + - - - - - - - l
I \
~ II "
~ 60% 1--------+----i-~r_+_------_t-----':_-_Pr_--_1
V') ,. \
~ I \
u I \
U I \
~ 40% 1--------+-----,f---i-....L---\:+--L..L.-.l,........r;;:.....:-f---r--,:t---+-r--_1
~
«
U
o
...... \
\
== 20% I - - - - - - - - + - - f -..........+---:+--------+\---~~--+--~~
3
---
1965 1970 1975 1980 1985
........
ANNUAL WILDCAT SUCCESS RATE (%)'
CUMULATIVE WILDCAT SUCCESS RATE (%)
ANN""0A~~.cWs __
Figure 23 Success rates change with maturity--annual and cumulative success rates compared with annual drilling,
Niagaran Reef Trend, Benzie and Manistee Counties, Northern·Michigan (Rose, 79920).
Chapter 3 47
has carefully reviewed and edited such data so that cash flows, are-without realizing it-actually
those ventures from which the edited trend success exhibiting risk-prone behavior with respect to
rate ~';\S cr:"ku!att~ are truly comparable to the such exploration ventures. Their managers des
·...... prospects in the trend you are exploring. perately want to find such large, new fields, so
they are prone to approve such ventures when
Historical Changes in Trend Success Rates
e\'er they appear. Ambitious geotechnical staff
In most exploration theaters (basins, trends, and are therefore prone to overestimate the value of
plays), wildcat success rates change through time. Suc such ventures in their efforts to respond to per
cess rates are characteristically high during the early ceh'ed management needs. This ~s especially truel
phases of exploration while larger and more evident when exploration organizations have the wron~
fields are being found, then decline as the industry priorities-when they confuse the need to drill
searches for fields that are smaller and/ or harder t~ {pells with the need to add vallie.
find. Figure 23 shows actual data from a segment of 2. Risk-analysis tools and methods for evaluating
the Niagaran (Silurian) Pinnacle Reef Trend of north such high-risk, high-potential ventures are oper
ern Michigan, Although the discovery probability a ting a t the extremes of the risk spectrum
--. should certainly be estimated based on the geotechni reserves anticipated are very large, whereas
cal characteristics of the prospect itself, the prudent chances of success are very small. Both theory
explorationist will also consider the trend's state of and e"1"-erience suggest that our ability to respon
exploration maturity in arrivingat a final estimate of siblv assess uncertain events begins to deteriorate
Pg, Pc, and Pe. at the extremes of either cllallce or 11lagllitlide. For
example, most geotechnical professionals are
The Trouble with High-risk Exploration
"',\
better able to judge whether the chance of occur
During the 1980s, large, international corporate rence is 50% or 25% than they are at distin
'- explorers, such as Shell, Amoco, and Mobil, experi
enced a disturbing result from their high-risk explo
guishing between a 5% chance and a 10% chance.
Because people are conservative processors of
ration ventures: within the category ofall NHVs having a fallible information (Edwards, 1982), we tend to
'",
predicted chance of success of 10% or less, less tllan 1% of set the ranges of our predictive limits too narrow
-.
those velltmes resulted in discoveries. Amoco (McMaste'r
and Carragher, 1996) indicates that, since 1982, such
(Capen, 1976). Also, technologic risk-reduction in
certain areas may actually encourage managers
~\
high-risk exploration ventures have destroyed corpo to take still higher risks elsewhere. Finally, Boccia
..... rate value, not created it. suggests that the economics of higher-risk \'en
The same pattern was reported by all three compa tbres are more sensitive to underrisking than are
nies independently, and the total number of wells in lower-risk projects. The relative negative impact
each company's sample was more than 200. These of overoptimism becomes progressively more
results cannot be ascribed to vagaries of random sam se\'ere as the probability of success decreases
pling-Amoco reports that the probability of such a "It hurts more, on a per-dollar investment basis,
result occurring by chance alone is only about 1%. to be wrong by the same degree when the risks
Clearly, major company geoscientists and managers are high. OUF ability to dearly discern risk is the
dealing with high-risk exploration ventures tend to be weakest for exactly those kinds of prospects
seriously overoptimistic in predicting chance of suc where it needs to be sharpest."
cess, so such ventures are often uvervalued as invest 3. In order to have an acceptable chance of making
ment ventures to the detriment of corporate economic such large discoveries, many trials must be
performance and the stockholder. undertaken. Such high-risk, high-potential ven
"'. tures are not easy to identify and are expensive to
...... Causes
carry out. "In effect, high-risk exploration may be
Some probable reasons for these difficulties have failing because companies are underestimating
been summarized by Boccia (1996): the time and money they need to commi t to the
high-risk game" (Boccia, 1996). Accordingly,
1. Most large companies (the kind that carry out such companies may have difficulty adequately
most high-risk, high-potential exploratory ven diversifying high-risk ventures within their over
tures) systematically favor such projects because all exploration portfolios.
they offer the potential for large, new reserve 4. A fourth possible cause, not identified by Boccia,
additions-big new fields that represent oppor is more fundamental-explorationists and their
tunities for long-lived, large-profit projects . managers have not come to grips with the reality
because of app!.:.~.,ti!)n..~ of ac!:..:D nced technology that the remaining world endowment of undis
and, efficiencies of scale. Such big companies, covered fields is getting smaller, simply because
havmg large reserves bases and corresponding petroleum exploration has tended to find the
Ii
48 Risk Analysis of Exploration Prospects
III
giant fields preferentially (see Figure 1). Accord success less than 20% would be undertaken, it might
ingly, exploration costs must be constrained, con only encourage geotechnical staff to find creative ways
sistent with the potential profits of smaller fields. to elevate what might legitimately be a 15% pros;>ect
H is not that exploration for large fields must to one rated 20% or 25%. Also, geotechnical staff must
cease-rather, it cannot proceed in the same be accountable, and any official prospect review com
ways it did in the 1950s through the early 1980s. mittee should not have power to approve or condemn
We must figure out how to be profitable while prospects presented to them. Otherwise, no one is
exploring for (and finding) smaller fields. accountable! The expected value concept is a useful
5. Yet another probable cause has to do with inade screen here. If the prospect-reserves distribution and' )
quate geotechnical verification of coincidence geologic-chance esl)mates are well documented and
(p. 34). All geologic chance elements must coin venture EV ,is'strongly positive (even though a scruti
cide in time and space if reservoired petroleum nized credible chance of success is only 10%), then the
accumulations are to occur in a basin or trend. prospect should probably be drilled.
Especially in frontier areas, it is essential to map
the areas where the various geC'logic chance ele
ments are present (or probably present) and to
restrict exploration to those areas of probable
coincidence.
Remedies
What are some remedies to counter these problems?
Exploration Ventures
·--
.~
\
i
-,
have a good working knowledge of economics and
finance integrated into their geotechnical expertise.
reserve projects. E"en though they may crea te large
"alues, such projects may depress the overall present
--. value of the portfolio somewhat, although not as much
-, Time Value of Money
as the short-term projects will (Capen, 1984). The
selection of a discoun t rate that is too low is not as
and Discount Rates
detrimental to portfolio value as the selection of a dis
count rate that is too high.
Corporations invest in petroleum exploration ven
Cash-flow Analysis
Discounting is "compounding in reverse." The dis
The cash-flo"Y model of a proposed exploration
--, c?unt rate has one use and one use only: to exprcss the
technical predictions are a professional goal. hence and the sL'lected discount rate. This is, in essence, i
I,
As previously stated (p. 24), discounted cash-flow an after-tax profit number, incorporating investments,
(OCF) models for most prospects should also be run costs, taxes, \\'L'llhead revenues, production decline
on the P10%, P50%, and P90% reserves cases, with the rates, operating costs, and the time value of mOf\ey.
.J
goal of generating a probabilistic distribution of net The cumulative NPV for any proposed prbspect
present values (NPV) for the venture. In other words thus represents the venture compared against
(because NPV IBOE may vary with reserve sizes), we the perspective of the firm's present performance,
want to determine the mean of all NPV outcomes i,e" How much better is this projrct than our present
rather than the NPV of only the mean-reserves out performance (or average weighted cost of capital)?
cc;ne. Naturally, geologic and engineering parameters
such as area, well numbers, net pay thickness, HC
recovery, reserves, well initial production (IPs), and Problems with DCF Valuation
field life must be compatible with the corresponding of Exploratory Ventures
PIO%, P50%, and P90% resen'es cases." The key result
from the cash-flow model is the discounted cumula For more than 50 years, the petroleum industry has
tive net cash flow (= "present value" or "present routinely used OCF analysis to value prodUcing prop
worth") over the projected life of the field, erties and exploratory ventures. One difficulty with
Exploration costs and development costs constitute the procedur.e has been that deterministic values were-" "
the lief il/vestmelll cash-flow strl'llll! (Megill, 1988); ana used, e\'en for parameters that were known to be ".
lysts are encouraged to calculate these costs on an highly uncertain. Multiple runs using different values
after-tax basis, taking depreciation and ill\'estment tax for some parameters gave a spread or range of possi
credits into account. The /let incollle CllS/I-f10W stream ble outcomes; this process has been called "sensitivity
(also on an after-tax basis) is intluenced by production analysis." But no probabilistic values could be ':
revenues (declining), net re\'enue interest, wellhead assigned to the ranges. Software add-ins, especially
(
"
taxes, operating costs, and income tax pro\·isions. For Lotus'Eo with @Risk®, or Excel® with Crystal Ball®, uti ,'" -
'
'
production-sharing contracts, further lllodifica tion lizing Monte Carlo or Latin Hypercube simulation, ,
may be required to correctly provide for the state's now allow probabilistic expression of such parameters ~"'
share of production revenues, cost reco\'ery, and spe so that the cumulative NPV can be expressed as a
cial tax provisions. For each year, the difference probability distribution. Even so, most firms at year
between afler-Iax il/vestmellt CIlslz flOil' and after-lax end 1999 still utilized multiple runs of deterministic
income cas!! flow is alllllwi after-tax lIel cas/! flow. Ordi cash-flow models.
narily, annual net cash flows are negati\'e in the first Another problem is that each cash-flow model rep
years of a project and positi\'e in middle and later resents only one sequence of events from project
years when the field is fully developed and \vells are inception to abandonment/often spanning 30 years or
in their producing (but declining) lives. Sometimes a more. It is not possible to build into one cash-flow
project may anticipate a second stage of im'estment in model the possibility of several alternative scenarios
a field's midlife to install an enhanced oil recovery that might develop during the life of the field. This dif
(EOR) program or additional infill drilling. By adding ficulty can now be partially addressed through a com
all the annual after-tax net cash flows for the life of the bination of "segmented" cash-flow models and
field, we derive the field's cumulative after-tax net decision-tree analysis.
cash flow (CNCF). Bu t the most fundamen tal difficul ty with OCF
Annual net cash flows are discounted llsing the analysis of exploration ventures is that the observed
appropriate annual present-value factor for the business behavior of most major companies indicates
selected discount rate (Table 10), By adding them for that they place much more intrinsic value on large
reserve, long-term prospects than is consistent with
their OCF valuation of such projects (Boccia, 1996). The
6To match the appropriate area, a\'eragc net pay, and HC-reco\'ery reason, of course, has to do mostly with discounting:
factor to the 1'90"10, PSoou, and P10"u resen'es cases, remember that using prevalent discount rates of 10% to 12%, PVs of ,
'~
multiplication of the product of the three P23~., case" for area, pay, annual production cash flows beyond the first 15 or 20
('. , and He-recovery results in the PlO"" reserves case, jllst as the prod
Li'L( of th"U1ree P77% cases for area. pay, and HC-recc",ery ~'ields
years of a project's life are reduced to practically-zero. --
the P90% reserves case. The product "i the three P50"" \'alm's yields But most companies specifically seek large new fields
the PSO'r" reserves (Appendix B). having long-term, stable production potential-just the
Chapter 4 51
~--
Present Value of 1
1
{1 + i)n
-
' .. _.-- .• _ _~-_ .. _--_ .. _-----------~. __ .. __ ._-----.._
l;
52 Economic Analysis of Exploration Ventures
kinds of fields that build companies and provide four parameters: time, price differential between cur
steadyt reliable t low-cost production revenue streams rent price and strike pricet interest rates, and variance
for many years. Yet these are the kinds of long-term t of price. But there are four important difference;
steady cash flows whose value beyond about 20 years is between options in the stock market or commodities
shown to be nearly zero by conventional DCF analysis! market and options in the oil business.
One solution to this difficulty has been to employ
artificially reduced discount rates when evaluating 1. When a stock option is exercised the benefit is
such opportunities. Of courset the choice of the proper realized immediatelyt but when an oil company
low discount rate is arbitrary. A better approach would exercises its option by developing a property or
be to run all DCF analyses for all candidate projects in installing new EOR procedures in an exi'sting
the inventory at a low discount rate that represenb the field t additional revenues may not be realized for
fundamental interest for the uhire of the money"-per one to four years. This can be handled by dis
haps 4'X, to 5%-one that does not include any provi counting those future cash flows back to the time
sion for infla tion of costs or prices. By using the cu rren t of the decision, but price fluctuations m,ly intro
mean of "real" historical oil pricest t]-tp rompany could duce substantial uncertainty.
compare all projects for its portfoli0 and rank them on 2. In the case of stock optionst economic benefits are
that basis. This would be internally consistent, reduc realized because the current price exceeds the
ing the negative effect that discounting has on long strike price. Howevert most large oil fields have
term cash flows and eliminating the possibility that lives of 20 to 100 years, and the price of oil or gas
some projects could be made to appear more attractive behaves as a typical fluctua ting commod ity,
through biased selection of elevated price-escalation oscillating widely through a max-min price enve
schedules. Use of the historical "mean of real oil prices lope. Therefore, any price differential at the time
(corrected to present day for inflation) recognizes the of option-exercise will change many times, posi
long life of most large oil and gas fields and tends to tively and neg'atively, during the field's life. Thus
correct for short-term price influences. Such a simpli the benefit in oil and gas field options usually
fied procedure would likely optimize the ranking of relates not to elevated oil prices but to improved
projects and would select for growth. operatil/g profits t which result from reduced costs
But the most promising solution to "the DCF prob to find, develop, and/or operate fields. In general
lem" lies in adoption of an alternate method of "aluing then, oil exploration option behavior depends on
exploration ventures: Option-pricing Theory. waiting for new data or technology that mav
lower costs and/ or reduce riskt not on sustaine~i
higher oil or natural gas prices.
Option-pricing Theory and 3. Calculating the variance (or standard deviation)
Valuation of Exploration Ventures in price of a given stock is easy. Moreover, it
would be easy to calculate the historical standard
The sequence of characteristic business decisions deviation in the real price of crude oil (approxi
that attend discovery, development, and operation of mate 2000 estimate = ±$7.00 bbl). What is not
large oil and gas fields represents classic option behav known is the variance ~n prices of oil properties,
ior (Figure 24). Companies acquire leases or contract such as exploration prospects or disco\'eries that
areas, then invest in addition,i1 geotechnical dil ta to have not yet been developed.
refine their risk/reward perception-which; if elKour 4. There is a large and continuous market for trad
aging, leads them to exploratory drilling. If drilling is ing common stocks-any time an owner wishes
successful t the company confirms ilnd delineates the to sell his or her common stock, a trade may be
discovered field, which, if it is judged to be economic, eHected almost instantly. But sales of oil proper
the company then develops and produces. Enhilnced ties commonly require months of preparation,
recovery projects milY be Clrlded during the field's analysis, and negotiation, during which ongoing
producing life t followed by ilbandonment \\'hen pro developments may cause major changes in the
duction revenues decline below operating costs. perceived value of the property (Lohrenz, 1988).
Depending on the previliling economics, technologic
developments, political trends, and contrilct terms, the Most major international oil companies are now devel
company may choose at \',nious decision points to oping some form of option-pricing procedure to
invest further, to defer ilction, to sell part or ,111 of their replace conventional DCF analysis for establishing the
interest, or to abandon the "enture. Each stage in the monetary value of oil and gas ventures, especially
venture thus represents an option. long-term, large-reserve project~J£=h.orn, 1999; Dixit
The Black-Scholes model (Brealey and \ [yers, 1988; and Pindyck, 1994; Lehman, 1989; Marl~"' -(::( ,~... 'i 99~_-;
Bernstein, 1996) for valuing il stock option depends on Paddock et al., 1983; Pickles and Smith, 1993).
Chapter 4 53
I
J Field Pro'eel
Figure 24 Option concept applied to E&P projects. FlO = farm-out; EOR = enhanced oil recovery.
-4
'~':j
moreover, it selects agaillst long-term, large-reserve In its simple form, ENPV makes no provision for risk
ventures. (as distinguished from dlnllee). However, ENPV does
form the basis of an expanded economic measure that
Maximum Negative Cash Flow considers project utility and corporate loss-aversion,
Maximum negative cash flow (MNCF) is the "turn called risk-adjusted value (RA V) (Cozzolino, 1977,
around" point on the project's cumulative discounted 1978), which is discussed further below and on page 92,
net cash flow profile--that point at which the cumula
tive net cash flows have reached their maximum neg Investment Efficiency
,,
ative level and will turn back upward toward the
break-even line and increasing profitability (Figure
25). MNCF is a discounted value, and for exploratory
Investment efficiency (IE) is the most discrimiJ~ating
form of profit/investment ratio, in which profit is the
project's NPV and the investment is maximum negative
,, . ,
ventures, where a new field may be developed if there net cash flow (MCNF) (Equation 4).
is a discovery, MNCF represents the net of invest
ments and partially offsetting production revenues. It IE = PRESENT VALUE (4)
is useful in capital budgeting and planning because it INV~STMENT (MNCF)
represents the amount of money that actually will be
needed frol11 the corpora te treasu ry. In firms tha t Figure 25 shows the relationship of the key terms
have short-term capital constraints, it may be useful involved in investment efficiency. Investment effi
for comparing and choosing projects. It also sen'es as ciency prod uces the same project ranking as does
a "time focus" for concerns about political risk another preferred measure: growth rate of return
because it represents the time of greatest economic (GRR) (Capen et al., 1976), in which net cash flows
vulnerability for project life. But the most significant are considered to be reinvested as received, at the
use of this parameter is as a proxy for investment in comptlny's real ra~e of return, and projected out to
investment efficiency, one of the most useful common some preselected target year (often 10 or 12 years). For
economic measures (see this page). longer-term projects, annual net cash flows extending
beyond the target year are discounted back and then
Expected Net Present Value combined with accumulated earnings. Even though
As introduced in Equations 1 and 2, ENPV is the generally acknowledged to be a superior economic
chance-weighted NPV of a venture, in which the prod measure, GRR has not received the wide usage of its
uct of the chance of failure and the cost of exploratory derivath'e, investment efficiency (Clapp, 1995).
failure is subtracted from the product of the chance of Another advantage of investment efficiency is that
success and the NPV of the mean-resen'es case (Equa it can easily be risked-that is, it can be modified to
tion 3). For prospects (rather than plays), it integrates chance-weight all outcomes (Equation 5):
prospect resen'es, DCF \'alues, and commercial
chance of success, thus allowing comparison of high Pc (NPV) - Pf (DHC) (5)
risk, high-potential prospects with low-risk prospects
Pc (I) + Pf (DHC)
having only a moderate resen'es potential. It is useful
in portfolio analysis primarily because the sum of all where
project expected values is the expected \'alue of the
entire ?ortfoliu. The c!r?whack of expected \'alue is
that it implies that the firm is risk-neutral: t\\'O projects I = maximum negative cumulative discounted
might have identical ENPVs even though one project net cash flow, and
requires a much larger initial capital il1\"eshnent (= DHC = net cost of exploratory failure
dry-hole cost) than the other (see p. 92). In fact, how (literally, dry-hole cost).
11 $11~~
10 f - - - - - - - - - - - - - - - - - - - . = = t : ; ; ; I : : : : t - i 10
9 -a- UNDISCOUNTED 9
8 8
7 7
6 6 )
5 5
4 Cumulative Net 4
3 Present Value 3
2 2
o1 -<>- DISCOUNTEDAT 4% 1
0
-1 -1
-2 -2
-3 -3
-4 -4
-51---------,M-----------------j -5
-6 L-----l'-----l---l.~____L___L~~~____.L_____L_____'____'__'__'___'_ $-6MM
2 3 4 5 6 7 8 9 10 11 12 13 14 15
YEAR
Figure 2S Cumulative net present value and maximum negative-net cash flow.
where
RT == risk tolerance
If the company's risk-quotient (r) is known,? every PY == prospect present value
venture can be sho\\'n to have an optimum working Cost == cost of failure
Pc == chance of commercial success
Pf == chance of failure
7Co7.zolino suggested thnt n rough npproximntion of (r) \\"as
$50MM nnnuni budget would be 1/50 or 0.02. bter observntions If a company wishes to include considerations of util
57
58 Exploration Plays-Risk Analysis and Economic Assessment
in substantial part at modern depositional processes, field development became more efficient because the
constituted a revolution in sedimentary geology, espe fields in a play usually had similar reservoir condi
cially with regard to industry knowledge about differ tions, thus similar development and production tech
ent types of reservoir bodies. niques could be applied to the related fields in the
In the late 1960s another new subdiscipline-sedi play. Learning facilitated profitability.
mentary basin analysis (MiaH, 198-!)-arose in But exploration economics tended to focus on the
petroleum exploration, abetted in part by the emerging single prospect as the ecol/omic /II/it ofexploratioll, even
revelations of plate tectonics (Dickinson, 1974; Klemme, though companies didn't ordinarily explore on th~
1975,1980; Bally and Snelson, 1980) and in part by com basis of isolated, unrelated prospects. Instead:they ;
puter-processed COP seismic data, ,'vhich could now directed their attention to the playas the operatiollal
;
sense and depict the subsurface configuration and Ill/it of exploratioll. And the unifying attribute of the
internal geometry of specific sedimentary sequences play was the overall similarity of traps-reservoir
(Vail et aI., 1977a, 1977b). Now geologists could relate rocks, structural and stratigraphic configurations, and
the distribution of reservoir and sealing rock bodies to sealing facies. The U.s. Geological Survey and U.s.
geologic structure, in terms of (1) origin, provenance, ~:ii-,erals Management Services employed the play
and regional setting, as well as (2) subsequent struc concept in carrying out several assessmelits of
tural evolution and present configuration. remaining onshore and offshore U.s. oil and gas
These geotechnical concepts and tools all contributed resources during the 1990s. Their databases are now
to the development of the explomliol/ plll~( cOl/cepl, first available to the public, at no charge, on CD-ROM
used by the Geological Survey of Canada in 1972 (1995, 1997, 2000).
(McCrossan, 1973; Roy, 1975; Miller, 1986) and later
well described by Baker et al. (1986): a grol/P ofprospects
as well as oil Illld/or gas fields, all Ilt7vil/g sill/ilar geologic ori Plays and Petroleum Systems
gil/s-n fllll/ily ofgeologimlly similar Imps. Fields compris In the early 1970s ongoing research in petroleum
ing a play contain similar reservoir rocks that arose geochemistry began to bear fruit, leading to the wide
from similar depOSitional processes, and the constituent spread recognition that oil and natural gas could be
field reservoirs exhibit similar production patterns. classified into distinct geochemical types, and that
Prospects and fields in a play have similar structural they were generated from kerogen-rich, generally fine
configurations and structural histories. They hm'e simi grained sediments in petroleum-generative depres
lar top seals and seat seals. Also, fields in a play form a sions (Demaison, 1984) through processes tha t were
coherent lognormal distribution of ultimately reco,'er time- and temp~rature-dependent(Waples, 1980).
able reserves. Other names for the same thing include Later geochemical work recognized that kinetics
genetic trend, geologic fairway, petroleum zone, com played a dominant role in the generation process
plex of fields, and producing trend. (Hunt et aI., 1991; Waples, 1994). Nevertheless, inte
Exxon geologists published important papers on var grated geologic methods of analysis and mapping
ious aspects of play anal ysis during the 1975-1995 could identify such thermally mature basins, or
period, especially David White (1980, 1988, 1992,1993) "kitchens," as well as the probable timing and direc
and R.A. Baker (1986, 1988). Sluijk and :\"ederlof (1984) tions of subsequent migration of generated petroleum.
described Shell's complex system for prospect and play With these geochemical capabilities, explorationists
analysis. Serial publications by the Texas Bu~eau of Eco could now demonstrate w'hat had long been sus
nomic Geology staff during 1985-1993 described and pected-that fields in a play also had similar histories
summarized many plays in Texas. Otis and Schneider of hydrocarbon type, origin, emplacement, and preser
mann (1997) published an excellent and detailed sum ,·ation. Armed with these powerful geochemical con
mary of Chevron's exploration risking procedure. cepts and analytic tools, explorationists began to
Perceptive prospectors recognized that when com exploit the play concept worldwide.
panies carried out exploration campaigns along struc By abou t 1975, regional explorationists began to
tural or stra tigraphic fairways, they were usually develop yet another unifying concept, employing the
engaged in play exploration. Geophysicists could lay concepts and techniques described previously: the
out regional seismic grids and stratigraphers could concept of the petroleum system (Magoon and Dow,
deal with depositional trends. This was a much more 1994). Here the focus was on the e!ltire petroleum-gen
efficient way to conduct exploration. Once the key erative basin complex of petroleum source rocks, car
geologic and geotechnical patterns and relationships rier beds and conduits for migrating petroleum, and
became understood-once a company had "cracked traps containing the reservoired hydrocarbons. A
the exploration code" and discovered the first field
discovery of other fields in the play ,,',15 greatly facili
--
pet~·<.;leJlm .system often contains multiple plavs with
..,. .
several different types of traps, all charged from a
tated. Thus, play exploration was a form of geo common petroleum source rock in a petroleum-gener
technical leveraging, ilnd it often led to multiple dis ative depression, or kitchen (Figures 26 and 27)
cO\'eries ilnd highly profitilble "core are,l~." Moreo\'er, (Demaison, 1984). Magoon's emphasis on the "critical
Chapter 5 59
\*
STRATIGRAPHIC
EXTENT OF .
PETROLEUM
.
... • SYSTEM
. I
\," ...
.. ' _ ' "
I
I'~ ~. ·1
I Fold·and·thrust belt: arrows
indicate direction of relative motion
Basement rock
Top oil window
* location used for burial history chart t) t) t) Top gas window
Critical moment: 250 Ma
Figure 26 Hypothetical geologic cross section of a basin showing the essential elements of a petroleum system,
From Magoon and Dow (1994).
Figure 27 Map of the hypothetical basin showing. areal extent of the petroleum system. From Magoon and Dow
(1994). Il
60 Exploration Plays-Risk Analysis and Economic Assessment
I
moment" (Figure 28) facilitates critical thinking and 400 300 200 100
investigations bearing on time of peak hydrocarbon
generation and probable migration routes to traps
available at that time. Accordingly, the petroleum sys
tem approach provides a very powerful framework fOf
understanding the basin's "plumbing system," and it /--_==""'...--_~ +-:R~eservoir Rock
stimulates and encourages the development of new Seal Rock
exploration ideas and targets (Table 11). However, it Overburden Rock
does not lend itself to economic analysis because of the /
Trap Fonnatioi'l
complexities attendant upon the multiple and dissimi
lar constituent plays (Figure 29).
Possibly because of their geoscience research focus,
Generation-Migration- i
Accumulation
Preservation nme
!
I'
leading petroleum system authorities define the geo
logic play differently than do most practical petroleum
Figure 28 Events chart showing relationship
cxplorationists. Magoon and Dow (1994), for example,
between essential elements and processes as weli as
conceive of the petroleum system as including only
preservation time and critical moment. From Magoon
those elements that are known-proven source rocks, and Dow (1994).
identified carrier beds and conduits, and discovered
fields whose analyzed oils and natural gases are clearly
related to the source rocks. Following that definition, a
play is only that part of the petroleum system that is any way hinder the employment of the petroleum sys
undiscovered (Magoon, personal communication, 1997). tem as an extremely useful exploration concept. The
To the practical explorationist, such a definition writer strongly urges active explorers to follow the
causes substantial problems: First, it allows se\'eral dif· above-recommended definition of the exploration
ferent families of traps, if charged from a common play.
mature source rock, to be included within a single play.
This presents extremely complex consequences for
companies that wish to analyze an entire playas a full Play Selection as the Critical
cycle business venture because of very wide variances Exploration Decision
caused by the diversity of trap types, geologic depen With the previous considerations in mind, and rec
dencies, and profitabilities. Second, it makes use of ognizing the prevalence in international exploration
field-size distributions (FSDs)-a very powerful pre of large contract areas obtained through various com
dictive tool in economic play analysis-much less dis binations of costly work commitments and/ or front
criminating because different genetic trap types must end bonuses, it is clear that the most important
be included in the same distribution. This is a serious decision in international exploration is not which
statistical drawback. Third, it artificially separates prospect to drill. Instead, the key decision concerns
prospects and leads from discovered producing fields which new play to enter because it involves much
in the play and thus hinders our use of field analogs larger commitments of money, time, and personnel.
and models in exploration. Finally, the preferred defin The economic consequences of choosing a bad play
ition of the play, nf(jllliIy ojgl'ologically rt!ll1tcdfidd~ . can be ''''rious tn a medium-sized international
prospects aJld lends, all of similar geological' origiJl Illld explorer; to a smaller firm such an outco~1e can be
charged frolll COI/l/l101I petrolel/lIl SOl/rce beds, does not in financially disastrous.
Table 11 Factor comparison in the four levels of petroleum investigation (Magoon & Dow, 1994).
...... • Development wells prove up the discovered field • Successive dis~overjes prove up the successful play
must construct detailed cash-flow models based on time value of money. But most successful plays involve
parameters estimated through objective and reliable several sequentially discovered and developed fields.
geotechnology and contract analysis to assess project Moreover, early production revenues may themselves
profitability on a full-cycle economic basis, first for the generate funds required for subsequent development
anticipated new fields (prospects) and then for the antic of fields within the play or core area. Construction of
ipated new core producing area (the play). pipelines and/or development of markets for produc
However, there are also some important differences tion may take several years. Accordingly, in some
between prospect and play risk analysis. For single plays there may be legitimate reasons for staging the
objective wildcats, the geologic chance factors used to exploration and development, consiste!1t with rflaxi /
estimate the chance of encountering flowable hydro mum long-term profitability. Often Stich delays may
carbons (= geologic success, Pg) are treated as if they constitute pragmatic, nonquarltitative applications of ti ·
were independent of each other. This assumption option-pricing theory (see p. 52).
allows serial multiplication of the individual compo
nents of geologic chance to yield Pg. However, for Integration of Geotechnology,
exploratory wells having ~uItipJe objec~ives, risk
analysis recognizes that some chance factors are COI/I
Economics, and Management
ilion to all the objective zones, whereas other chance
in Play Analysis
elements t'ary among the different objective zones (p. There are many ways to convert a viable play into
41 and Appendix D). The chance factors that are shared either an economic loss or an unappreciated, rejected
(that is, common to all the objective zones) frequently investment opportunity. Whether the error has to do
include petroleum generation and migration, includ with bad geotechnology, bad engineering, bad eco
ing the elements of timing and presen'ation. The inde nomic evaluation, or bad acquisition strategies, the
pelldwt chance factors typically include reservoir, consequences are the same. That's why successful risk
closure, and containment. To correctly estimate the analysis of exploration plays requires thorough syn
chance of various combinations of discoveries for such thesis of multidisciplinary topics: geostatistics, geol
multiple-objective prospects, it is necessary to calcu ogy, geophysics, reservoir engineering, drilling and
late the chance of success of the shared factors sepa completion technology, economics, decision analysis,
rately from the indepelldCllt chance factors. contract analysis, political analysis, industrial opera
This same principle is also consistently employed in tions, and business strategies. These specialties take on
risk analysis of plays, simply because there are, char greater or lesser importance a t different stages in the
acteristically, elements of geologic chance that apply life of the successful play, but their successful integra
equally to every prospect in the play (:::: "shared tion is required if the play is to succeed as a full-cycle
chance"); similarly there are also elements of geologic economic risk venture.
chance that vary among the different prospects (:::: Accordingly, the most common organizational pat
"average prospect chance" or "local chance"). To tern adopted by the modern petroleum industry for
make things even more complicated, we may also deal analyZing and exploring new plays is the multidisci
with partial dependencies, where a gi\'en geologic plinary exploration team. The composition of such
chance factor is partly shared and partly independent. teams may change in both numbers and specialties as
In carrying out risk analysis of plays, especially new the play evolves.
plays, the elements of geologic chance associated with
hydrocarbon generation, migration, and timing are
usually of paramount importance to address and sat Important Geologic Concepts
isfy. However, for plays hadng greater exploration for Play Analysis
maturity or plays in known petroliferous basins, con
cerns about hydrocarbon charge comrnonly are dimin AppendiX E outlines topics, procedures, and key
ished, whereas concerns about reservoir, closure, and steps in carrying out a geologic synthesis of an explo
containment are greater. Ne\'ertheless, all chance fac ration play. Such a synthesis-Htotal geology"
tors are theoretically of equal importance in the sense integrates all geologic elements bearing on the likeli
that all must have been satisfied in order for a reser hood of petroleum accumulations, their possible vol
voired accumulation to exist. As pointed out previ umes, spatial distribution, and geologic character.
ollsly (p. 34), the chance factors should be thought of Such integratea studies form the basis for any subse
as if they \'\'ere equal links in a circular chain: if one quent legitimate attempt at risk analysis and economic
link is broken, the chain is broken. assessment of a new exploration play, or entry into an
Another significant difference between plays and ongoing one.
prospects concerns the strategies inYoh'ed nfFr.;;= disL;'~~ o. It is far beyond the scope of this book to delve deeply
ery: generally, it is important to develop the newly suc into the subject of geologic basin analysis or even to
cessful prospect promptly and begin generating summarize such a complex topic. AppendiX E is, how
production revenues as soon as possible becCluse of the ever, a useful flow sheet to guide explorationists
Chapter 5 63
Established Correlation
---:--r:~""-~---------Y:~=r-- - 4
Marine Shale .- -
Shelf Marine
Limestone Shale
Figure 30a Traditional facies interpretation indicates that the shelf limestone and marine shale are coeval.
Established Correlation .
_________________r.:--=_c-"-=_..----
Marine Shale
Shelf
Limestone -~~~~~iit~1;=} ---- -
~.
Marine
Shale
-
-----"...... - - -
" " ,,':.-.--
,"' ~-----_ . .::=-<
-
' .... ~'..::: ...:~-- - -:
~-~-::::-~-""'- ~
Establishe -~:;~~:.~
- -
Marine Limestone
--
d Correlation - -
Figure 30b Depositional topography interpretation indicates that all of the shelf limestone is older than any of the
marine shale. The only facies relationship is between the shelf limestone and marine limestone, with basin starvation
inferred.
through the process/so that they will end up with the concerned with efficient evaluation of basins and plays,
ke~ maps, cross sections, and data necessary for the sequence stratigraphy is important for three reasOns.
estimates and calculations required in the sequential
ste\?s of formal play analysis. In addition, several geo Regional Mapping
IOglC concepts are especially important to consider in One of the fir~t tasks facing the geoscientist who is
cond ucting a sound play analysis. entering a ne\~' basin or play is to identify natural
cycles-unconformity-bounded stratigraphic units
Stratigraphic Sequences
and to make maps and cross sections using such natural
in Play Analysis
sedimentary packages and especially the unconformable
surfaces bounding them. The true spatial and geologic
The topic of sequence stratigraphy has been a relationships beh~'een source rocks, reservoir rocks, and
rapidly expanding field of geologic study since the sealing rocks are best understood in context with these
mid-1970s. Utilizing the principles of sequence stratig natural chapters of earth history. In particular, sedimen
r~phy is .extremely important in carrying out geotech tation related to contemporaneous structural events and
mcal reVlews and syntheses of new basins and plays. sedimentary provenance is most readily deciphered at
L.L. Sloss (1963) published the first comprehensive a regional scale when geoscientists think in terms of
treatment of continental-scale stratigraphic sequences, stratigraphic sequences and subsequences.
identifying six unconformity-bounded "packages" of
sedimentary rocks on the North American continent. Depositional Topography
Sloss/s original sequences each represented about 50 to Traditional stratigraphy has not dealt comfortably
120 million years of geologic history. Later work by with depositional topography in sedimentary
SI~s/s student, Peter Vail, and his Exxon associates led sequences. Figure 30a shO\·vs how the characteristic
to the recognition of a hierarchy of ordered sequences changes in rock type and thickness are conventionally
cycl.es spanning dimir~ishing intervals of geologic time interpreted as facies changes of coeval sedimentary
(Vall et al., 1977a, 1977b). Iior petroleum geoscientists deposits, whereas Figure 30b deals more realistically
64 Exploration Plays-Risk Analysis and Economic Assessment
It
with the same rock patterns, as a combination of dif depressions" and recorded four important characteris
ferent facies and depositional topography. Deposi tics of such generative basins:
4
allows the geoscientist to understand the true spatial exploratory success ratios;
relationships of various constituents of the petroleum 2. They are mappable by integrated geologic/
and seals. These are most readily detected where the 3. Large petroleum accumulations tend to occur
cross section or mapping datum is a regional uncon near the center of such petroleum-generfitive /
formity, either at the base or the top of the subject basins, or on adjacent structural high trends; and
sequence or subsequence. 4. Migration distances commonly range in tens
Predictions Based on Sea-level Stands rather than hundreds of miles and are limited by
the geologic past, and-in context with depositional tration of these basic ideas (Figure 31). Murris (1984)
topography-has produced characteristic depositional and Sluijk and Parker (1986) demonstrated that synthe
models and lithologic patterns. Stratigraphers may sis of these integrated geochemical/geologic methods
look for and predict certain combinations of petroleum (in combination with trap-size considerations) results
facies associated with highstand, transgressive, and in greatly improved exploration results (Figure 32).
lowstand deposits. In particular, just the awareness of Later refinements of Lopatin's and Waples's ideas on
the possibility of lowstand events can be an insightful time-temperature index as an indicator of petroleum
guide in the search for new plays, and modern seismic generation, through realizatiOlI of the strong kinetic
surveys are a powerful tool for detecting and delineat influence on the process (Hunt et aI., 1991; Waples,
ing key reservoir and sealing units. 1994) and sophisticated probabilistic computer model
Petroleum Source Rocks the time and relative quantity of petroleum generation,
Arthur and Schlanger (1979) recognized the exis as well as the probable direction of migration out of the
tence of "global anoxic events," of Cretaceous age kitchen toward the basin margins.
(Albian and Turonian), and speculated that they rep Explorationis'ts engaged in play analysis must
resented Widespread organic-rich marine deposition. understand thoroughly the genetic relationships
Klemme and Ulmishek (1991) identified six such between the family of related traps we call a play and
events (which occurred in the middle Silurian, Late the kitchen from which the traps were charged with oil
Devonian, Late Pennsylvanian-Early Permian, Late and gas. Fortunately, the petroleum system concept
Jurassic, middle Cretaceous, and the Oligocene facilitates that analysis and understanding.
high priority should be given to those basins and plays nical predictions for purposes of exploration in oil and
containing marine sedimentary rocks of <lny of those gas plays, geoscientists must integrate all aspects of
At the sam'e time, the geoscientist must not ignore rocks), structural geology (burial history, tectonic
petroliferous source-facies that are more prO\'incial or events, structural styles), geochemistry (source rock
even local in origin and extent, such as foundered richness and volume, generation amount and time [the
structural basins, euxinic back-reef depressions, and "critical moment"], migration routes and impedance),
lacustrine deposits. However, such secondary source geophysics (trap detection and measurement), and
rock units may often b~ understood only in context petroleulII occurrence (fields, production data, shows).
\vith sequence stratigraphy, basin tectonics, and geo We call such synthesis "total geology." The petroleum
/
Deplh conlour> 01
souret rock Itvtl
(m. subseo)/
Droinsheds ot
this levtl
oreo (VR/E)
(a)
0
I ,
IOkm
0
A PROSPfCf A' ~-j-o +
-=:.,::.=:.:.:....-...l..---=---_ _- l - CENOZOIC
I depth
----- -------~ KU
I
---- Kl
--
..,.,.,. / ' Source rock. Il!vet
4000 ..
financiill, and operational considerations. Moreover, and uncertainties t,hat bear on a play's definition,
the procedure should be carried out in consistent prospectivity, uncertainties, and risks. Each company
ways throughout the organization, for purposes of should adopt a consistent geologic style by which
com~leteness and comparison with other plays being essential information may be synthesized objectively
conSIdered. Accordingly, many firms have devel for n1anagement's consideration.
oped play-analysis checklists to guide geoscientists
in the investigation, analysis, and representation of
such plays and trends. Appendix E is an example of Use of Analogs in Play Analysis
such a checklist. When such a checklist is properly Exploration play analysis requires geoscientists to
used, the result is that all available significant geot make informed geologic and geostatistical predictions
echnical information has been gathered for use in under conditions of great uncertainty. Frequently,
the risk analysis of the playas a full-cycle business direct evidence bearing on key geologic criteria is sim
venture. Such a play risk analysis, however, usually ply not available. However, knowledgeable geoscien
is carried out as a sepetrate flow sheet or software tists may make surprisingly accurate indirect forecasts
program . by using appropriate analogs.
. It is also desirable for geoscientists to develop con An example of the effective use of analog FSDs,
sI.stent, expressive symbologies and mapping tech adjusted appropriately for geologic differences and
I1Iques that contribute to such synthesis, represented exploratory maturity, is presei1L~' lat<.:.-:-in a west
as play maps (White, 1988), which convey to explo Texas case where the two subject areas were indeed
ratIOn decisionmakers the key geologic relationships geologically analogous.
Ii
66 Exploration Plays-Risk Analysis and Economic Assessment
6000,.----------------------------,
I
7200 ,
THEORETICALLY BEST
POSSIBLE
FIRST 4~ DISCOVERIES IN
DECREASING SIZE, TIIEN 120 DRY
DRILLING SEQUENCE:
HOLES
:£:
i= -.
u
l/)
<t
;:=:._. -
)---_.
4800 1U
0: )------
0 --{
...J u..
_.
r--._._. =---'--=::===I
[[)
[[)
0~
'"2 0
Q
?::
n.
I
F-'
<5 'WOO
I
(fJ
~. . . . . .- {
>
a:
" F- -------7
w
> ,--......=-_-_---- .._-====1
---.
o
u
(fJ
f-
i3 3200
w
>
i= PROSPECT RANKING
<t
r:====;:=:~'--~'~~§~
...J
::;) BY TRAP SIZE ONLY:
:::E
::;) ~-'':::_..----f:----::=:::::I FOREC.t.STlNG EFFICIENCY 28 %
u
-.----.~ j------(
I
I WOULD LEAD TO
DISCOVERY SEQUENCE
SCATTERED AROUND
+-----
\
t===~-~"l-' DIAGONAL LINE:
1600 .--1 I FORECMlTIN$ EFFICIENCY 0%
------·W
----- '1
THEORETICALLY
WORST POSSIBLE
-II DRILLING SEQUENCE:
800- ~ f20 DRY 1I0LES FIRST
TIIEN DISCOVERIES IN
INCREASING SIZE
0-fJ--------,r--------.--------,~----__r-----.____--J
Figure 32 Integrated methods produce more efficient exploration results. From Sluijk and Parker (7986).
Chapter 5 67
In a broader context, explorationists should always 3. The relath'e efficiency with which industry can
be alert for documented analog examples of the basins, sense and discover larger, richer fields eafly and
plays, and fields they are studying in the course of lea\~e sm'aUet, 1:?'~.5 profitable fields to be fOlmd in
evaluating new plays. The international geologic liter the later stages of the exploration cycle.
ature is rich in classifications and documented studies
of basin types, classes of plays and trends, types of dif Discovery Process Modeling
ferent fields, FSDs, and other geostatistical data. Item 3 above touches directly on the concept of dis
To maximize the effective use of analogs, three con covery process modeling, first expressed in 1958 by
ditions must be met: Arps and Roberts: Ifre likeliJlOod of discoverillg afield of "I
givell size class is proportiollal to tile gcollletric probabilit~
1. Geoscientists must have ready access to the inter of its discot'ery alld to tile efficicllcy of exploratioll (Equa
national geologic literature. A qualified geotech tion 8).
nicallibrarian, with keen organizational and
computer-search skills, greatly facilitates such F;\ (11') == FA (oo)(I_e-CAidB) (8)
capabilities;
2. Statistical databases, organized to be geologically where
discriminating, allow geologic class and type to
be related to pertinent geostatistics; and FA(w) == cumulati\'e number of discoveries estimated
3. Geoscientists must adopt a working philosophy to be made in size class A by the drilling of w wells
that recognizes the likelihood of substantial FA(oo) == ultimate number of fields in
uncertainty and embraces the use of analogs to size class A that occur in the basin
help make estimates where direct evidence is B == area of basin
absent. A == average areal extent of the fields
in the given size class
w == cumulative nlimber of wildcat wells, and
Key Concepts and Techniques for C == efficiency of exploration
Risk Analysis of Exploration Plays It is important to understand that C is a dimension
less number estimated for each size class, and that C
Field-size Distributions may vary, being higher for the larger size classes and
The topic of FSDs was introduced in Chapter 2, rel declif\ing gradually as size class diminishes unless new
ative to prospect risk analysis, where the FSD was concepts or technology cause temporary increases in
used as a reality check (p. 12). In play analysis, knowl exploration efficiency.
edgeable manipulation and titili'zation of FSDs, in When.C = I, discovery is random. For 1.5 to 3.0
combination with estimates of total fields to be found MMBOE fields in the Dem'er Basin and Texas Gulf
in the play, is an extremely powerful technique that coast, C '" 2.0 (Drew, 1990). The commonly observed
allows surprisingly accurate forecasts of total play pattern is for C to increase (as size class of fields
reserves, and prediction of approximate field sizes increases) ultimately to values approaching 6 or even
pre;;ent in new plays. greater for very large fields. Four very important con
ditions ~ust be emphasized:
Concepts and Principles .
The FSD is the statistical expression of all fields 1. Sampling without replacement is assumed;
found within a basin or play to date. It represents the 2. The area of the play remains constant;
effective natural synthesis of three factors: 3. Exploration technology is either constant or its
improvement is regular; and
1. The natural petroleum endowment of a play, 4, There is a consistent proportion bet\\'een fieid
including ultimate recovery and field productive area
• the abundance and efficiency of hydrocarbon that is, fields ,"'ith large reserves have large sur
generation and migration, and face areas. Generally this is a valid assumption.
• the number and capacity of traps
2. The relative maturity of exploration in the trend, A good way to estimate Cs for each of the significant
including field-size classes of a fl/ture play would be to calculate
• numbers of exploratory wells and discoveries, C for analog field-size classes for several well-explored
and analogous plays or basins, and to apply it to the antici
• optimization of current geotechnology in pated exploration campaign. Obviously, you must
exploration, drilling, and development. have some idea of the natural endowment of oil and
68 Exploration Plays-Risk Analysis and Economic Assessment
...
0
lOOO
here are some documented values for C: '0
(a) 0
10 Oiscovo,y numb.,
10
u
500
~
;; 400
~
u
Discovery process modeling can be useful in play ~
> 300
risk analysis because it provides a method by which ~
An interesting example of efficiency variations in surface geology relates to numbers of prospects, land
international exploration was presented by Shell sat images may indicate the density oJ.,!nomglies. •
(Murris, 1984) and previously introduced as Figure 32. • -:M-
Here, random exploration (C =1) is represented by the Probabilistic Range of Field Number Estimates
straight, sloping line from lower left to upper right. Of course, we cannot know just how many petroleum
Use of only the selective factor of estimated trap size accumulations are present, awaiting discovery, in a play
generated forecasting efficiency that was 28% better fairway. But we can develop plau&ible predictions by
than random. However, use of geochemical parame using analog basins and trends, employing different inde
ters plus trap siz~ produced forecasting efficiency that pendent reality checks, and expressing our estimates as J
was 63% l">etter'than random. Note that this expression probabilistic ranges. For example, we might be quite con- •
of effici~ncy can be calculated at any stage of maturity, fident (P90%) that at least one field is present (given that
and it applies to discovery of all size classes, not indi an acti\'e petroleum system exists in the play area), with a
vidual size classes. Efficiency is expressed graphically best guess (P50%) of four fields, and be 10% confident in
as the percentage of all space between the diagonal as many as 16 fields. Note that this distribution is proba
random line and the optimum efficient case on the left. bly lognormal.
By analogy, exploration efficiency in the Gippsland Such a probabilistic expression can be combined with
Basin was clearly greater relative to trap volumes than the projected. FSD br the play to give a probabilistic dis
to trap areas (Figure 33). tribution of play reserves, a key goal for play analysis.
Shift of Field-size Distributions Field Density in Analog Plays
As previously pointed out (p. 11 and Figure 5), when Careful counts of the numbers of fields in one or
we separate a given play into earlier and later groups of more analog plays usually yields characteristic ranges
field discoveries and plot them on a single log-probabil of field numbers per unit area (i.e., 1.5 to 5.0 fields per
ity graph, we can see the "daughter" FSDs shift steadily 100 square miles [260 km~l or the reciprocal expres
toward smaller field sizes during thelife of the "parent" sion,l field per 20 to 67 square miles [52-174 km 2]).
play. Figures Sa and Sb illustrate this tendency, where Obvioush', the desired value represents the ultimate
more than 100 fields in the Minnelusa Trend (northeast number o'f fields, so the stage of exploration of the ana
ern Wyoming, U.S.A.) were separated into (1) the earli log play must be taken into account. Note I.hat such
est one-third of all discovered fields; (2) the middle ranges lend themselves to probabilistic expression.
one-third; and (3) the latest one-third of all fields found.
Although there is considerable overlap in field sizes Field Area versus Field Reserves
among the three daughter distributions, the overall in Discbvery Process Model
reduction in field size is significant: mean field sizes For analog plays and trends, graphs can be made
drop from 4.58 MM (first third) 'to·2.17 MM (second shOWing the relationship between documented field
third), to 0.76 (last third), for complete, untruncated dis areas and the associated projections of ultimate recov
tributions. For distributions truncated at the commercial eries from those fields. Use those graphs to assign
limit (0.1 MMBOE), the respective differences in mean appropriate areas to the different field-size classes
reserves are larger: 5.51 MM, 2.98 MM, and 1.06 MM. (Table 13). For different field-size classes, use the pos
sible exploration efficiencies as outlined on page 68.
Estimating Field Numbers The final stf'P is to uti!i;:~ the apFopriate area for each
field-size class in the discovery process model (Equa
One of the essential g~0t~chnical tash required to tion 8), which will yield the expected number of
carry out meaningful play analysis is to esti~ate how remaining fields in each size class. The resulting value
many petroleum accumulations remain to be found in should be viewed as an independent reality check of
the play area. Although the development of such esti your other probabilistic estimates of field number that
mates may be rE:'viewed with skepticism by novice explo you derived by analog experience. '
~ationists or non-geoscientists, useful projections can
Indeed be generated using several tested approaches out Pragmatic Approach for
lined as follows. Field Number Estimates
Counting Visible Anomalies A useful rule-of-thumb approach to estimating
field numbers has been suggested by Dr. Jeff Brown
Ii
70 Exploration Plays-Risk Analysis and Economic Assessment
Table 13 System of graduated size classes used FSD for the subject play, whether it represents an
currently by the U.S. Geological Survey (after Drew, 1990). undrilled play in a frontier area or a later-stage entry • ~
into a proven play that is thought to still have eco 'l
Size Size Range nomic potential.
Class (Million ROE'" recoverable oil and gas) The basic principles outlined earlier provide the
basis by which explorationists can analyze and shift
1 0.0 to 0.006 projected FSDs for those plays that are being consid
2 0.006 to 0.012 ered for possible entry and participation by their com
3 0.012 .' to 0.024 panies. Manipulation of FSDs can be classified and
discussed usefully in relation to play maturity.
4 0.02{i to 0.047
5 0.047 to 0.095 Where Production Is Well Established
6 0.095 to 0.19 Companies may consider entering an existing play
7 0.19 to 0.38 where they identify remaining profitable exploration
8 0.38 to 0.76 potential, often anticipating changes in
9 0.76 to 1.52
• regulation, taxation, or oil prices;
10 1.52 to 3.04 • new technology enabling higher success rates, or
11 3.04 to 6.07 increased discrimination regarding field reserve
12 6.07 to 12.14 sizes;
• increased profitability via reduced drilling costs/
13 12.14 to 24.3
more effective stimulation/completion tech
14 24.3 to 48.6 niques/ or mQre efficient production operations;
15 48.6 to 97.2 or
16 97.2 to 194.3 • new geologic concepts or perspectives about the
trend.
17 194.3 to 388.6
18 388.6 to 777.2 In such areas, the existillg play FSD should form the
19 777.2 to 1554.5 basis for the projected FSD for purposes of play analy
20 1554.5 to 3109.0 sis. Where new areas of the play appear prospecti\'e,
the parent FSD itself can be employed. But where
'BOE = barrels of oil equivalent anticipated additional exploration will take place in
the existillg play area, the projected FSD should be
compatible with the most recent 10 to 20 discoveries, '.
trend/ area to consist of productive closures. Given depending on the timeframe involved. It is important
common incomplete trap fill-up, averaging perhaps that analog FSDs reflect the use of similar geotechnical
40% of the closure areas, this suggests that about 1% of techniques and concepts. That is, if new technologies
any productive play should be occupied within pro demonstrate or promise a dear improvement in aver \
ductive field outlines. Utilizing the graphs showing age discovery size, it is permissible for explorationists
the relationship between field reserves and field area to shift the FSD toward 3~ightly larger average field
(p. 68)/ and the expectation of lognormality for the sizes. However, in most cases, projected FSDs in estab
FSD, we can "back out" reasonable estimates for the lished basins should reflect recent discoveries, or shift
field numbers and their associated sizes. TI1is can often toward slightly smaller average fields. In such cases, a
be done by trial-and-error methods as well. reliable and complete catalog of producing fields pro
vides the basis for constructing the projected FSD.
Ouserved Estimating Patterns
Where Limited Production Exists
As pointed out previously, most geoscientists tend
to underestimate the numbers of fields that will ulti In petroliferous trends where previous production
mately be found in a prospective play (gi\'en success) consists of small fields because of poor reservoir perfor
and overestimate the sizes of the those fields. mance/ new tools (that allow areas of higher porosity
and permeability to be detected) may generate disco\"
Use of Field-size Distributions
ery of larger fields for a limited time. Similarly,
improved stimulation techniques and / or more optimal
in Play Analysis
welllocation.s...':'1'\Y allow greater per-well recoveries,
""",,,.:.>J. ... ."k
It is not possible to carry out a geologic and eco resulting in larger fieius. Sum expectations, if prudent
nomic play analysis without generating a projected and documented, should be reflected through manually
Chapter 5 71
shifted FSDs. New tools such as 3-D seismic and/or resulting from a deficient HC-eharge, absence of effec
direct hydrocarbon indicators (OHIs) may allow explor tive seal, fracture leakage, or extremely thin or poor
ers to sense traps that were previously invisible. quality resen'oir rock. Recommended values forsuch ,"'-
Although the population of the remaining trend FSO situations must be credible; they should generally lie
will not change, it is likely that, for a limited period, dis in the range of 0.01 MMBOE to perhaps 2 MMBOE,
coveries will preferentially include the larger remaining depending on geometry and structural/stratigraphic
fields, rather than the smaller ones, as previously dis style. At the upper end, the projected Pl% value
cussed under Creaming Curves. should represent the largest field that could credibly
Another common class in this category applies to be expected in the trend, based on exis;ting geologic /
those trends in which only a fe'''' fields have been dis knowledge. Generally this value will range from a!:,out ,.
covered but have not been developed to the point that 100 MMBOE to perhaps as much as 1 to 4 BBOE, if
reasonable predictions can be made for their approxi such a large discovery is indeed remotely possible.
mate reserves sizes, or the key data bearing on reserves Both the P1°" and P99~" "alues must be credible. A
sizes are proprietary and not available. In such cases, straight line drawn on a cumulath'e log-probability
geoscientists must use whatever data can be obtained graph then represents the projected FSO for the subject
to make range-type estim<ltes of key parameters such trend. This method should be used Oll!!f for new, non
as areas ofclosures (land-sat images, air photos, geologic productive frontier trends where there'is very limited
maps, seismic data, etc.); a7)ernge /let pay thickl1ess (out information, and the projected values must be compat
crop studies, stratigraphic projections, well logs, ible with a"ailable geologic knowledge. Again, this
reported perforation intervals, seismic data); and He method represents a last resort.
recoveries (sources reported above, plus analog fields, As a broad reality check the reader is referred to Fig
test reports, and extrapolations made therefrom). It ure 3-1, which shows the FSO of all oil and equivalent
may be poss:ble to estimate a range of possible field gas fields in the world. It is instructive to note the prob
numbers based on anomalies present on air photos, abilistic distribution of world field sizes: P99% = 0.018
geologic maps, or subsurface seismic maps. It is often MMBOE, P90"o = 0.25 MM, P50'X, = 5.4 MM, P10% =
helpful to identify other producing areas that appear to 143 MM,P1",,=1,917MM, with the mean = -100 MM.
be geologically analogous, utilizing FSOs, field num The writer is indebted to Richard Nehring and Petro
bers, and field densities from such counterpart areas, as Consullnnts, Inc. for these data.
well as more specific analogs bearing on prospect Manipulation of an analog FSO-by shifting it to
areas, reservoir thickness and character, and producing reflect the expert's opinion that the subject play will be
characteristics. The final product of such studies should either more or less petroliferous-may reflect the fol
be a catalog of projected field sizes that follow a lognor lowing expectations:
mal distribution and a probabilistic distribution of pos
sible field numbers. 1. More aud largerfields, because traps are thought to
be larger, reservoirs thicker and/or more perme
Where Production Has Not Been Established
able, and top-seals more numerous and effective;
Three approaches can provide guidance here. The or (ewe'r alld smallerfields, for the opposite reasons;
first uses ranges of geologic parameters from available 2. M~Jre alld largq fields, heca use petroleum source
outcrop or well data, air photos, seismic lines, and the rocks are thought to be thicker, richer, or more
like, as discussed previously. The second method favorablv situated with respect to migration
relies on analog trends and plays in adjacent basins or ·pathway·s and/or timing, or fewer alld smaller
geologically similar basins elsewhere, even on differ !'jdds for the opposite reasons;
ent continents. The key requirement here is that the 3. 'More alld larg('/' fields, because the subject play is
subject basin and the analog basin must be geologi thought to be at an immature stage of exploration
Gilly similar with respect to traps (structural style, compared with the analog play; or fewer MId
reservoir lithology, and origin), scaling rocks, and slllallC'r fields, because the subject play is thought
petroleum endowment. The third approach miJY be to be more mature than the analog play.
iJppJied in desperation-only when time or data limi
tations prevent the construction of a projected FSO for Prediction Using Analog Trends and
iln unexplored, nonproductive new trend as described Interpretive Shifts-An Example FSD
eMlier. In this approach, the geoscientist constructs a The Permian Basin of west Texas and eastern New
generic FSO, represented at the low end (P99%) as a Mexico is a world·c1ass Oil-producing province. In 1945,
small, noneconomic or noncommercial field. Concep its northern extent (Area X) was poorly known, very
tually, this value might represent a small trap that was lightly drilled, and essentially unproductive. Adjoining
not correctly sensed by explorationists or a Jilrge trap Area X on the south, however, was an area of similar
containing only a thin attic-type oil accumulation stratigraphy and structural history and comparable size,
72 Exploration Plays-Risk Analysis and Economic Assessment
•
GLOBAL FIELD FAMILY
P1 P1
·P2 -~ - -- - 1-- r- ~- - I P2.
kiJiP
P5 -I -I
-- - - I - - pY 1-- P5
-P1 o f- '- - I - - f - ~ ~o- -- I
P10
I
- - ilIl
P20-- UI_I
-
..... ~ ...
-l- I '-- I I- P20
·P30- - - - - - -- - -- ' - P30
P4 0-- - - -- I- I - - I ~ - - I- - P40
P50 P50 P50
-
f!!- --II'
P60-- - - I- I - e- - P60
P70- - - - -- - - --
~-
I -I I P70
PB0-- - l- I I- - 1--' l- I-- 1 - -- - PBO
I
-P9o . I-
~
~90 - - I- - - -I ' - - --- P90
P9 5- r ~ r- I - I - 1--1 r- I-- , - 1-- P95
I e.
P9 B-r;;;~ - I- I -- - -- - - - - I - -!- P9B
-P99/ P99
0.01 0.1
1.0 10.0 100.0 1,000.0 10,000.0
GLOBAL FIELD POPULATION (n = about 30.000)*
Area Y, which contained 32 producing fields, nearly all similar for the Area X FSD, the geotechnical team then
with structural-stratigraphic traps in Permian-Pennsyl constructed a predicted FSD for Area X, using P99% =
vanian shelf-sandstone and carbonate reservoir rocks. 1,500 BOE and P1% = 400 MMBOE (Figure 35b).
In order to construct a predicted FSD for Area X, the In 1945 the original Area Y FSD had an arithmetic
FSD for analog Area Y was first adopted as a proxy mean field size of 45 MMBOE, with a median of about
(Figure 35a). Then the proxy was assessed for plausi 1.5 MMBOE (see Table 14, column 1). Swanson's
bility. In this case the geoscientists concluded that mean field size was 22 MMBOE. After downward
Area X was probably not as petroliferous as Area Y adjustment for the aforementioned geologic reasons,
because (1) one of two regional source-rock horizons the predicted FSD for Area X had a median field size
was absent; (2) Area X was less deeply buried and of 0.8 MMBOE and Swanson's mean field size of 8.0
cooler, and therefore less petroleum might have been MMBOE (Table 14, column 2).
generated, and then migrated and emplaced; (3) many Compared with those 1945 predictions, what was
of the Middle Permian porous carbonate formations of the actual outcome in Area X (Table 14, column 3)?
Area Y seemed to have been replaced in Area X by Fifteen fields were actually discovered in Area X dur
evaporites and tight dolomites through facies change; ing the next 12 years (1945-1957). The largest field was
and (4) finally, the geoscientists did not believe Area X Anton-Irish (212 MMBOE); the smallest was West
would prove to contain a fielrl. as large as the largest Petersburg (30,000 80E). The actual FSD for these 15
field in Area Y (Slaughter, 1,000 MMBOE). fields is shown on Figure 35c, for comparison wi th the
Based on these geologic considerations, and work original Area Y proxy as well as the Area X prediction.
_ ing from the Area Y FSD, the geoscientists chose a pro All 15 fields are formed by structural and structural
.•,- ject~ PI%·f:eld size of 400 MMBOE as plausibly the stratigraphic traps in Permian-Pennsylvanian shelf
largest field that Area X was likely to contain. Recog sands and carbonates located along an irregular east
nizing tha t the low-side field size observed for the west regional transcurrent fault trend, the Mata
Area Y FSD (P99% = 1,500 BOE) probably would be dor Arch.
i , :1/1 ! I '1: !
,
!i 1:. i:i11 I
i I
.. ;
I
'i
'I
i
t r! i!. I I ; '!
"
, I,·,1
: II I , I I I :/
Fi~ure 35a
)
, :;.j
";~I i:11111 'I ' I ,
,i ./"II!I :! Iii, I :; I
'"
, : ;;!t
! i!:l
,
: I
i
'Ii
1;1 , II I;1 :!11
; ,,',/ ,1 "il.1
, i ,!
:11 I
I!.;
I!"
.1:1 I I i 11;1
,
;i/ I I I iii! ,
,I
I I I iii
I
:11 'I
:1: : ; : ~ !: '"
Figure 35b
~ . ::jii:l 'I
1--"':·=-4~·:7· ',lJl::-;-;~':-;:;-;! ~:::l'-; !!,j.!- - , - ' C - - - ---------+----'71-----
!..:.i
g: ,,.,-?;i P1 %=1500BOEt: .. J
Ul
..." N 01 N U1 Ul -
'" Ul - '"
Figure 35c
f
I
(2)= 1st precktion, Area X FSD (;3)= lsi-stage actual. Area X FSD
CD ® ® @ ®
Reserves 1945 Area Y 1945 Area X 1945-57 Area X 1945-57 Area Y 1957-95 Area X
Parameters FSD FSD FSD FSD FSD
(BOE) (1st Analog) (1st Prediction) (1st Actual) (2nd Analog) (2nd Actual)
P99% 1,500 BOE 1,500 BOE 4,000 BOE 650 BOE 800 BOE
P90% 30,000 BOE 25,000 BOE 50,000 BOE 8,900 BOE 10,OQO BOE /
-
P50% 1.5MMBOE 0.8MMBOE 1.3MMBOE 0.34MMBOE 0.24MMBOE
Swanson's 22MM 8MM 9.5MM 2.4MM 1.58MM
Mean
Arithmetic 45MM N/A 15.4MM 6.76MM 0.76MM
Mean
P10% 72MM 23MM 30MM 7.1MM 2MM
P1% l,100MM 400MM 400MM 225MM 55MM
How well did this described predictive method sizes: the Area Y FSD shifts to the left, toward smaller
actually perform? How do the predictions compare field sizes, by roughly one order of magnitude, in the
with the actual outcomes? Table 14 facilitates compar second stage of exploration.
ison by showing statistical parameters of the Area Y With such supportive data, the geotechnical profes
analog FSD (Column 1), the Area X Predicted FSD sional staff ofthe exploration team began to have con
(Column 2), and the Actual Area X FSD (Column 3), as fidence that Area Y could indeed be used as an analog
recorded 14 years later. In fact, the median field size for Area X, given the appropriate adjustments for the
was actually 1.3 MMBOE (\'5. 0.8 Mi-.lBOE predicted), geologic differences. To provide an independent test
and Swanson's mean field size for Area X was actually of this approach, we might reasonably expect that the
9.5 MMBOE (vs. 8.0 MMBOE predicted). The arith shift in Area X FSDs from the 1945-1957 discoveries to
metic mean field size for Area X pro\-ed to be 15.4 the next cycle (1957-1995) would be comparable to the
MMBOE. So the method generated reasonably accu observed shift in analog Area Y FSDs between 1945
rate predictions, as documented by Table 14 and illus and 1957. In fact, Figure 36b confirms this expectation:
trated by Figure 35c. the correlation is striking-the downward shift is of
very similar magnitude, and slope changes are also
Second-stage Example quite comparable, providing compelling evidence for
Additional confirmation of the \"aliditv of this the applicability of this method.
method for predicting FSDs in new areas is provided During the period 1957-1995,52 new fields were
by the next stage of exploration in the region. In ana discovered in Area X. The arithmetic mean field size
log Area Y, 292 additional fields \-..,ere discovered dur was 0.76 MMBOE and the median was 0.24 MMBOE.
ing 1945-1957. Arithmetic mean field size was 6.76 Swanson's mean field size was 1.58 MMBOE, based on
MMBOE, and the median field size was 0.3-4 MYmOE. the projected trend line of the FSD data points; based
Swanson's mean field size was 2.4 tvL\IBOE (Table 14, on actual field sizes, Swanson's mean field size was
Column 4). So, as Figure 36a demonstrates, fields dis 0.82 MMBOE (Table 14, Column 5).
covered in Area Y (FSD -l \·s. FSD 1) were notably
smaller in the 12 years after 1945 than previously. Four Additional Examples
Comparison of Figure 36a with columns 1 and 4 of Further evidence of the overall validity of such FSD
Table'14 qL":t:"ltifie..,-i.his substantial redtT(llo;, ;...;,. c;eld __ manipulation is provided by the consistently downward
Chapter 5 75
c
I ;I! r1j =t-! fTfW -,! aughter
I I ! I I!I i I i 1III i r : II A. ;I' IIIU'
(I II 11/1 IJ4il V~ y!~~,'
I j III I I f rtH I I iI ! I IIII Vf U-L A
~". O(1)JII
I I II I I r II rII ! 1 II I I I I iJ{ I V ..rrI I ~nton:lrisli'
I I i I I I I I III : I
" ,• II Iw II i ...... v I I
I i I II I II I~ • ...Y I I I II! I I ! II if
I ! IIII I I~ I..t, r.', I I I 111111 I il , I I 11111
I I III I P I lit' I ! I' I " 1III I I , I I I I III!
I I I I J! I ~ ( I I III Ii I I Ii IIII I I : III I ! I : ! !i
I I I i .,'A
17
v~
I II I' II I~ III J~ !I• I Ii! ~ I I I I iii!'
, 4
1.1 I
I I
.I
.~ ~~! I IIII "I I! ,I ! Ililil
I!: i;/
I I 1111. I, ,.I I Ii'!I! 'Ii
~
....
Ij.,,;.., : I I! I; " I I I!!! i i i I ! ; : II , rI . 4
'1'1
.1
...... f\)
" In...... f\) <.n ...... f\) (]I...... f\) In f\) c........... f\) U1 ......
Figure 36a
! i I; ·111;lil;~~~~~·~~~ughter
, 1
; ; 1)1 II' 1111,'
1 i . Ij! r I "I I I! III : I I ~fff1 Vi I ~ IT ~ PC f Jill
I I I III I I I , IIII I I I , I iI I » .Hl1l Vi ,..rrl _c; I • • .: . , I
~~~~:v~! I!!:.!!!!I!
0> ....... N '"t,JUl ....... f\) In -..L f\) Ul......
! il!!!!:
1\) Ul
! !!!!!;
1\) (Jl......
I
t\)
lt~!:;!
Ul -..L
Figure 36b
-~.II Figure 36 a, b " Use of analog HOs in forecasting actual FSDs: second stage.
-"
Ii
76 Exploration Plays-Risk Analysis and Economic Assessment
: , i :~i
: . I;"'
: . "
~ : i . :i
J
Ii·
:~r I ~ .
; I .
! i .'
I .
O.001MM -
O.01MM
~.
O.1MM 1.0MM
• .....
10MM
~.
100MM
...
1,OOOMMBOE
Figure 37a
Figure 37b
Figure 37 a, b, c, d Downward shifts in F5Ds in area Y between first cycle of exploration (pre-1945: ar,b"c" and
d r) and second cycle (1945-1957: O2, bJ, c2' and d), for four plays.
shifts in FSOs of four plays in Area Y between 1945 this type, there was also an unexpected sharp il/crel1se
and 1957 (Figure 37a, b, c, and d). Changes in the in the size of 51111111 fields in this play. The result is that
median values are substantial, averaging about 85% the mean field size actually increased slightly (0.18
reduction between the two exploration stages. One MMBOE ~ 0.35 MMBOE) in the second exploratory
partial departure should be '11otl:'d:~T....~ona!·). Oolomite cycle. This anomaly may be caused by the small sam
FSOs show a marked steepening from the first to the ple size (n = 5) of the first cycle FSO; in any case, the
second cycle, as expected. However, although there ccollOmic significance was borne out that fewer large
was a sharp decrease in the size of largdields found of fields were found, as predicted.
Chapter 5 77
I;;
;j
I I ::
:
! iI Ii~ ~
~ ;;,! "i ; ; : I
i!----~~-4
1--.-.:...---:..--'.J....
3: ..
! ;:! : ; ;
..----'----'--'-"'--'±..: -----'----Ai----:...
i!
i - - - -...
0.001 MM 0.01 MM 0.1MM 1.0MM 10MM 100MM 1,000MMBOE
Figure 37c
,,
'.! i;i , -~
!: ~ I
.;
~
.. !: ~
.
! i ~
...
'; iii
~ ...
10MM 100MM 1,oOOMMBOE
Figure 37d
Figure 37 a, b, c, d (continued)
. I
Minimum Economic Field Size
exploratory capital investments required to establish
The reader was introduced to the concept of n"1ini the playas a profitable venture-or an economic fail
mum economic field size (MEFS) cnrlier, and its impor ure. Ordinarily, the investments recovered include
~ance in pIny annlysis has been discllssed briefly. Early geologic and g~ophysical project costs (including seis
In the iCc. ),nss.of COl~d1lcting a modern riskanaly'si:-; of a mic), lease or contract-area acquisition costs (such as
pro'po~ed n~w play, it is necessary to estimate 1\1 CFS, bonuses, fees, etc.), costs of exploratory drilling and
\~hlch IS defmed as the threshold amount of producible completion, and overhead, plus required interest. If
. I
oil and gas sufficient to re);over~witl1interest-all one or more delineation or confirmation wells are
78 Exploration Plays-Risk Analysis and Economic Assessment
Concept: How large a field (or fields) must we find to recover all exploration investments, including interest?
Considerations:
1. Money has a time value. .
2. Some discovered small, noneconomic fields may be developed as commercial discoveries.
3. Several small but economic fields are a more likely outcome than one large field.
4. $4.50 PV /BOE determined from PlO%, P50%, and P90% cash-flow models, considers contract terms~ /
required to warrant field development, those costs requirement that one large fielc be discovered. In fact,
must also be included along with the anticipated costs however, if a company is successful at all in a new play
to develop the discovered field. In other words, the dis or large contract area, the usual outcome is the discov
covered new field(s) must be at least large enough to ery of several new fields. Accordingly, it may be a more
generate, through production revenues, an after-tax likely scenario to require three lO-million-barrel dis
NPV sufficient to pay for all investments required to coveries or two IS-million-barrel discoveries rather
establish the fact that profitable production does exist than one 30-million-barrel new field. This leads to the
(PV > $0).·' concept of minimum economic reserves required
The use of $PV /BOE is a practical convention that (MERR), to provide minimum profi ts sufficient to
includes development costs of successful discoveries cover exploration investments. The practical result of
because it is directly derived from DCF analysis of the such an enlightened criterion is to prevent more large
discovered and developed field. Accordingly, where area projects from being rejected because of unrealis
$PV /BOE is being used to establish MEFS, no addi tic, conservative biases.
tiollal provision should be made for de\'elopment costs
unless they are required to prove that an economic
discovery has been made. In some cases, necessary
Economic Truncation of
tion. Reasonable-and expectable initial production In mature onshore U.S. petroleum provinces, FSDs
rates and percentage deL;ine should be employed in of plays or basins contain constituent "fields" ranging
the cash-flow analysis, which should also utilize the down to as small as a thousand barrels each, or even
legal provisions, terms, and price structures of the less. In fact, such fields really represent shows of oil or
anticipated contract. It is important for all estimates gas that were completed only for business reasons, or
used in the cash-flow analysis to be rigorously and from geotechnical or financial imprudence, but never
objectively evaluated, thereby elimina ting bias of any theless must be included among all production, which
kind-whether optimistic or conservative. Estimating is regulated by the state.
ranges may be used where there is uncertainty about The useful terms commercial and economic were
certain parameters, and probabilistic project NPVs defined and discussed on page 38, as they applied to
may be calculated through Monte Carlo or Latin prospect-reserves distributions. Procedures for trun
Hypercube simulation. Equation 9 pro\'ides an exam cating such distributions were also reviewed. The
ple calculation of MEFS for an onshore play in a same procedures apply to FSDs. Appendix F provides
mature trend, assuming two fields. additional detail.
One of the most harmful conservative biases con
cerns the prCi~',k~ ~::. requ'!.j,'1g a single dTsco\'e:-jto .... Pmcfs and Pmefs
cover all exploration investments-that is, for MEFS to By estimating MCFS and/or MEFS and finding the
apply only to a single field. This usually results .in the location of those reserves values on the FSD, we can
Chapter 5 79
1% W 1%
f-4-+- 0
CO
, -
10%
(ij
"eQ)
•
I
xci
:2
(1):2
cu lO
--' •
.••
10%
c
III
:6
iii
EW
EO
oeQ
uo --
~
(1)-
Q)o
$"5
II
[,./
/"
Ii'
iii
(1)0
cuO
Xci Q)U V"
Q)
i3 50% ~~
(ii.S:?
EE
L 50%
/
~ - 0 .§ g V .'
0
g
(1),-
Q) II
$:1::
'-0
V V
-
0.0
:0
III
.0
Q)o
cu-, ~~
0 E ::J
.' ."
a: 'x U i
."
.• .
90% 0 90%
'-
a.
a. •
<{
•
.. ' .
99% 99%
10,OOOBOE 50k 100k 0.2MM 0.5MM 1.0MM 2.0MM 5MM 10MM 20MM 50MM 100lvlM200MM 500MM 1.0BBOE
Figure 38a
10'10
10%
c
III
10%
:6
<0
iii
~
(9
0~
50%
50%
~
:0
III
.0
e
(L
90%
90%
99%
10.000BOE 50k 100k 0.2MM 0.5MM 1.0MM 2.0MM 5MM 10MM 20MM 50MM 100MM 200MM 500MM 1.0BBOE
Figure 38b
estimate the probability of a commercial field (or Figure 38a shows an FSD of all fields in part of the
l~rger) or an economic field (or liHger) in the play,
gIven at least one discovery of sufficient reservoired
~ydrocarbons to at least sustain flow. Those probabili
prolific northern Midland Basin, \A,'est Tex.£l;: fields in
the distribution rilnge from 10,000 BOE (P9n;,) tot"JuO
MMI30E (Pl %). Generally, a well that encounters an
--
ties are Pmcf and Pine£, respeclively. accumulation of at least 10,000 to 20,000 BOE will flow.
Il
80 Exploration Plays-Risk Analysis and Economic Assessment
In that area, any well that flows will probably be com seem to overlook it. Reconstruction of the parent FSD
pleted for production, if a reasonable profit can be is possible, however, by reviewing all pertinent well'
expected on the investment in stimulation, tubing, and data and logs in the play area, and by distinguishing
production equipment. Of course, such a commercial dry holes that were uncompleted show holes from
well would not recover the investments in seismic, dry holes that did not encounter flOWing hydrocar
leasing, or drilling-they would be classed as sunk bons in the reservoir objective. Appendix F provides
costs. Accordingly, the commercial threshold in such a more detail on the particular procedures involved in
province might be approximately 10,000-20,000 BOE; such reconstructions.
naturally, that threshold might vary depending on the / I
details attending any particular well, especially flow , I
rate. So, for the Northern Midland Basin, Pmds equals i . Chance of Play Success
I
about 99% to about 96'Y" (Figure 38a).
However, for an exploration venture to be ecolloJllic Geologic Chance Factors ,
it must recover all investments involved to bring it A system of five discrete geologic chance factors ,
into production, plus a reasonable profit, taking into was presented in Chapter 3, pertaining to prospect
account also the time value of money. To estim.:.te the chance of geologic success (Pg), including definitions
proportion of fields in the northern Midland Basin that and descriptions of subcomponents involved in each
are economic, we must first estiinate the approximate of the five main categories (p. 34-36). The importance
reserve size required to generate such a profit (about of the chance factors applying to plays as well as
500,000 BOE),S and then determine the proportion of prospects was emphasized. Table 15 summarizes this
the FSD having that reserves value, or larger (Pmefs = recommended system.
75%). By excluding the noneconomic segment of the
parent FSD, the remaining segment has been trun Coincidence of Geologic Chance Factors
cated economically, just as previously described for In order for the geologic chance factors to work,
prospect-reserves distributions (pp. 39-41). The their influence must be operative in a common area
smaller, noneconomic fields have been excluded, that is, their effects must coincide in time and space
therefore the mean of the remaining distribution of (pp. 34 and 48). Figure 39 depicts a hypothetical basin
economic fields will be larger than the mean of the to illustrate the problem. Petroleum source rocks are
parent FSD, just as previously described. present only in the eastern end of the basin bottom.
Because buoyant migrating hydrocarbons move updip,
Offshore FSDs Are Already Truncated at right angles.to structural contours, it would be diffi
Now, suppose the Midland Basin actually was cult (even impossible) to charge the faulted closures on
located offshore, in waters up to 600 ft (-200 m) deep. the southwestern flank of the basin. Moreover, primary
The parent endowment of accumulations would be no reservoir sands do not extend far enough southwest
different, but the actual FSD would be greatly attenu (updip) to be present in the area of the faulted closures.
ated. Because of the expense of offshore production Also, the regional evaporate seal is not present in the
facilities, FSDs of offshore plays reflect dl?fnclo trunca western half of the basin. The result is a greatly
tion (Figure 38b). Ordinarily such truncation is com reduced chance of success because the key geologic ele
mercial rather than ecollomic. However, because the ments controlling the occurrence of reservoired
cost of offshore production facilities is generally very petroleum accumulCltions do not coincide anywhere in
large in relation to exploration costs, the' commercial this hypothetical basin.
threshold usually approaches the economic threshold Coincidence is one of the most significant elements
for offshore plays. The same principle holds true for that explorationists must assess, especially for new
all plays in which very substantial financial commit plays in frontier basins. It is probably one of the most
ments are required to bring a new field into effective commonly overlooked problems, which leads to very
production-for example, remote, l:!lra-deep, haz low industry success rates for high-risk new plays.
ardous, or hostile provinces. It is important to Play maps, as described previously, provide an excel
emphasize this point because many explorationists lent way to address the coincidence problem.
Shared versus Independent Chance Factors
Earlier work by Baker et ai. (1986), Baker (1988), and
RNaturally, we must recognize. that 500,000 BOE should not be con White (1993) discussed independent and shared
sidered as a precise value. The economic threshold will vary chance factors, and the concept was reviewed on
according to the iniluence of depth, completion costs, flow rates,
transportalion, wellhead price, and operating costs, including taxes.
pages 4J",;mr.! ~~ ._ -c. .
Nevertheless, it is pOSSible to generate an approximate minimum Consider the five elements of geologic chance~
economic fit'Jd size for fields in a play bec<ltlse pl<l\'s are generally some will apply to all prospects in the play area,
constrained by similarities in the abo\'!' f<lclors. whereas others will vary among all prospects in the
Chapter 5 81
Table 15 Geologic chance factors required for play and prospect success.
Confidence (%) that thermally mature HC-source rocks are present in adequate volume, richness, and type to
provide an HC-charge to the play area. Components:
• quantity [thickness, extent, richness]
• HC-type [oil, mixed, gas]
• thermal maturity
Confidence (%) that hydrocarbons have migrated, utilizing conduits and carrier beds, along IT)igration path- J
ways into the location of existing closures in volumes adequate to charge them. Components:
• conduits [carrier beds or zones]
• migration routes
• efficiency [concentration and transmission versus dispersion]
• timing
Confidence (%) that reservoir rock is present in adequate volume, poroSity, and deliverability to support one
or more floWing wells in the play area. Components:
• storage capacity [thickness, extent]
• porosity
• reservoir performance [permeability, drive mechanism]
Confidence (%) that structural and/or stratigraphic closures of adequate area and vertical relief are present
in the play area and can be detected. Components:
• adequate closures exist
.'" • confidence in mapping
Confidence (%) that effective sealing rocks are present and that emplaced hydrocarbons have been
preserved (= containment). Components:
• effectiveness [thickness, differential permeability, absence of open fractures]
• preservation from subsequent spillage [fault leakage, fracturing, breaching, tilt-and-spill]
• preservation from degradation [biologic, oxidation, thermal]
- - . .. .. ------ _._-_. __.----"_ _- --_.--_._ ...• _-- - _ _--- - _._._-.-_._--_. __ .. _ - --_._----.-_.-
..... ~
.,-.,
;
",
Figure 39 . Effects of operative ger;Jlogic chance factors must coincide in the prospective area.
,
82 Exploration Plays-Risk Analysis and Economic Assessment
•
play area. Examples of frequently shared chance factors
include the presence of mature hydrocarbon source
Integration: Calculating the
rocks, hydrocarbon migration into the play area, and Chance of Economic Play Success
timing. Independent chance factors commonly include The purpose of this section is to explain and illus
reservoir rock, closure, and containment. In some trate one of the mathematical procedures involved
cases, subsidiary elements of each main chance factor utilizing various geologic and economic estimates
may vary-for example, seal effectiveness (such as -to calculate the estimated chance of economic play
fault leakage) may be independent, causing some success (Equation 10).
prospects in the play to be dry, but not all, whereas the Part 1 of Equation 10 shows the basic form o(.the equa
regional preservation history (absence of biologic or tion, with the shared chance factors separated from the
thermal degradation) may be common among all independent chance factors. Given that the play exists
prospects and therefore shared. (shared chance = 1.0), the steps shown for the indepen
This leads, then, to the separation of geologic chance dent chance factors allow successive determination of:
factors for purposes of play analysis. Those elements
are shared that affect all prospects in the play area. If a 1. average local chance of a prosper:t's geologic suc
valid test establishes clearly that a shared chance factor cess (one test),
does not exist in the play area, it condenms all prospects 2. average local chance of a prospect's economic
in the play area-or it forces significant re\'ision of the success (one test),
play area outline (area of possible coincidence). " 3. average local chance of a prospect's economic
Finally, it is common to recognize partial depen failure (one test),
dencies-elements of geologic chance that are par 4. chance of all economic failures in 11 local trials,
tially shared (= dependen t) and partially local and
(independent). 5. chance of at least one economic discovery in /I
local trials. .
Chance of Play Success Is
Economic, Not Commercial This equation expresses the generally recognized
As previously discussed, the criterion for success of truth that, if the play really does exist (i.e., the product
the exploration play is that the play must be economic, of the shared factors = 1.0), then the chance of at least
rather than commercial. The reason, of course, is that one discovery increases as the number of wells testing
no responsible organization is going to continue to the independent chance factors also increases.
drill for fields that are mereh' commercial-after sev Part 2 of Eq1!ation 10 lists the estimates and assump
eral such marginal discoveries in a play, management tions required" to carry out the calculation. For most
will perceive the folly of continuing to explore in a established plays-even "undercharged" trends
play that cannot be economic on a full-cycle basis. That there is usually more than enough hydrocarbon source
is, they will not continue to drill for fields that do not rock and generated/migrated oil to supply the traps in
generate production revenues sufficient to exceed the the play, i.e., the adequacy of source rock and gener
invested capital plus interest. Accordingly, such a play ated/migrated oil is usually not a critical problem. \
will be abandoned even though some marginal fields However, ill frail tier basill~ and plays, especially those
\
have been found. reqllirillg very large discoveries in order to be economic, the
A further complication is presented by the play play analyst may be justified in assignillg a probability to the
which contains at least one economic field, but addi likelihood that SOl/rce rock and migratioll pi'ocesscs have bec/l
tional dry holes or seismic cos ts incurred in subse adequate to Sl/pply at least aile ecollo11lic field.
quent exploration for other counterpart fields add Part 3 of Equation 10 tracks the actual calculation
sufficient expense to transform the play from eco using the equation from Part 1 and the values provided
nomic to commercial on a full-cYcle basis. The solution in Part 2. Play (= shared) chance is estimated to be 0.54:
to this problem is to identify, for various reserves the analyst is quite confident of source rock presence
cases, the maximum costs for additional drilling and (0.9), and moderately confident about generation/ \
seismic that can be incurred before the entire project, migration adequacy (0.6). Given that the play does
on a full-cycle basis, deteriora tes from economic to exist, the average chance of prospect success (flovving
commercial. After disco\"en' of the first economic hydrocarbons) is 0.28, and the average chance of eco
field, each subsequent discO\'ery shouJd bear the costs nomic prospect success is 0.112. Therefore the average
of the number of dry holes that lNould be consistent chance of economic prospect failure is the comple
with the independent chance of economic success; i.e., ment, 0.888. The chance of drilling four consecutive
if Pe (independent) is .25, each anticipated discovery economic fai!~re? from such prospects is 0.622. There
--- should carry the cost of three dry holes, in addition to fore the chance ofTitakingat least one economic dis
the cost of the discO\'ery well. covery among four such prospects is again the
Chapter 5 83
1. BASIC EQUATION:
PSHARED X PSHARED (1 - [1 - (PINDEr x PI:\DEP X Pli':DEr x PMEFS)]NO.TRIALS)
I""z
,
, V'"
d": I
PLAY CHANCE I
I Al'g Local CII,lJIct' (~f SIlC(t'~~
I
I
I ~ I I
(= shared chage-es) I I.
I
I I
~
..._-----""""'~------~",.:
..... I
Ai'S Local ClwlCt' of £cOllOllli( SIICCt'S~ I
I
I
I
I
L ¢I'
¥
Al'S Lt'cal CIII1IIe',' of £conolllic Failllrt'
. 1IIlIIiIl_"IlIIIIIIIIIIlI_"_~"'_Ii!lI
"'-lllI_· Y' ImI_"#"':,
I
Cllanct' of All £Ct,,/(llllic Faillln'~ in n Local Trials :
, • ¥
., I
I
2) HC generation/migration = 60'X,
1) Reservoir = 70%
2) Closures = 80%
3) Containment = 50'Yo
d. NUMBER OF SUCCESSIVE DRY HOLES BEFORE PLAY IS ABORTED = oJ
= [.54] x [1 - (I - [.11LD~]
= [.54] x [1 - (.888)~]
=- [.54] x [1 - .622]
= [.54] x [.378]
f·
"i
I '
Table 17 Form: flow sheet for play risk analysis (modified from Baker et 0/., 1986; White, 1992, 1993).
(= Confidence - %) !=Confidence - %)
_ _ _ 3. Thermal Maturity
_ _ _ I. Conduits Exist
_ _ _ 3. Efficiency
_ _ _ 4. liming Correct
_ _ _ 2. Porosity
_ _ _ 2. Confidence in Mapping
geologic chance factors that are indepelldent (that is, Cost of Exploration
that may vary locally, among all the prospects in the The cost of exploration is the sum of estimated pru
play) are, 011 average, satisfied in the area of the play. If dent expenditures for seismic sun'cys (including pro
the play really does exist (play chalice = 1), thell prospect cessing), other geophysical data, other necessary maps,
~'hallce is the average geologic success ratio of all prospects surface data acquisition, geologic and geochemital
III the play area. It is the average product of the inde analyses, leasing costs (including bonuses and fees),
pendent geologic chance factors. Common indepen exploratory drilling and testing, and o\'erhead. It is the
dent chance factors include reservoir, closure, and total amotlnt that must be recovered (a l interest) from
some elements of containment (Table 17). Remember future production revenues if the veniurc is to be at
that prospect chance equates only with discovery of a least minimally profitable (= PY > $0). This is the
reservoired accumulation c<lpable of sustained flow. It amount of money you are going to have to spend to
has no economic or even commercial implications know whether the project is a success or failure. In
those constraints are derived values considered later some conservative project evaluations, this amount is
in the analysis. '
projected to be recovered through production revenues
Ii
86 Exploration Plays-Risk Analysis and Economic Assessment
from only one discovery; in more reilsonilble evahlil what is learned from each dry hole and the antici
tions, it may be recovered from several discovered pated capacity for geotechnical learning, adjustment:
fields (see p. 78). Conceptually, it also Ciln be considered and improvement.
as the cost of a failed exploration plity-and itS such it
should be tied to your minimum exploration progrum. $PVper BOE of Discovered Reserves
The play analyst needs to be able to translate pro
Projected Field-size Distribution
jected prospect reserves into present value in order to
The construction of an estimilted FSD for the subject be able to estimate the present value of the entire play
play has already been described at length, whereby as a full-cycle economic venture. In some plays, a use
informed manipulations (= shifts) consider geologic ful approximation of this parameter can be made from
attributes and exploratory maturity and efficiency. the median or mean field-reserves case and applied to
Regardless of the methods employed, the projected fields of all sizes in the play. In most plays, however, the
FSD represents a responsible estimate of the fields that analyst is advised to prepare economic analyses (cash
will be found, givell tlrl! discovery of OJ/I! or J//Orl? fidds, flow models) of key field sizes (PlO°!." PSO%, P90'Yo,
and it is a key parameter for risk analysis of any explo Pmean), using the contract terms and tax schemes
ration play. It is expressed as a lognormal cumulative expected in the area. This is because in most plays, espe
probability distribution. cially those with production-sharing contracts or off
shore are(ls requiring expensive production platforms,
Projected Numbers of Fields $PV (in dollars) per BOE is not a constant. Variables
Another required primary estimate is the number of such as initial production, decline rate, and wellhead
fields that will ultimately be discovered in the play price can be considered via such reserves cases. The
area (p. 69). Note that such fields need not be eco ultimate product required is a simple graph showing
nomic. Ordinarily this is expressed (IS (I lognorm(ll field $PV as a fu~ction of field reserves and/or $PV per
cumulative probability distribution. The projected 130E for different field sizes.
field number is usually determined by comparison
with analogous areas, taking into account relative Minimum Economic Field Size (MEFS)
maturity of exploration, but it milY also be estimated MEFS has already been discussed in detail, and
using the principles of discovery process modeling or Equation 9 shows the method of calculation, which
other methods. Obviously this estimate assumes at utilizes 0) cost of exploration; (2) $PV per 80E; and
least some future success in the plity. (3) minimum number of required discovpries.
Number of Exploratory Tests
Secondary (Derived) Parameters
Given that the shared chance factors turn out to be
satisfied in the play area (play chance = 1), then the Probability of Minimum
probability of success is a function of the independent Economic Field Size (Pmefs)
or local chance and the number of consecutive unsuc Pmefs refers to the probability, givell n discoVl!ry ill
cessful trials (dry holes) the firm is willing to drill the piny, that the discovered field will be equal to or
before giving up. Usually, this number is partly polit larger than the MEFS. If ~everal equal-size fields are
ical, based on management's preferences, but it is also anticipated to reach break-even economics, Pmefs
a function of what specifically is learned from each is the probability of either of them (but not both).
trial. This involves considerations of Bavesian mathe-· This parameter is derived using MEFS in conjunction
matics9 (see Baker, 1988). In general, cO~1panies expe with the projected FSD for the play, as described on
riencing two consecutive tests that indicate deficiency page 79.
in the shared chance factors are probably justified in
abandoning the play. Consecutive dry holes that con Projected Numbers of Economic Fields
firm the play chance factors, but appear deficient This parameter is expressed as a cumulative pruba
regarding the independent chance factors, might be bility distribution, ordinarily lognormal, derived from
allowed to number two to four or fiYe, depending on the projected number of fields distribution (p. 69),
reduced by multiplying by Pmefs.
Projected Economic Field-size Distribution geotechnical staff must be trained to understand the
_This FSD is derived froll'l the original projected FSD principles of play analysis if they are to use software
for the play, economically truncated at Pmefs and (or tlow sheets) responsibly.
redistributed through 98%. It is necessary to deter
mine the new PI0%, PSO'X" P90%, and Pmean of the Recommended Procedure
truncated di~tribution as described previously and in The follo\'\'ing discussion, which also draws on
Appendix F. prior \\"ork by Baker et al. (1986) and White (1992,
1993), is intended to guide the exploratjonist/analyst /
Projected Distribution of
through the proper steps of economic play analysis.
Eco-nomic Play Reserves
This distribution is the product of two distributions: Model and Map the Playas if It Already Exists
(1) projected numbers of economic fields, and (2) pro Depict all the requisite geologic elements in the con
jected economic field sizes. These two distributions templated area, 11:: if t!ley already exist-the following
are combined by tvlonte Carlo or Latin Hypercube analysis is then designed to estimate the likelihood that
simulation or by employing the analytical!grilphical your picture is essentially correct. In particular, you
method described in Appendix B. The result is a must show the areal and spatial distribution of geo
cumulative probability distribution of economic play logic features, on maps and sections, in reiatiun to the
reserves. The mean of that distribution (mean eco play fairway and to locations of specific prospects-if
nomic play reserves) is statistically the best represen you ha\'e that much data. Remember what you are
tation of economic play reserves. risking geographica-lly: the chance that at least one
flowable petroleum accumulation (hadng the geologic
Present Value of Economic Play Reserves
attributes described) really does exist in the play area,
This distribution is the product of two distributions: in the stratigraphic succession identified.
(1) $PV /BOE, and (2) distribution of economic play
reserves. The mean of the distribution is statisticallv Estimate the Play (=Shared) Chance
the best expression of economic play present \·alue. Based on Table 17, circle those geC'logic chance fac
tors that are s!lared, i.e., common to all prospects in the
Play Expected Present Value
play, and enter your confidence (probability = decimal
The expected value concept was discussed previ fraction) that each factor will indeed pro\'e to operate
ously. It is the chance-weighted present value of a pro or exist in the play area. If you are essentially certain
ject-it represents the play "s an entire, full-cycle that a shared factor exists throughout the trend area,
economic venture. It is one of the important economic assign it a probability of 1.0. Then multiply the circled
measures by \",hich competing play? can be usefully chance factors-the product is play cIlal/ce or s!lared
compared.
•
It may be derived usin<>0 (1) ch"nce of eco c!lal/ce. :\:01£: Individual chance factors may be partly
nomIC success; (2) mean $PV of economic play shared, and partly independent, i'ndicating partial
reserves; and (3) cost of exploration (= cost of a failed
, play). As explained earlier, its weakness is that it
geologic dependency.
Estimate the Average Prospect Chance
assumes the firm is risk-neutral.
Under Item 1 of the flow sheet (Table 17), circle
those geo!;:;gic cha;-,ce factors that are il/depel/del/t,
Process for Systematic Risk that is, thilt vary among the prospects in the play
'\
Analysis of Exploration Play:; Mea. Assign to each your confidence (= probability)
that the chance fact~rs will be satisfied for average
Now we wiJJ set out the various steps and calcula prospects, and multiply them. Some chance factors
tion.s r.eguired to carry out a geologically responsible, may be partially dependent, so you may enter deci
statIstIcally sound assessment of the economic value mal fractions in both the shared and the local
of a. given play, as previously discussed (p. 84). Note columns. If you have sufficient data, determine from
agalI1 that the values used in the flm\' sheet must be several prospects in the play area your average confi
based on a sound and thorough geotechnietli re\'jew of dence (= probability) in the existence of each of the
the subject plays, as outlined in Appendix E. independent chance factors and multiply them. 1O
Many exploration firms utilize comprehensive Their product is nl>ernge prospect dlal/ce (prospect Pg).
computer software, or a stancbrdized flow sheet, to If the play proves to exist (i.e., play chance is 1.0),
carry out systematic play risk analysis. Either then the average prospect chilnce will be the overall
npproach c<ln bring companywide consisten<;Y,Jo the
process, which then leads to pl<lY inventories that, ij~
turn, yield ]egitil11il[e optimized values to those plilYs
selected for execu tion-the play portfolio. However, '''liO\\'l'\'l'r, scc pal;l' -16 for..1ddilionaI pcrspcctivc on this sccnario.
Ii
88 Exploration Plays-Risk Analysis and Economic Assessment •
t,
A ::: grade for highest-chance prospects } l·
C ::: grade for intermediate-chance prospects
F ::: grade for lowest-chance prospects { Consider all
Geologic Controls •
l I,
, I,
(b) Step 2: ,
l.
.
I··
Approximate
Future :::
1 + (No. of A Prospects) + (No. of A + C Prospects)
Total no. of A + C + F Prospects
,
Success Ratio
I
success rate of prospects in the play. Some reality the project. Reality check: Following Baker (1988), this
checks are: number ordinarily should be more than one and usu
ally no more than four or five. Key questio/l: What do
• If the play is a future projection of an ongoing you expect to learn from each dry hole? IllIportmltllote:
play, is the historical exploration success ratio This assumes only one play concept to test.
compatible with your estimate?
• If the play is undrilled but you have a comparable Estimate the Cost of Geotechnical Exploration
productive analog, what is the average prospect 'What is the area of the play? How much will
success rate of the analog play? Be sure you take regional gravity and magnetics cost? Interpreted land
modern technological capabilities into account. sat imagery pr air photos? Surface geologic mapping
• Use White's "prospect grading scheme" if you and lor sampling? How much seismic data (regional 2
have enough data to identify the location of many 0, local 2-D, and 3-D data) will be required? At what
of the prospects in the play, as shown in Equation cost? How many successive exploratory dry holes
11. The problem with this approach is that com would your company drill in the play before declaring
monly, if you have sufficient data to be able to the nnture a failure? What will they cost? Remember
identify and grade so many prospects, you're to include nonrecoverable bonus and lease costs here \.
already ill the play! That is, you han already as well as overhead, and make all estimates on a~
invested capital for geotechnical information after-tax basis. Reality check: Another term for this
the decision to enter the play has already been valuE' is the cost of ,':;ploratioi; fail/lre.
made!
• If you have absolutely no idea about the average Estimate the Number of Fields in the Play '"
prospect success rate, remember that the world Based on the size of the play area, structural and
average for the petroleum exploration industry stratigraphic grain or complexity, and field density in
has been consistently about 25% for the past -1:0 analog trends, forecast the total number of fields that
years (Pcommercial, not Pgeologic!). will be discovered in the play area, given that tl,ere is at
• Reality check: If you ha\'e an inventory of leasl ol/e discol'ery. This forecast should take the form of
prospects in the play, be sure the average Pg of a probabilistic range with a lognormal distribution
such identified prospects is consistent with aver (see p. 69). An alternative method is to employ discov
age Pg for the play. ery process modeling principles to derive these esti
mates (see p. 69). Remember the common biases: Most
How Many Consecutive Exploratory
explorationists l/Ilderestimate the /lul1lber of future
Dry Holes before your Firm Quits?
fields in plays and overestimate the reserves sizes of
BClsed on your firm's risk propensity and past histo~\. those fields, especially in onshore areas.
and assuming your wells establish that the play chance
factors appear to be satisfied (play chance = 1.0), esti Analyze the Terms of the E&P Contract
mate the number of consecutive dry holes your man Be sure to incorpora te all.relevan t terms, including
agement would be likely to drill before abandoning taxes, into cash-flow models that you develop for
Chapter 5 89
possible project scenarios. Identify particularly criti estimates represent statistical values, it is permiss\ble for
cal, sensitive, and useful provisions that should influ these values to be decimal fractions.
ence your negotiations and/or business decisions.
Find the Chance of Economic Play Success
Construct a Field-size Distribution for the !,Iay
This calculation has already been discussed and
Details have been discussed (pp. 67-76 and Appen illustrated (p. 82 and Equation 10). It is derived using
dix F). Subject the projected FSD to reality checks. the following:
Find Net Present Value for Different Field Sizes
• play chance = product of shared fac'hJl's
. /
Run cash-flow models on PI0%, PSO%, P90%, and • a\'erage prospect chance 7" product of local or
Pmean reserves cases under the operative contract independent factors
terms and using the current company discount rate. • number of consecutive dry holes before failure
Construct a graph showing the relationship of NPV to • probability of MEFS (Pmefs)
recoverable reserves (p. 86). Also run cash-flow models
-', for the total failure case and at least one partial failure
Construct the Economic Field-size Distribution
.... case to determine the negative NPV for those outcomes. This distribution of economic field sizes is derived
All values should be estimated on an after-tax basis. by truncating the original FSD constructed for the play
(p. 87). It is essential to determine the new PI0%,
Estimate Minimum Economic Field Size (MEFS)
PSO%, P90%, and the Pmean reserves values for the
Use the graphs and data from the previous item to ecollomic part of the FSD, as described on page 79, and
estimate the minimum economic reserves required in AppendiX F.
, (MERR) to recover the cost of exploration at interest.
Then, given at least one economic discovery, what is a Find the Mean Economic Play Reserves
reasonable low-side outcome-one field discoverv? ll1e next step is to construct a cumulative probability
Two? Three? Whichever is the most reasonable o~t distribution of economic play reserves. This is accom
come will determine MEFS. Reality check: Large play plished by combining the distribution of economic field
areas and geologic complexity both encourage field sizes and the estimate of the number of economic fields
numbers larger than one, but prudence suggests that in the play, ordinarily through conventional Monte
-\ this field number should not exceed three. Short explo Carlo simulation or the Latin Hypercube method (p. 94).
ration time periods (contractual), limited exploration An analvtical method to combine either two or three
budgets, and risk-averse management all drive the such distributions, using graphical procedures, has been
estimated field number downwilrd toward one. developed and described by Capen (1992), Megill (1992),
and Rose and 1110mpson (1991), as discussed in Appen
Find the Probability of a Field .
This value is the chance-weighted statistical value of cost of geotechnical failure), it may be necessary 4
the play. Play ENPV utilizes the following parameters: to perform several iterations to develop an appro
• chance of economic play success and chance of (= winning bid). Such iterations may impact val
• play cost of failure To provide a basis for evaluating the £.]ay as a busi
ness venture, or to compare it with other opportuni
It should be calculated on an after-tax basis. ties, it is useful to calculate the anticipated values for
key economic measures, such as:
Determine the LeaselAcquisition Price
This determination will vary widely, depending on ~
discounted cash flow rate of return (DCFROR)
the method by which contract rights are to be acquired (calculate at PI0'Yo, P50%, P90%, and Pmean eco
sealed bid, oral auction, performance bid (= work com nomic reserves level). DCFROR is only useful as a
mitment), serial negotiations, or priYate treaty. This minimum qualifier, or hurdle-it is not useful to
tion on Acquisition Strategies (pp. 93-99). There are • investment efficiency (calculate on a risk-adjusted
• If there is a reasonable likelihood that the host i.e., include t.he costs of dry holes).
\
Risk-adjusted Value, Risk Tolerance,
Practical Expressions and Applications and Optimum Working Interest
\
The term "risk" implies the threat of loss. A pro
The economic parameter risk-adjusted value (RAV)
posed project becomes a risk venture only when we
was defined in Chapter 4 (p. 55) and shown to be a
assign economic conseq-uences-gains and losses-to
modification of the expected value (EV) equation, in
",
the various possible outcomes. The fundamental
which the function that expresses the risk preferences of
attribute of all risk ventures is that the threat of loss
the decisionmaker is exponential. The relative degree of
" outweighs the prospect of gain-people will take a
risk a\"ersion expresses itself as variations in "r"-larger
greater chance to avoid a loss than to make a gain of the
r's indicate an increasing degree of risk aversion.
same size. This is the essence of utility theory. Except
Calculated ,'alues for RAV have been intuitively
for compulsive gamblers anci those who vurposefi..lly
hard for planners and decisionmakers to deal with
remain ignorant, the normal human condition is to be
because they combine objective economic values with
risk-averse whenever a proposition involves things
subjective behaYioral responses. Subsequent develop
that are precious-life, health, money, reputation,
ments made applied utility theory more useful to
human rel~tionships,social status, and so on. exploration management:
Accordingly, we try to "hedge" our position in such
91
92 Management of Exploration Projects as Business Ventures It.
Table 18 Biases affecting risk decisions (modified after Tversky and Kahnemann, 1974).
Framing effects Decision makers will take a greater gamble to avoid a loss than to
Existence of a prior account Decision makers are more inclined to take a risk at the beginning
Maintaining a consistent reference frame Decision makers are most likely to invest during a "run" of good
Wrong action versus inaction Managers prefer to take a risk by not making a decision, rather
Number of people making decision Groups are more prone to take risks than individuals.
Workload and venture size Large-volume ventures are preferred over smaller ones, especially
Personal familiarity The "comfort bias"-decision makers are more risk-prone in deals
The power of the OWl calculation is that it allows negotiations, or by leverage, because of an inherent
I
EACH SUCCESSIVE DECISION WEIGHS
INVESTMENT LEVEL •
~
J I.
, J
,
CONCEPT
LEAD PROSPECT DEVELOPMENT
Figure 41 Sealed bidding far uncertain reserves leads ta the winner's curse.
for success: The goal is I/ot to wil/ tlte tmct; the goal is to After area-wide sales began, overall bid prices
add vallie-to //lake II/Ol/el/. Second lv, limit the bid to dropped to 12.s'~;, of their former level.l~ Some compa
Some deep discount of ti1e tract expected net present nies, such as Shell, Chevron, and Amoco, continued to
value (ENPV): the more uncertainty, the deeper the bid efficienth' and kept acquiring offshore acreage for
discount. Naturally, this presupposes that anticipated proportionat~ly fev.'er dollars (Figure 43). Others, such
competitive bid levels for the particular sale are as Exxon, Texaco, and Mobil, had apparently learned
expected to rise above threshold trend per-acre prices from their prior bidding experience and switched to
of, say 5-10% of ENPV or anticipated minimum bid efficient bidding. On the other hand, a few firms,
levels, as they did in the 1996-98 Gulf of Mexico sales. notably Arco/Vastar and Unoca!, apparently lost their
Third, bid widely: recognizing the large uncertainties corporate memories and nmo\' became inefficient bid
inherent in exploration, firms should bid on all blocks ders. Some latecomers, such as Kerr-McGee and
perceived to have positive expected value. Fourth, Amerada, bid \'ery aggressively but inefficiently.
encourage a detached, disciplined bidding attitude, to
wit, "!fwe call't get tltis tmct for ol/r price, we dOIl't want The Ubiquitous Overbid
if." Fifth, as our ind lIstry was able to do in 1983, The second operative drawback of sealed bonus
encourage the state to reduce the intensity of competi bidding is the Ubiquitous Overbid of Megill and
tion by offering very many tracts, as the U.s. Minerals Wightman (1984), defined simply as "the money left
Management Service (U.S.M.MS.) does through area on the table"-the difference between the winning
wide sales. Finally, seek other ways to acquire min bids and the second bids, as a percentage of the win
eral rights, such as with private treaties and through ning bids. From the company's point of view, the
farm-ins, trades, or acquisitions. overbid represents a totally ",..asted investment.
Adoption of the are<\-\·vide leasing procedure by the Although counterintuitive, the fact is that \·vith less
l!.S.M.M.5. in 1983 reduced the intensity of competi competition the average percentage overbid tended to
tion for Gulf of Mexico leases. Before 1983, competing iI/crease, from about 50'1.. pre-1983 (Figure 44) to about
companies like Shell, Arco, c\J1d Chevron v,'ere using 7S'}';, post-1983 (Figure 45). When we deal only with
deep bid discounts 'lI1d bidding \o\'idely (Figure 42).
They were bidding efficiently--acquiring their
--- acreage for fewer doHars than others like Exxon,
Mobil, Gulf, Tenneco, ilnd Texaco, who were cono:'n
12This compMCS ~\'l~r~ge bid prices on ~ nominal doll~r basis; on a
"rc~I" b,lSis, i.e., t.. king infl~lilln into ~ccount, the ~\'l'roge per-~cre
reduction of bids is much gre~tcr, probably to about -t% to 5% pre
trating their biljs and p{\~'ing top dollar. aren-\\,jde snles avcrilgc .
•
96 Management of Exploration Projects as Business Ventures
a
12
10
I % OF # OF BLOCKS WON
Co
D% OF $ SPENT TO WIN BLOCKS
8 I-:--
I
6 f- .- .. '
; i.
..
4 -- - f--
'.
t-
2 , .'
~I - f---
.'. f---
t-- -
,
'!".)f, f
,
-.
'" ~~ .{1
.,
--
o : t~
i~~
~i ','
-c-
'0.,..
,
~ Z < < ~ z w
~
w
0..
>:Q
0
U
~
.
--
Z
....,
~
::: 0
><:
0
u
<
0
<
::r:
~
0
u
0
<
u
0
::r:
w
c;
-..
::z:: < > ><: X ~ ~
,.... :E 0 E-o u
en ~ :E
~ ~ W
E-o
t:J
~
N < Z
;:> <
P:: ~
:.J < < c:::
:E ~
::t:
Figure 42 Net purchases by company: Gulf of Mexico sales, 1972-82 (figure courtesy of Robert Clapp). Compare
12
"
I % OF # OF BLOCKS WON
10
~
o
L....-
% OF $ SPENT TO WIN BLOCKS
._
6
"-
4
-
~
_":":,'
Jt
"..
~
;l
.~
~I
2
'.'f--- ,-' "
<-
,.- i',f---
't, .....
"
~
':,t-.
Iz
.1·
""
n
0'"
.~.
.;.
. "~.'
.~; .....: ~ ",
' ~~i· .~ ~
,\)' .
,
...~.;'
0
...:J
...:J
0..
l:Q
0
U --::::z
-..I
-=:: ,
Z
0
0
u
<
0
<
::z::
0
u
~
< 0
~
~
~
::z::
P::
< > -...., X
X
<
X
<
:::: ::t:
,....
...:J
0
~
u
0
::r:
E-o
r ...
'-'
u
__ C". •
en
-
:=::
...:.J ~ w w
E-o
~
~
<
N < z
;:J
<
~
<
~
~I
,.;
0'.:
~
::t:
Figure 43 Gulf of Mexico biddim) efficiency (1988-95). Compare with figure 42.
Chapter 6 97
8000
- 60 4=~
~
7000 50 ~
~ ~
-f:F) 6000 40
;,
0 0
/
CD 5000
(f)
30 ~~
~
Z 4000
0
CO
UJ 3000
(f)
«
w 2000
--.J
~
~ 1250 +--=-+-:-::J-:--:~hI-:r.~:"':+:X=----+--:~~~~-=---:::+~-=-L~~~
Y}
o-
1000
OJ: AVERAGE OVERBID::: 75%
if)
::> 750 -r-t----+---+---f--t-1f--+---I--+-+-,--I--+--+-----'+-"''--Ir--H
z
oOJ
W 500
if)
<{
UJ 250
-.J
TOTAL
SECOND BIDS
0
1 21 sales
LO
.. N LO If) .. C") a If) ..CD eo ..
N ~ ~ CD t--.
..N
C") LO
.. .. .. .. ..
LO
.. .. .. .. .....
GOM SALE NUMBER AND-6EQ'JEi-+SE
Figure 45 High bids and second hids; area-wide GOM lease overbids averaged 75% post-1983.
iii
98 Management of Exploration Projects as Business Ventures
multiple-bid tracts, the percentage overbid becomes confidentiality discourages the outside input thnt
smaller-45% pre-1983 and 57% post-1983. Now, might indicate overestimation. Near-term deadline~
overbids are characteristically large, simply because promote similarity to sealed bonus bidding. How
bids-based mostly on reserves potential-are log ever, if the deal seller sells a fractional share, a: limit is
normally distributed, hence the difference between thus set on the upside potential, and the inherent
the first and second bid is generally large, compared advantages of competitive confidential bidding have
with, say, the fifth and sixth bids, because of the been eliminated.
log scale. As Megill and Wightman pointed out,
prospectors and their managers must understand_ J
Oral Auctions
that the overbid is intrinsic to the sealed bid p'"ocess.
Overbidding is part of the mathematics and' can't be The third common method is the oral auction, which
eliminated. is still used by some U.s. states and onshore federal lease
It is instructive to compare levels of overbidding sales, and in privately managed sales of producing
before and after the adoption of area-wide sales: properties. Theoretically this method should be more
Figure 44 was made by Bob Megill in 1983 and shows attractive to deal buyers because there is no need for
that the average overbid before arc-a-wide sales was more than a fractional difference between the high bid
about 50%, whereas Figure 45 (Rose, 1999) shows that and the second bid, thus eliminating the overbid and the
the average overbid rose afterward to around 75%. inherent waste associated with it. However, competitive
However, since the entire bidding level had been so egos often seem to produce the same effect as the Win
greatly reduced, as it was starting in 1983, the pain of the ner's Curse, generating bids that indicate greatly over
increased overbid could be accommodated more easily. valued tracts. So success in oral auctions requires
disciplined bidding, as well as the recognition that the
Summary behavior of other bidders may provide useful on-site
Thus sealed bonus bidding has these two main information about your own valuation of the tract
drawbacks: the Winner's Curse and the Ubiquitous whether it may be too high or too low.
Overbid. However, their negative effects can be ame
liorated by substantially discounting the tract's Performance Contracts
expected value to determine appropriate bid levels, Performance bidding (or so-called work-contracts)
and by reducing the intensity of competition on indi should eliminate both the Winner's Curse and the
vidual tracts so the net impact is economically tolera Ubiquitous Overbid because, theoretically, opera tors
ble. Nevertheless, even after area-wide sales began, the should limit their bids to whatever a prudent explorer
key to success in bonus bidding remains to persuade would spend to explore any given block, based on
executives that their intuition cannot beat the effects of tract size, geology, and geophysics, not on perceived
the Winner's Curse and the UbiquitousO\·erbid. reserves potential. This sounds good in principle; in
The actual effect of the Winner's Curse and overbid practice, however, the competitive desire to obtain the
ding on the overall profitability of U.s. oil companies block often translates into extra wells or better terms.
operating in the Gulf of Mexico is chilling-Lohrenz Such measures are not necessarily bad, as long as deci
(1988) reported that the total industryinHstment in sionmakers recognize that the Winner's Curse is oper
the Gulf had not yet paid out and likely would not ating on all such overages and extras. Obviously, the
ever achieve actual profitability. Furthermore, if area worst of both worlds is represented by acquisition
wide bidding had not beep adopted, it is quite possible procedures combining performance bidding as well as
that the economic development of deep-water discov competitive sealed bonus bidding.
eries in the Gulf of Mexico during the late 1980s and
1990s would never have been possible. Private Treaties
The best method for the deal buyer to acquire
Serial, Time-constrained Auctions petroleum rights is through private treaties or simple
The second common method for acquiring
bilateral negotiations, where a given deal is shown to
petroleum rights is the serial, time-constrained confi
only one buyer at a time and multiple buyers are never
dential auction, in which a deal is shO\\'n individually
looking at the same deal simultaneously. This is the
and privately to several potential buyers over a short
method traditionally employed in conventional
period, with an announced deadline in the near
onshore U.s. exploration. It is best for the deal buyer
future. Such deals function like sealed-bid sales, with
because it avoids the Winner's Curse as well as the
the same drawbacks to purchasers and advantages to
Ubiquitous Overbid, and it's worst for the seller for
the deal seller. Whichever potential buyer valu~s the
the same reasons. Such deals may_:;.:'!-;roach staged
deal the most is likely to submit the best offer, and
exploration if the deal involves a new concept ortool
Chapter 6 99
and competition is as yet minimal. Such situations, Prospect and Play Portfolios .
however, don't last long.
Corporate Acquisitions
-- . ~-
Requirements
A prospect portfolio (Rose, 1992a) is selected from
Our final method for acquiring petroleum rights is an ilwentory of exploratory prospects. The portfolio
through corporate acquisitions-buying another com listing displays their respp.ctive costs, chances of com
pany's reserves by buying the company. Friendly mercial success, estimated mean reserves and mean
takeovers function like private treaties, generally being NPVs, risked mean reserves and risk~d NPVs (the laP
most favorable to the buyers. Unfriendly takeovers that ter are derived by multiplying estimated mea~
lead to bidding wars function more like sealed-bid reserves and NPVs each by chance of commercial suc
sales-the more bidders, the more favorable to the com cess), and preferred economic ranking measures such
pany being bought. However, the impact of the Winner's as ENPV and risked investment efficiency.
Curse may be relatively less because there is generall~' Each of the prospects in the inventory may be effec
less variance surrounding a company's true worth than tively compared and ranked ,;\,ith the other prospects
there is around an offshore exploration block. because the same processes for estimating resei-ves,
chance of success, and profitability have been used for
Sanctity of Contracts versus
all. Consistent corporate hurdles and discount rates
have been app.Iied to each prospect, and meaningful
Subsequent Renegotiations
economic measures have been consistently used to
All of the observations and comments have assumed rank the various ventures in the inventory. So the
the sanctity of contracts-"a deal's a deal." If it's possi inventory represents those prospects that the firm is
ble to renegotiate a de;)1 later, after disappointing cOllsidering; the portfolio represents the exploratory
results have been confirmed, then all bets are off and wells that wiII actually be drilled.
advantages and disadvantages of the previous six 111\'entories become another selective screen in
methods are obviated. This by itself may suggest a sev choosing prospects for the company's annual or semi
enth method of acquiring rights through subsequent annual exploration portfolio, but only if the various
renegotiations. Such methods are anathema to most candidate ventures considered for entry can be evalu
Western corporations, perhaps because when such sub ated quickly and efficiently, both geotechnically as
sequent renegotiations took place in the past they repre well as economicallv.
sented violations of existing contracts. However, Similarly, a portfolio of explora tion plays may be
perhaps this is a limiting Western cultural value-sup selected from an inventory of candidate exploration
pose we approach international exploration, from the plays. Ordinarily, however, play inventories and port
start, as an uncertain business that should 10(~icaIIy be folios are considerably smaller than prospect invento
carried out through agreements that expresslyOallow for ries and portfolios; moreover, prospects and plays are
changes in terms, depenu ing on wha t the results of segregated into separate inventories and portfolios.
exploration turn out to be? Because of their inherent differences in scope and time
This may represent the last, best way to improve frame, plays and prospects should not be included and
exploration profitability-to develop new, flexible compared in the same inventory. It is far better to
t~pes of contracts that allow us to approach the effi maintain two separate inventories, one for plays and
cIency of staged exploration. Of course, such contracts another for prospects.
presuppose that an informed landowner or state must
b~ able to ve~lty the legitimilc)' of the crilicai geotech Benefits
~lCal.expendltures and findings that impact the chang
If these conditions nre met, the assembly of a
Ing nsk-versus-reward picture, and thus, the changing
prospect or play portfolio from an inventory of qualify
terms. Some aspects of production-shnrin\~contracts
contam . suc I1 fIexibilities. tJ ing prospects or plays cnn significantly improve corpo
rate exploration performance, for at least six reasons:
Conclusion Optimizing Capital Allocation
. Small companies or independent operators espe If the selected portfolio contains those projects that
cl;"ll.ly shou~d realize that the method of sale itself plays rank highest, llsing risked investment efficiency (see p.
n bIg part In the profitability of the purchase or sale. 54), it will produce the high9St possible cnpacity to cre
Stnted simply, when you're selling, try to utilize the ate value for the firm. A r<1l1king bused on ENPV will
scaled bid or auction model. When you're buying, buy produce n portfolio hnving the highest ENPV, but
through private treaty or performance bidding. because investment costs may vary i1l11ong ventures,
100 Management of Exploration Projects as Business Ventures
that ranking may ignore venture risk; at the same time predictions. So the inventory / portfolio process helps
it may not maximize value. Any economic measure promote systematic performance review. This
that includes a provision for risk aversion, such as cal approach is not as readily applied to play analysis,
culation of OWl, will necessarily reflect value reduced however, simply because of the long time elapsed
from optimum, in consideration of the reduced vul between forecasts and discernible results common to
nerability to loss. play development.
Forecasting Performance Eliminating Predictive Bias
Prospect portfolios (Table 20) are one of the most Motivational bias expressing overly optimistic or / .
effective tools to improve exploration performance. overly conservative estimates of reserves, chance, and
Objective professional estimates of an individual profitability, can, with a reasonable number of trials
prospect's chance of commercial success, in combina (wells), be detected, analyzed, and corrected through
tion with reliable forecasts of mean reserves (and thus feedback and subsequently modified procedures by
prospect mean ENPV), provide the basis for predict geotechnical staff. By reducing such bias we can
ing the following: impmve portfolio performance and create added
value for the company.
1. approximate number of discoveries from a given
multi-well program; Scheduling Future Work
2. approximate total new commercial reserves The selected portfolio forms the basis for planning
added (p. 33) and their present value; and and scheduling the work necessary to carry out the con
3. approximate program cost-of-finding, using pro~ stituent ventures over the period of the portfolio.
ject cost forecasts.
Naturally, the accuracy of such program forecasts is
Lognormality and Performance
keenly sensitive to (1) average prospect discovery of Prospect Portfolios
probability; (2) variance in indiYidual prospect Most knowledgeable explorationists and many of
reserves distributions; (3) the predictive skill and lack their managers now accept the principle that prospect
of bias of the geoscientists; and (-1) the number of reserves distributions are lognormal, reflecting natural
prospects in the inventory. Accordingly, the predictive processes of multiplication [area (acres) x average net
ability of the play portfolio is much inferior to pay (feet) x He-recovery (bbllacre-foot)J. Accordingly
prospect portfolios. However, play inventories are the distribution of most corporate balc1l1ced portfolios
quite useful because they facil~tate the comparison of is also approximately lognormal.
different exploration plays. ," What is remarkable is that many corporate officers
and high-level exploration managers have not grasped
Guiding Geotechnology the implications of this principle as it impacts the mag
Provisional risk analysis may be carried out on nitudes and timeframes of corporate exploration
emerging prospects and plays before they are ready results. In particular they do not seem to understand
for drilling to identify those ventures that seem to the expected natural pa.ttern of annual portfolio out
have the greatest economic promise. To maximize comes: predominantly mediocre annual results punc
cost-effective use of geo~cientists nnd their tools, spe tuated occasionally by exceptionally good years and
cific exploration technologies s]1ould tnen be focused bad years. It can be demonstrated that such fluctua
on the highest-ranked anomali.es, leads, and trends, tions may have nothing whatsoever to do with geo
and especially on the critical geologic chance factors technical or managerial skill; rather, they may be the
(see p. 38). natural consequence of repeated sampling from nat
ural lognormal prospect-reserves distributions.
Assessing Predictive Performance A common manifestation of management's misun
In order to construct an inveritory (from which are derstanding of the lognormal principle is their continual
selected the best ventures for the company's drilling and excessive reorganization of ongoing exploration
portfolio), estima tes of reserves l chance of success, programs, in the well-intentioned but mistaken belief
critical risk, initial production rates, percentage that such adjustment and interference (= "tweaking")
declines, and drilling and completion costs must will improve year-to-year explora tion results. Bll t
already have been estimated for each venture. Thus exploration is inherently a sustained, long-term process
half the task of geotechnical performance evaluation plagued in most corporations by short-term interfer
h~iS ~lready been carried out-a,nd preserved! After ences. There are indeed effective criteria by which
the resWl:s oreach venture are known-successful or exploration performance can be judged, to distinguish
unsuccessful-the actual results (!an be assembled by luck from skill; however, prediction of ,mnual discov
th.e exploration team from drilli-~g and completion ery volumes is not an effective criterion unless the
reports and postdrill reviews and compared with the exploration portfolio contains 60 to 100 or more triClls.
-
.- _._._--~----------------------------~--
Table 20 A model prospect portfolio (prospects are ranked in this list by ENPV).
A 6.2 0.05 5.0 50.0 112.5 275.0 2.5 15.0 0.03 7.86 7.86 1 15.0 = 0.52 0.103
B 5.4 0.10 2.4 24.0 54.0 120.0 2.4 8.8 0.07 7.14 7.141 8.8 = 0.81 1.224
C 3.2 0.20 2.2 8.0 16.0 32.0 1.6 6.0 0.26 3.84 3.841 6.0 = 0.64 '0.445
D 4.0 0.15 1.0 10.0 22.5 45.0 1.5 7.2 0.14 3.35 3.351 7.2 = 0.47
..10.214
E 1.8 0.15 0.6 6.0 13.5 ;1'1.0 0.'1 1..1 O.B 2.07 2.07/ 3.3 0.63 0.305
F 2.0 0.20 1.1 4.0 9.0 14.0 0.8 2.0 0.35 1.20 1.20/ 2.0 = 0.60 0.197
G 0.8 0.25 0.3 ;.0 4.4 6.0 0.5 1.8 1.00 0.90 0.901 1.8 = 0.50 0.516
H •j 1.5 0.20 0.5 3.0 6.6 9.0 0.6 6.1 0.39 0.60 0.601 6.1 = 0.10 0.111
0.5 0.30 0.2 1.0 2.2 2.5 0.3 0.8 1:00 0.40 0.40/ 0.8 = 0.50 0.309
0.4 0.40 0.1 0.5 1.0 1.0 0.2 1.0 1. 00 0.16 0.16/ 1.0 = 0.16 0.137
N = 10 25.80 2.00 108.5 52/l.5 11.3 52.0 ~.57 27.52 IE pro9rilm = 0.53 2.56
Avg =.2 Avg = .46
------_._.-----_._ .. -~ ... - • • _" _ . ~ ••••• _ _ ~ _ ••• 4~ •• _ _• • •
-". _.- ..... -._ .. _---_ ..• _--._.---------
NOTES: 1) Dry-hole cost includes exploratory drilling 8t cOl1"pletion, I~nd, G8tG, and overhe~d.
2) Firm's r=5150MM=.I.
3) Firm's RT=1/r",1/.1=10.
4) Order of prospects changes if ranked on investment efficiency (IE) or RAV.
ANTICIPATED RESULTS: This portfolio of 10 exploratory wells is a bC1lanced program including three lower-risk extension wells (G, I, 8t J), five medium-risk trend wildcats (C, D, E, F, 8t H), and
two high-risk new-field wildcats (A 8t B). The most probable outcome of this program is: two discoveries, totaling 11.3 MMBOE reserves, having a total mean program expected value of
$27.52MM. Cost of finding should be about 25.8/11.3 = $2.30/BOE. Progr~m EPV/lnvestment = 27.52152 '" 0.53. 1"'1
:::r
QI
...
"g
..."
01
...
...o
....
to
102 Management of Exploration Projects as Business Ventures
••
t.
Table 21 Simulation of results for a prospect portfolio.
We have an inventory of 20 exploratory prospects. The prospects have varying chances of success, and each has a
lognormal distribution of reserves (if successful) with P10% divided by P90% = 13. In this exercise we will use the
,
"
given probabilities and simulate the results of drilling the inventory. To sample randomly from the lognormal l
reserves distribution of each successful prospect, multiply (Mean Reserves if Successful) by (Multiplier).
Accordingly, assessing exploration performance based forecasts of portfolio outcomes. However, in providjn~
on annual portfolios may require 1-5+ years, depending a proper answer, management must be asked, \Vhnl
on the size and aggressiveness of the firm and the num level of confidence do YOll reqllire?-50%? 68%? SO'Y.,?
ber and character of ventures in the annual portfolio. 90%? 95% ? or 99%?
A useful answer can be provided in at least two dif
ferent forms:
Predictability versus Portfolio Size
Assuming that the staff's estimates of prospect 1. A range of new reserves or NPV added, such <lS
reserves and chance of success are geotechnically "80% confidence in new reserves of 3.5 to IS,::;
responsible and unbiased (utilizing the estimating con million BOE (barrels oil equivalent)"; or
cepts and procedures described predouslYJ, the num 2. 50% confidence (or some other probabilistic con
ber of prospects in the portfolio influences the precision fidence level) that "at least 8 million BOE Dr
and reliability of forecasts about portfolio results. The $25MM NPV will result."
average chai'i>£.? I:\f !?.,!;.ospe.rtsuccess and the prospect -C" . .
reserves variance also influence portfolio predictability. Table 2fl\::presents a model 20-well exploratory
But understandably, management commonly wants portfolio for a domestic U.s. firm. Prospect discovery
, to know how many wells are required to make useful probabilities are 10%, 20%, 30%, 40°/<" ,(,ld 50"';· (iour
Chapter 6 103
CfP.
/ OJ>.
prospects in each chance class), with a portfolio aver alternative to a computer-driven Monte Carlo simula
age chance of Success of 30%; all chance categories are tion, which may not be understood.
represented by a mechanical spinner (Figure 46). All Figure 47 shows that as such a portfolio increases
prospect-reserves distributions are lognormal, with from 20 wells to 100 wells, the confidence in forecast
estima ted prospect mean reserves of 11.5 MM bbl and results improves, the P10%-P90% range becoming reJ
an 80'}'0 range (P90%-PlO% estimates) of 3.0 MM bbl to ativelynarrO\\'er with respect to the mean.
I
22.6 MM bbl. Median (P50%) is 8.8 MMllOE.
All prospects have P10%/P90% ratios of 13, as
Practically speaking, as many as 80 trials (four suc
cessive Table 21 portfolios) may be required to provide
I established by the outer multiplier ring on the spinner 80'};, predictive confidence in new discovered reserves
r (Figure 46). This represents a minor inconsistency volumes egual to ±50% of the predicted mean of 4 x 11.5
I because variance typically increases among high-risk, = 46 MMBOE, if that is the level of confidence manage
large-potential prospects. To operate this simulation, ment desires. If a more conservative portfolio were
each prospect is first tried for success or failure, using selected, one in which the wells had lower-variance
the graduated inner ring. If a prospect is a discovery, reserves and higher chances of success, perhaps as few
the next spin determines the amount of new reserves as 40 wells would be enough to deliver an equivalent
found (outer ring). If the prospect is a dry hole, the next level of predictive confidence. At the other end of the
spin determines success or failure on the next well. scale, for a company involved in high-risk (Pc = 10%),
ObViously, employing a spirmer is a visually satisfying large-reserve (lOO to 500 MMllOE), high-v<Hiance
iii
104 Management of Exploration Projects as Business Ventures
100
Z 80
~
U.'l
60
~
~ 40
~
~
20 )
~.'·.P90
U.'l
U
0 1------+------....-, -G:l; P10
Z -20
U.'l
~
U.'l -40 ,,_.,~__," I!I.'_''''M'''-'''''''''-''''._~'' Ii iI'-'·-"~- ~",.,.",.~~, I II
~
~
~
-60
... ......
~
.. ~. -' ' .• .J~~." ...,~ ..--
Cl
-80
~
-100
.. ~.
0 ,20 40 60 80 100
NUMBER OF WELLS IN PORTFOLIO
Figure 47 Predictive accuracy of portfolio performance improves with the number of wells in the portfolio.
prospects, an inventory of 250 wells or more might be The key point for management to realize is that any
required to provide adequate confidence in the forecast portfolio can be routinely analyzed by Monte Carlo
outcomes (Schuyler, 1989). simulation to provide various confidence levels associ
If the portfolio's size is inadequate to deliver the ated with corresponding reserves outcomes. Then
confidence in forecast outcomes that management management can decide how the portfolio should be
expects, at least six possible solutions exist: adjus~ed to provide the predictive confidence they
require. Obviously, if prospect parameters are biased,
1. Add more wells to the portfolio (which of course the portfolio will lose much of its effectiveness.
will increase exploration expenditures propor Fo~_companies operating in a variety of exploration
tionately); theCf~fs, especially those desiring a diversified, bal
2. Expand the portfolio by drilling morejomt-venture anced'r.6rtfolio, it is important to recognize that divi
wells (this increases sample size without signifi sions oF-erating in mature provinces may be expected
cant increase in total exploration expenditures); to provide prospects for the smaller-reserve, low-risk
3. Consider the portfolio over a mUlti-year period, end of the overall portfolio, whereas divisions exp!or
i.e., the forecast might cover a 3-year or 5-year ing iD frontier basins may provide ventures for the
prediction with running averages; high-potential, higher-risk part.
4. Modify the character of the portfolio by includ
ing more low-risk wells at the expense of some Principles of Exploration
high-risk wells (cf course this usually also entails
substantial reductions in reserves potential); Portfolio Management
5. Focus geotechnical exploration tools on high-risk It is far beyond the scope of this book to review the
prospects, to either: principles of modern portfolio management (Bern
• improve confidence in critical risk geologic stein, 1996; Markowitz, 1952). The key point here for
factors and raise Pc; or both explorationists and their managers to realize,
• condemn such prospects in favor of other, however, is that the same principles of risk-reward
more attractive ventures; or optimization apply to an exploration portfolio that
6. Management may revise its expe~!,~ti?ns for the apply to portfolios of various common stocks and to
level of confidence required in portfolio flii~casI5,~ other financial ventures.
i.e., they may accept a more realistic, increased level Exploration management can be provided with a
of risk consistent with real exploration variance. risk-reward diagram showing many pOSSible portfolios
Chapter 6 105
19\~X-
explu-r~ition"t"ton.!:rr..;:ts require that certain wells be
drilled regardless of how those wells compare with
other prospects in the portfolio.
x x X
X X
T
l\lost firms include such wells in the portfolio for pur
poses of forecasting portfolio performance and for
assessing staff performance in making geotechnical pre
x dictions. If they include such obligatory wells in the)
ranking process, it is only for purposes of general com1
....
106 Management of Exploration Projects as Business Ventures
Table 22 Uncertainty leads to common underperformance of exploration portfolios (after Horner, 1990).
prospects they wish to include, and then try to secure of return (DCFROR), and he realized that actual per
such a model portfolio as the year unfolds. ll1e result is formances of the individual constituent ventures
to diminish some of the theoretical advantage would vary, both up and down, from predicted n'ean
of portfolio selection. Another aspect of dynamic port performance because of the substantial uncertainty
folios is to preselect preferred participation levels in cer that attends exploration ventures (Horner, 1990). That
tain classes of prospects, or to employ 0\ VI to indicate is, some estimates would turn out to be too high and
the appropriate share for the company (pp. 55 and 92). some would be too low. To model what may actually
happen in nature if estimators are unbiased, he con
Timing Considerations structed a table comparing estimated rates-of-return
A second possible goal of portfolio management (ROR) with actual rates-of-return for a portfolio in
has to do with optimizing tilllillS of cash .flows. Ideally, which the economic cutoff was 15% ROR (Table 22).
projects should be timed so that excess cash from pro As Horner pointed out, the portfolio cutoff (vertical
duction revenues will be available when large de\'el line) must be made under conditions of uncertainty,
opment projects are expected to begin, and large-scale where prospects estimated as less than 15% ROR are
exploration projects are available when the company excluded. Because predictions vary from actual
has cash flows to use in increasing ,1sset value. Such results, however, some included ventures will
timing may require precision beyond our present lev inevitably underperform expectations, and some
els of geotechnical and predictive skill, however, and excluded ventures would have ou·tperformed expecta
some companies endorse a simpler procedure: rank tions. Thus the vertical cutoff is wrong, but inevitable,
the portfolio to maximize value, then if money is whereas the horizontal cutoff is correct but unattain
needed for development, it can be bOfluwed or able. So those ventures shown in bold type in the
derived by selling existing, less profitable properties. northeast and southwest quadrants of the diagram
An alternative approach to the dual corporate will be incorrectly dealt with. The consequence is that
needs of (1) cash flow and (2) growth could be to set the actual performance of the original portfolio will
up two noncompeting portfolios, one composed of inevitably be lower than the estimated performance
low-risk ventures that could generate needed cash filCII if cstillwtes are unbiased. The prevalent tendency of
flows near-term; and the second, composed of higher explorationists to overestimate prospect reserves
risk, large-potential projects that could provide (optimistic bias) has the effect of aggravating this
growth. Then the question is, What is the appropriate problem.
relative level of funding for the two portfolios? Although Horner indicated this to be a problem
without a practical solution, there may be at least a par
"Theory of Inevitable Disappointment" tial solution that can be described as the "pilot-fish"
Identification of this fascinating phenomenon is concept (pilot-fish are small fish that accompany
ascribed to Dr. Dennis Horner of Roval Dutch Shell. sharks, deriving their living off the crumbs and morsels
He re·cogriiL.'t"!:': that c.:>mpanies aSj':l:lb!~ ~)ortf~ios of not swallowed by the shark during feeding). If large
drilling prospects on the basis of some form of pre companies can identify small exploitation firms that
dicted economic ranking, such as il1\·estment effi may be willing to develop small or marginal fields, and
ciency (IE), ENPV, or even discounted cash-flow rtlte can prenegotiate basic deal structures with them, then a
Chapter 6 , 107
(:<;;)"ery that is recognized to be below standard rna\, unnecessarily long and expensive efforts and with
)'mmptly conveyed to the smaller, more efficie~t
-
tantalizing but ultimately fruitless results.
,\<'
'LtT1 Such business practices allow the large compa"!;}'
. "'
Negative Impact Is Inversely
;('hOld the development cost of a marginally prof
',i ilbJe field, and may allow it to recover some incremen Proportional to Firm Size
':-(J cost as a transfer payment from the small firm. Generally, the negatiw impacts of a bad play choice
are inversely proportional to company size. To very
large firms like Exxon-Mobil, Shell, or BP/ Amoco,
iVlanaging Exploration Plays which evaluat~ dozens of new plaY$ each year and)
Matching Play Attributes
enter perhaps' five or ten, a bad play can be shrugged.
off ana balanced against other successful plays, which
to Business Strategies
are expected to carry the cost of unsuccessful ventures
Successful exploration is the lifeblood of the inter so that the owrall annual new-play portfolio creates
"Ihoned oil and gas business. New fields must be substantial new value.
,md on a regular basis in order to replace the firm's To intermediate-sized firms, which may evaluate
';I<:'iidily depleting producing fields. But petroleum five to ten new plays each year and enter perhaps one
:!ccumulations usually occur in geologically related or two, a had play choice usually causes a long-lasting
'" milies, which we call plays, and modern petroleum reduction in the company's annual production stream
c\l'!oration is characteristically carried out at play and puts more pressure on the organization to find and
'lcalc rather than prospect-scale. Accordingly, the criti enter successful plays or even to purchase existing pro
,:i,'xploration business decision concerns which new ducing properties. Thus the consequences are inconve
"r'ly to enter, not which prospect to drill. It follows, nient or even serious, but ordinarilv can be tolerated.
+<':':::fcrc, that any dedicated modern oil and gas For small firms, which may eval~ate only a few new
',);lTany should have a strategy and process under plays each year and actually ~nter a new play every tv"o
i"eh it systematically and continuousl" identifies to fi"e years, the consequences of a bad play choice can be
lei Ci'i::>!uates candidat~ new plays. J
disastrous. As a result, smaller companies tend to choose
ne\\' plays in more established petroleum-producing
,;:;iness Consequences of Play Choices
areas, participating as minority partners, or choose sim
(mnpetent exploration play analysis and play selec ply to drill a series of independently submitted individ
, ,11' involve not only the synthesis of petroleum geo ual prospects in areas where they already have expertise,
:-:::"'1U" statistics, and economics but also require thus foregoing the efficiencies of regional exploration, as
:i'j;I.;idcration of (1) present and future business condi well a~ the likelihood of making large new discoveries
::e,';:; (local and international), and (2) business pat (see the discussion on creaming curves, p. 68).
: ,',Tto' (md requirements that are unique to the firm. Of course, the consequence of a successful new play
.; is clear that successful development of new is to add a new core producing area, which can then be
; ploration plays generates a steady supply of new expected to provide a significant and long-lasting
""] attractive prospects, which in turn lead to the suc incremental increase to the company's production rev
e"sive discoveries of profitable oil and gas fields to 'enues, BecauseI~.rge new fields are typically found
'1,1,',1 i, and even increase the company's reserve base, early in the exploration cycle, large firms have usually
I thel'C are consequences of participating in a bad placed a high premium on exploration in new trends
.-.. ;('{
and basins. Unfortunately, during the 1980s overall
industry participation in such high-risk, high-potential
l. poor return on invested time, staff, and capital new plays did not add to corporate net worih-it actu
an actual loss of capital);
(,-,[len ally destroyed value,
. 'J!i2 company may miss out on good profits from
(lj-, :llernative play in which it did /lot participate;
Learning to Make Money
l i ;(j Finding Smaller Fields
, '>rnpetilors benefit from a successful play in Incorporation of exploration statistics and the prin
'·our company's absence. ciples of exploration risk analysis into decision making
" has had a substantial, positive impact on the selection
f--c:,unsLlccessful plays, there are two polar end of new plays by modern petroleum companies. Espe
"(;'1/)C'(2O: every company should fervently wish that cially compelling are data indicating that, despite con
..e'., fCli!ure will occur quickly and unambiguously, tinuing development of superb new geotechnical tools
,'.:';i;I):; ,i"t'oscicntists and management content that a and concepts that have kept average exploration SllC
:,'.::(i':lg idea has been evaluated and disproved
--~. .
cess ratios constant at about 25% (Figure 22), annual
it'c!e",!y 2nd relatively cheaply. The alternative neg volumes discovered world\'vide have been decreasing
, ' [ c o n t C is correctly dreaded-the play that was since about 1965 (Figure 49a). Moreo,'er, the new
'; iy cho:;cn and improp,.erly negotiated, requiring fields being found by the international industry are
108 Management of Exploration Projects as Business Ventures
I\III)EAST
O ......SHORE IN (OENERAL
150
Q
Ii:
c::
w:.r.:
>0
c= 100 J
u::Q IlUR(;AN
en
5 i
o
~#~###~##~#######~~~,###
_~ FIEU, YEAR
Source: Petroconsultants
I
Figure 49b Discovery data, 1990-1999 (excluding U.S. and Canada).
I
II
steadily decreasing in size (Figure 1), so that the decreasing since the 1960s, successful oil cOJJ\panies I
chances of making a very large discovery ,ue getting must organize themselves and conduct their E&P
very small indeed (Figure 49b). The message is clear: business to make money by discovering and develop I
._,;,Given that 0) most of the world's possible petroleum
basi~ are increasingly well-known, and (2) the rate of
ing oil and gas fields in the 10-100 MMBOE categories.
They /Ill/st become more efficient ill ordl!r fo find n!uf pro I
I
discovery of super-giant fields ("elephants") has been duCt: s/llallerfields profifnbl!J.
Chapter 6 109
Fortunately, the new principles of exploration risk 4. For smaller companies that typically generate new
analysis (and especially play analysis) provide meth plays, promote them, and thereafter opera\e as
ods by which a firm's entire eAtilorai~Thl eff~J:t can be minority partners, certain types of plays may be
made more efficient, greatly reducing the likelihood of currently in or out of favor in the marketplace. For
selecting a bad play and providing a greatly improved example, it is currently very difficult to sell a play
basis for predicting reserves, chance, costs, and prof or prospect in the U.s. Gulf Coast unless it
itabilities of new ventures being considered for entry. involves wide 3-D seismic co\'erage.
5. Given that taxation of oil and gas production has
Special Business Requirements substantial impact on project profitability, Clnd )
May Dictate New Play Attributes that tax ramifications may apply to a firm's other '
Different companies have differing financial circum ventures, certain plays that offer special tax cir
stances and business constraints, These special condi cumstances may be preferable to other equally
tions may place special limitations on their require \'alid plays that do not. In the U.s., there has been
ments of play choices. However, it is important to dis an unfortunate tendency for firms to pay more
tinguish between short-term (one to three years) and attention to such "if-success" tax benefits than to
long-term (four to eight years) constraints, remember the geotechnical and economic merit of the p\ClY
ing that the time required to identif", evaluate, secure, itself. Workillg /"Ill'.?: First priority sllOlIld go to gl'O
prove, execute, and develop a new 'play is usually not teclllliclllly SOlllld, ecollollliclllly superior pillys; t!Jell
less than three years and often as much as six to eight cOllsidcrllllcillllry tnx bCllcfits.
years. Obviously, the astute exploration firm will not
allow. short-term constraints to pre\'ent it from takinab Concurrent Geotechnical, Economic, and
part 111 a play that is likely to be entirely compatible Business Evaluations of New Plays
with the business circumstances expected to exist In order to achieve efficient, thorough, and objecti\'(~
when the play actually comes to fruition. assessments of new exploration plays, many geoscien
Nevertheless, here are some special business condi tists have recognized the need to adopt a consistent pro
tions that may influence new play choices: cedure that uses either a manual flow-sheet or
computer software to perform all the geotechnical tasks
l. A company with legal contracts that require it to and attendant calculations needed to carry out a risk
deliver speCified volumes of natural gas to cus analysis of a new exploration play.
tomers may attach higher priorities to gas plays HOI-\'e\'er, a common and unfortunate industry pat
capable of increasing deliverable gas in the short tern ha~. been to conduct the geotechnical evaluation
term, especially if the company fears a possible first, and perform an economic assessment only at the
shortfall. end of the investigation. This often leads to poor
2. A comp'any with a limited "exp"loration budget and/ or delaved business decisions. A much better
may give preference to plays located near pro approach is {or the geotechnical, economic, and busi
ducing infrastructure, or to those whose primary ness assessment to proceed simultaneously, with fre
objectives are rela tively shallow. Alternatively, quent and detailed integration of such information
such a company may choose to invest in special among team members. This approach results in more
ized professional staff and state-of-the-art geo reliable estimates and more innovative business plans.
technical tools and methodologies, which can be It is also more efficient because proposed plays that
eApected to develop new' play concepts that can have severe economic or business limitations may be
then be leveraged into substantial shares of mul detected early, and expensive but unpromising geo
tiple new-play ventures. The problem with the technical work may be curtailed, allowing scarce man
latter approach, however, is the substantial time power and a limited budget to be used in more promislllg
required to build such staff and generate the nec projects elsewhere. An excellent general summary of
essary new concepts. long-range exploration planning is provided by Megill
3. Some companies may be constrained from operat (1985).
~ng in certain countries or geographies, either by Specific business circumstances that should be
mternal organizational restrictions or limitations explored and assessed include:
?r by external circumstances. Similarly, any lim
Ited geotechnical skills of the professional staff Product Transportation
~ay prevent consideration of some plays. Some Are there pipelines in the areas? Do they have avail
tImes a previous negative experience by senior able capacity? Will their owners allow your product
management in a given location or in a play of a to be shipped in the areas? Are there transportation
_r: . .
-.' c.er!~T1 type may prejudice them against objec tariffs? Are there calendar restrictions? If a pipeline
tIvely considering an otherwise valid play venture. must be constructed, what is the lead time? cost of
110 Management of Exploration Projects as Business Ventures
•
construction? minimum new reserves to justify experienced international petroleum experts, assess the
pipeline construction? Can you make additional prof degree to which the real infrastructure of the host COun- •
its by shipping gas for other producers as well as your try (not just politicians) really wants the firm to partici
own? What is the cost and capacity for temporary pate in the long-term economic development of their
transportation (truck or ship)? hazards? country, as opposed to just wanting the firm's money.
• a perceptioll of excessive profits by the company, in the selection process. The most important criterion,
which production potential of the contract area however, is to estimate the likely profitability ~f the
turned out to l:ir1nuc1'r~w~a.t.~than originally play venture, given discovery. Other criteria then
The problem in choosing new plays is that many sig largest sources of inconsistent corporate decision mak
nificant criteria must be assigned in order to optimize ing (pp. 55 and 93).
112 Management of Exploration Projects as Business Ventures
•
Designing Exploration Campaigns
The most practical publications dealing with the
design of exploration campaigns-how to plan and
carry out successful play exploration-have been writ
ten by Megill (1985), Downey (1992), and St. John
(1992).
LEVEL
Well-planned plays save money by identifying
unsuccessful ventures quickly and efficiently. They 6
make money by identifying potentially successful 1
plays with high probability, optimum participation,
and minimal investments, before success can be
claimed with certainty. SI/ccessjitl piny exploration repre 5
If the play appears clearly to be inferior to other are, in fact, estimates made under varying degrees of
ventures being considered, it should be filed for later uncertainty. This places a heavy professional burden
"'-fevi~ (on~.Jo two years). If the play appears to be on the corporate technical staff to consistently gener
marginal, a brief analysis should be carried out to indi ate responsible, unbiased estimates, and on explO
cate those factors that could be improved and the ration decisionmakers to utilize such estimates wisely
likely degree of improvement. .' and consistentlv. When technical estimates are overly
Superior plays should be selected from the play optimistic, the firm is encouraged to invest in inferior
inventory to form the annual play portfolio. This selec projects. Owrly consen'ative estimates discourage
tion should give strong weight to the risked invest the firm from realizing the full profit potential of I
ment efficiency of the constituent projects, but it is also underestimated projects. Thus either error has the>
appropriate to consider other criteria as well. Required potential to cause loss; either error reduces staff credi
investments will be apparent at this stage. bility. ''''hen company decisionmakers do not know
Once the play portfolio is selected, budget alloca the relati\'e reliability of technical estimates, they are
tions must be made. These allocations should be encouraged to rely on their intuition (which usually
planned prm'isionally O\'er at least a five-year period, causes great inconsistency) and to improperly use
covering several possible outcomes. Decision points expensh'e geotechnical data bearing directly on pro
should be clearly denoted in the five-year budget plan. ject risk \'ersus reward.
following contingent developments. . The onl~' way for corporate decisionmakers to
The final step in the planning process involves man improve the o\'erall exploration performance of the
power planning, which has three essential aspects: firm is to monitor and preserve, on a systematic and
routine basis, the technical and economic predictions
1. What skills are likely to be needed, for what time made by their professional staff, and to compare them
periods, and where? against clctual outcomes. Without such ongoing
2. Who iHe the existing staff members who repre calibrations, technical and economic forecasts are
--. sent the required expertise? What vacancies must analogous to one-way rockets launched into outer
be filled by recruitment, and when? space-there is no feedback! Without comparisons,
3. Construct a provisional plan for staff succession exploration companies continue to make the same
and promotion. mistakes year after year, and performance of neither
". staff nor decisionmakers can be measured properly;
The final step' involves manacrement's
o decision to carn'. accountability is thus greatly diminished. What gets
out the new play, with early priority given to both advan lllca~ltrl'd, gets dOl/e.
tageous negotiating tactics and the formation of a play
team based on the previous manpower considerations. Technical Performance versus
Economic Performance
"\ Overview of Successful Play 'M~nagement
It is important to recognize that positive results
Successful, cost-effective play development is a con from excellent technical predictions can be nullified by
~nuously evolving, multidisciplinary staff effort that inept economic forecasting, and vice versa. Technical
r:guires dedicated professional skill, unselfish coopera staff should not b~ penalized for the poor performance
tIon, thorough communication, objective evaluations, of the economic staff and managers. Accordingly, for
and courageous execution. Properly organized and both plays ill1d prospects, I\ovo different aspects of per
man"ged profitabIlity, play selection, and development formance should be monitored. Technical perfor
. provide---on a regular basis-new prospects to populate mance compares all geotechnical, engineering, and
'"-'. the firm's annual drilling portfolio, and therefore new cost/price forecasts with actual outc0mes. Some
discoveries of economic reserves to replace produced examples are reserves; chance of completion; initial
reserves. Orderly new play development is impeded by production rate; decline percentage; drilling, comple
frequent reorganizations or personnel instability. Prop tion, and operation costs; and ,,,,ellhead prices. Eco
erly executed, the exploration play is the ultimate form nomic performance measures forecasts of project
of leveraged geotechnical risk venture. But an essential profitability in relation to actual profitability; both
aspect of successful play generation and analysis is forecasts and actuals must be related to preselected
another ongoing process-the rigorous assessment of corporate standards. Some useful economic measures
geotechnical as well as economic performance. That is adaptable as success criteria include predicted project
the subject of the final section of this chapter. ROR in relation to actual ROR and to corporate hurdle
rates; projected venture NPV and IE compared with
Assessment of Exploration Performance actual NPV and IE; actual early-term investment costs
compared with predicted investment costs; projected
'. Mo;~f the critical parameters that influence valu exploration cost of finding (COF, expressed in $/BOE)
a tion and ranking of explo~a tion plays and prospects compared with actual COF and established corporate
II
114 Management of Exploration Projects as Business Ventures
COF goals. Making economic comparisons on the Criteria Indicating Competence in Play Analysis
basis of actual monetary value (rather than present
Because indirect and subjective criteria must be
monetary value) may facilitate the process. Also, cost
used to assess the technical performance of play
of-finding resul ts should be normalized by distin
teams, the writer has developed 10 criteria that help to
guishing between increased drilling efficiencies,
indicate the relative skill level with which such teams
stimulation and completion efficiencies, and explo are functioning in the critically important activity of
ration skills. Finally, try to separate economic influ generating and evaluating new exploration plays. He
ences that lie within the purview of professional staff has used these criteria for more than eight years in
and managers from external influences over which
revieWing play development te<!.ms for many dilferent
technical and management staff had no control. inteqlational exploratiOl~ (Ompanies.
For all prospects, scenarios outlining various techni J
I
cal and economic outcomes alld their collseqlll?llccs 1. SOllie tl'alll lIIelllbers have previollsllf beell clOSe/If
should accompany any project recommendations. For associated with discoveries. Successft;1 past experi
possible economic events exfl'l'Ilal to the project (world ence brings an essential sense of scale and scope
price fluctuations, international political de\'c!8p to the geotechnical process, especially to keeF'
ments, new technologies, etc.), project planners should the effort focused on practical (rather than scien
also outline scenarios, impacts, and possible responses. tific) problems. Play analysis is best carried out
by seasoned, capable professional geoscientists
Measuring Performance: Plays versus Prospects and engineers.
Papers by Rose (1987)/ Clapp and Stibolt (1991)/ 2. ROlltille lise is made Ofthe petrolelllll systelll approacll.
Capen (1992)/ Otis and Schneidermann (1997)/ Alexan All geotechnical aspects of hydrocarbon occur
der and Lohr (1998)/ Johns et al. (1998)/ McMaster and rences are routinely integrated, with special
Carragher (1998)/ and McMaster (1998) presented dif emphasis given to understanding the HC kitchen
ferent but complementary methods for evaluating and (see Appendix E). Integration of petroleum sys
expressing staff performance in geotechnical predic tem thinking with regional tectonic expertise and
tions related primarily to prospects rather than plays. broad knowledge of basin classification provides
For many corporations that participate in 20 to 100 additional creative insight and power.
exploratory prospects each year, such methods pro 3. ROlltil/e lise is made of geostatistics alld databases.
vide an acceptable basis for assessing and improving Thoughtful, thorough acquisition, organization,
geotechnical predictive performance, either on an and regular analysis of critical and current sta
annual or multi-year, mo\'ing-average basis. tistical data is necessary, such as:
Assessing predictive performance for plays is much • field production catalogs;
more difficult than for prospects. Variance (uncertainty) • routine construction of FSDs, including FSD
within plays is generally much greater, and the number shift studies (see pp. 71-75);
of predictions at play scale are far fewer than predictions • success-rate studies in analog trends, includ
at prospect-scale. For plays, elapsed time between tech ing show-holes, economic discoveries, and
nical predictions and measurable' results is commonly dry-hole causes; and
four to eight years; economic forecasts commonly • field number and density studies (including
require six to 12 years before useful comparisons can be creaming curve data) for analog trends.
made. Normal personnel changes o\'er such timeframes 4. StaI!has easy access to cOlllprehellsive petroleulll refcr
reduce staff continuity and presen'ed histories of pro mct's. Staff must have ready access to the intel:na
jects. Fear of criticism (on the parts of both teclu1ical staff tionalliterature on petroleum exploration,
and managers) leads to incomplete presen'ation of key production engineering, economics, finance, and
records and reports. Changes in technical and economic Ic.';'v·~ lvlaIlj Jclla d.l~ j~r~~, in th~:- puhlic dO!i!...ti~~.
definitions produce inconsistent data sets. Proprietary reports and databases provide addi
Moreover, there is great organizational pressure for tional information. Successful play analysis
early and correct assessments-for example, that an requires efficient sifting through of copious quan
emerging play is mediocre, or that it has great poten tities of data. A competent technical librarian is a
tial. Understandably, companies do not want to wait ,·aluable member of any play-generating group.
four to eight years to decide that a given play is not 5. A cOllsistCllt process is rOlltillely used forforl/lal play
worthwhile. Opposed to this pressure is another truth: III IIllys is. Rout.ne use is made of a simple, consiy
Beware the prematllre exit, tile abal/dOJlmel/t ofa promisil/g tent flow sheet or of software to estimate chance
play after ollly one dry hole! Naturally, the key issue here of economic play success and reserves potential.
is to focus on what was learned from that dry hole 6. There is demollstrated colltilWOllS illtegratioll of pro
does it apply throughout the play area (as a shared pisiollal ecollolllic criteria alld busilless collsidera
attribute) or does it apply only to the specific prospect tiolls with tlle elllergillg geotechllical pictllre. It is
(as an independent factor)? important that each play group have a modicum
Chapter 6 115
of economic and business expertise, or ha\'e routine 3. the acceptablL' k'\'el of irreducible uncertainty
access to (and interactions with) such expertise, and risk, gi"en pr(lject costs and poteMial
.... .,._7. All opell, dyllnmic piny illvelltory is /1/ailltnil/cd. A rewards.
continuously changing inventory of new and
..... existing plays should be ranked consistently by Play Analysis: Organizational
investment efficiency and show status of the
~.
Patterns and Principles
projects. Appendix G is an example. The con
sti tuents of the inventory should include plays Common Patterns and Procedures
having reserves potential that is appropria!..e to .J The writer has obsel'\'ed and noted several common /
the company's needs, in diverse gf'ologk set- .',.' biases, patterns, and useful procedures of corporate ~
tings, and of an adequate numb!!r to ensure a\ risk analysis that are practiced by capable explo
- satisfactory annual flow of new trends into the'
firm's exploratory effort.
rationists around the I\'orld:
8. Techllical work is ol'"allized IIlId orderl1/. There are 1. lVlost geoscientists with mediulll to large CL))ll
visible planning calendars and evid~nce of: panies tend to ()\'erestimate the reserVeS sizes of
• efficient and timely wf)rk performance; future disco\'L'ries, especially in onshore plays.
• accurate forecasts of costs; 2. Conn?rseh', thl'\' tend to underestimate the num
• geotechnical costs that are competitive with ber nf futt;re fieids, especially in onshore plays.
industry standards; and 3. Plllets for nl'\\' plays commonly lies betwel'n
• systematic preservation of regional geologic P30'~;) and P70"". As J>lllefs rises above 50-60'\"
compilations, in a usable format. the chance of economic play failure incre,lses
..... 9. Syste/1/atic criteria are used for measllrillg the aCClIrtlcy rapidly .
of tec/1IJical foreCII:::ts. There is an ongoing, open 4. If the required geotechnical data arc available
process for recording forecasts of emerging data and organized (see pp. 65 and 84 and Appendix
focus in the areas and comparing them with actual E), a discriminating play analysis can be carried
results, as well as periodic revie\-v of predictive out in one or two work days.
performance and sharing of lessons learned. 5. ~.'Iany of the critical supp~)rting dahl on explo
10. ROlltille group procl'dllres arc used for peer reUiell'::: ration plays are in the public domain and there
al/d the gelll'mtilJll of llrw idea:::. There is an open, fore are quite inexpensive-usc them!
". interactive working atmosphere where profes 6. Reasonable estimates of future field numbers
sionals informalJy review each other's emerging can be derived from analog producing areas
\\Iork and stimulate new ideas through routine ai1d \'erified using discovery process modeling
group interactions. or other pragmatic methods (pp. 69-70).
7. Intelligent manipulation and utilization of FSDs
Attributes of Good Play Managers (combined with estimates of field numbers)
Good managers of the process of exploration play allow useful forecasting of play reserves.
generation and analysis Clre by nature generalists <md 8. Items 4 through 7 demonstrate the importance
integrators. They must be able to motivate individuals of having, in,e\'ery play-analysis organization,
", ~f diverse technical expertise and personality types, to qualified technical staff-especially a librarian
"
~lsten and communicate effectively, and to keep pro and statistical technician-to assist in data
\ Jects on schedule and within budget. They must strike a acquisition, organization, and'analysis.
working balance between accepting large, irreducible 9. Full-cycle economic analysis of exploration
uncertainties and using expensive state-of-the-art tech plays aJJows discriminating business decisions
nologies to reduce risk where they are cost-effective. to be made early in the exploration cycle.
Good play managers must be willing to make decisions 10. Play analysis does not provide perfect ,1nswers,
on imperfect information, being cognizant that some of but it will prevent serious mistakes at either
their decisions may turn out to be wrong. They must extreme, both the positive and the negative, and
understand clearly that early rejection of a considered that level of precision is usually sufficient.
play for technical and/or l'conomic reasons docs not 11. The play only works where all the genlogic
constitute failure. Such managers know that mOlley chance factors coincide. Your job as play analyst
can be made in two ways: by finding profitable new is to delineate the Mea of coincidence (and pns
plays and by staying out of unprofitable ones. Such sibl)' the more apparent prospects in it). Make
milnagcrs mu<;t hil\'(' elll inherent sense of: overlay maps showing degrees of confidence in
the various clements of g~~lo~ic chance.
1. the known fact,.; thai are critical' 12. Pay very close attention to the kl:~ ..en-"-';;7ld rec
2. til<! unkno\'1I1l fllcts thelt need t~) be determined; ognize that "shows arc the footprints of migrat
and
ing oil."
i
13. Using the petroleum system concept, think criti 12. The expected value concept is alwilys a useful
cillly ilnd imilginiltively ilbout potentiill reser yilrdstick for consistency, When EY is positive, •
voir/seill couplets to help recognize potentiill you're investing-when it's negiltivc, you're
new plilYs, gilmbIing. Invest!
14. Identify ilnd focus on the criticill diltil thilt will 13. There Me two simultill1l'ouS ilnd interilctive
prove or disprove the plilY. eVilluCltions: geotechnical ilnd economic-don't
15. "The eildy bird gcts the worm"-usuillly, the ignore either one.
lilI'ger fields in ,1 new plilY ilre found eilrly in the 14. In competition, the Winner's Curse illwilYS
explomtion cycle. .looms. Know how to ilvoid it-ilbove all, avoid ,,
16. PlilY exploriltion requires routine ilppliciltion of . blind competitiveness. .
geoteehnicill ilnillogs, /Jilt /7(' slIr(' YOllr nllnlog is i? It is not as if you know nothing-you lIsuilJly
llfllid! know more thiln you think you do!
17. Try to nurture creiltive ilnd/or unconventionill
idc<ls in the filee of contmry opinioIls, especiillJy
in older trends.
The "Prospector Myth" It is also not surprising that most explorationists are
versus Systematic Exploration: overly optimistic about their basins, trends, and
prospects. After all, such dedicated prospectors could
Dealing with the Dilemma hardly be expected to be rigorously objective about
Introduction their prospects! The Prospector Myth is the primary
reason why explorationists persistently overestimate
The Prospector Myth is the petroleum e'Cplorationist's the reserves potential of their prospects.
version of the Hero Journey. We are informed and But the process by which oil and gas prospects are
inspired by the image of the courageous lone prospec translated into economic ventures also contributes to
tor who struggles against Mother Nature, financial overoptimism. In the early days of petroleum explO
hardships, skeptical associates, and repeated rejection ration-the glory days-many prospectors were
by investors, before finally succeeding through persis indeed individuals or small firms, and their investors
tence, faith, and luck, to achieve vindication, wealth, were private third-party investors and corporations,
and fame. Most of us know personally-or know of more or less knowledgeable about petroleum explO
one or more such individuals. We call them "wildcat ration. en-ileal emptor was the prevailing ethic because
tel's," "oil finders," "visionaries," and other dramatic the subscribing sponsors were expected to be able to
names reflecting the respect they are accorded in our judge the true merit of each deal. So .the operative cri
industry. terion for success was to sell the deal, to get the well
drilled. And that pattern still today characterizes
Uncertainty, Intuition,
many smaller firms that generate prospects and turn
and Overoptimism"
them over to larger companies to be drilled. "Have
faith-success will evel1tuate, given persistellce mzd suffi
Geoscientists select (or are assigned) basins or cient trials: olle real success will carry II dozell failures."
trends in which to explore for petroleum. Such Unfortunately, the same value system still operates in
endeavors are characterized by daunting uncertainty, some offices-even thrQughout entire companies
which can be reduced (but not eliminated) through even though the professional employees of these pub
costly geotechnology and seasoned judgment. Explo licly owned corporations are ~elling their deals to their
. rationists must peer through Nature's fuzzy lenses, own managements (and thus their own stockholders).
searching for cryptic clues. They must invest intense Because of our sympathy for the Prospector Myth, we
physical and intellectual energy, over extended peri tend to dismiss more benign examples of this as moti
ods, patching together possible portraits of the subsur vational bias. More flagrant examples, however, we
face, then selling and defending the proposed risk are bound to label as conflicts of interest.
ventures that arise from their imagination £Ind labor.
Because exploration is dominated by subjectivity and
uncertainty, it invites the exercise of intuition. And, of Times Change
COurse, when geotechnical intuition is rewilrded by The increasing employment of geoscientists and
exploration success, the prospector's ego is affirmed the rise of large, publicly held corporations after
and eVen C'xtolled. But most exploration projects f£lil, World War II gradually began putting a crimp in the
so most seasoned prospectors have learned to live freewheeling presen tation of oil and gas prospects.
with repeated failure. It is no wonder that petroleum Scientific objectivity, professional ethics, the declining
. ~
exploration cherishes the> Prospector Myth. petroleum resource base, and the need to deliver
II
117
118 Petroleum Industry Practices of Exploration Risk Analysis
• F~EQUENCY
_ J
1%
1\1
/-"1
&. 10%
- ·V
• CUMULATIVE :2~
PROBABILITY lK
'"'
~
:J
:J
-
- VI-
f
I 5()%
')()%
')')%
10K lOOK 1M 10M 100M IB
Reserves (Log Scale)
(in this figure, K = thousand, M = million, B = billion)
promised exploration performance together collided from which they select their annual dr'illing portfolio,
with the Prospector lvlyth and salesn:anship. E\'en so, which comprises only those prospects that together
influenced by the Prospector Myth, the world maximize economic value consistent \vith company
petroleum industry wasted a lot of money on explo goals and risk tolerance. We try to manilge exploration
ration in the late 1970s to early 1990s. Shell, Mobil, by managing the exploration portfolio, With this con
and Amoco independently reported that exploration cept comes
.
the realization that, if portfolio mana"'e
0
for high-risk, high-reserves targets in this period ment is to succeed, each prospect must be assessed
destroyed value, rather than creating it. While we consistently and objectively. The inherent uncertain
kept looking for elephants, sizes of discO\-eries were ties can be dealt with via improved geotechnolo o.o"\' and
steadily diminishing (Figure i). We lost credibility with geostatistics. What kills the portfolio is bias,
with directors; we lost money for stockholders. By the which overvalues some prospects so that the \'alue of
1990s the industry was becoming global in scope and the portfolio is not optimized (Figures 15 and 16). The
needed to become much more efficient. To be sure, stockholder is shortchanged by the Prospector Myth.
improvements in seismic technology and drilling and Systematic portfolio management is more eHective
completion methods improved our success ratios sub than intuitive prospect selection-or"cherry-picking."
stantially. But consistent, objective, technically sound 111is is a blow to the egos of most prospectors, as well
procedures for assessing prospects also led to the as many managers.
adoption of systematic prospect risk analysis proce
dures. The recognition of the lognormal distribution
as the prevailing natural pattern of oil and gas field Imperfect Remedies
reserves distribution (Figure 51), together with the Since the 1950s, our industry has tried to reconcile
development of refined methods of estimating geo the dilemma in various ways. A common approach
technical uncertainties, began to constrain the opti thankfully now diminishing-pitted geoscientists
mistic exuberances that had been indulged by against engineers, tacitly accepting and reinforcing the
intuition and the Prospector Myth. proposition that geologists were expected to be OH?rly
Today, most modern petroleum exploration is car optimistic, which required engineers to be corre
ried out by multidisciplinary geotechnical teams, not spondingly overly conservative. Another remedy was
individual prospectors. Most substantial companies to artificially inflate economic criteria-r"tur,,\.J:·'. -'he
consider an inventory of many candidate prospects, discount rilte-under the mistaken notion that "those
Chapter 7 119
SMALL FIELD
-'
THERE'S A SMALL
,... ~.
CHANCE OF A VERY
Arithmetic Scale
LARGE FIELD
PROSPECT RESERVES
Figure 52 Exploration reality.
prospects that still have positive expected monetary Response of Systematic Exploration
values (EMVs) under inflated discount rates must be Companies eventually began to employ institution
better prospects than those that don't." A third tech alized systematic procedures for continuous improve
'.
nique was to employ hidden hurdles (p. 123) in the ment, ,·vhich required objective comparison of
higher levels of the decision chain: managers at head geotechnical predictions versus actual outcomes. Geo- .
quarters routinely cut prospect values by half (or scientists began to learn from their mistakes and to cal
more), based on their observations of past exploration ibrate their predictions. Corporate explorers began to
'. overoptimism. Tn response, explorationists became employ a different criterion for success: adding mIlle
adept at sniffing out such arbitrarv, screeninO'0 measures versus gettillg ti,e prospect drilled. The need for objectiv
and devising ingenious ways to generate geotechnical ity generated a long-overdue appeal for geotechnical
numbers that were adequate to get their prospects professionalism as a requisite to objectively identify
drilled. Tn particular, many explorationists in top man ing and selecting those projects that together maxi
agement exercised their own version of the Prospector mize the value of the portfolio, consistent with the
Myth by applying their privileged intuition-and organization's strategies and risk tolerance. Geoscien
egos-to the prospect selection process. Understand tists could take pride in being professional.
\ ably, such managers found it difficult to surrender the Today, some prospectors mourn the diminished
int~itive style that had advanced their successful explo influence of the Prospector Myth, even though the log
r~hon careers to date, in favor of probabilistic expres normal distribution still allows them to dream (Figure
slon~ and reliance on systematic prospect selection. 52). But increasingly, astute explorationists are recog
Smce almost no one kept systematic records docu nizing the.necessity for heightened exploration effi
menting actual pr.ospect results (compared against ciency as the global resource base continues to shrink.
geotechnical predictions), everyone in the chain Systematic risk analysis, professional objectivity, and
prospectors, engineers, local m~nagers, and senior performance tracking must go hand-in-hand with
executives-usually did not have to confront the con sophisticated geotechnical methods.
sequences of systematic biCls. The urgency of drilling But the Prospector Myth still lingers, and properly so
. the next well far outweighed the valu~ of ubjectively if it can inspire our courage, persistence, and imagina
. and purposefully assessing our mistakes from the tion in petroleum exploration, without biasing portfolio
~ast well. We were too busy drownillg to take ti1l1e to selection. All too often, however, the entrenched intu
11I1prOVe 0/11' swill/millg allility. Instead we put our itive preferences of management as well as some geosci
greater reliance on geotechnology (especially seismic entists-and inappropriate concern about preserving
surveys!, \~hich often did reduce the impact of lilrge budget share-warp exploration performance through
uncertall1hes regarding prospect reserves and pres bias or double-risking. The ongoing challenge for geo
ent value and the discouragingly low chances technical professionals is to harness the energy of the
of prospect economic success. But most of us stead Prospector Myth without compromising the scientific
fastly refused to address the glaring problem of bias -C"int'eb:"'Y ar;,.::i-:.JUsiness objectivity now required for
(Figures 15 and 16). successful management of the exploration portfolio.
120 Petroleum Industry Practices of Exploration Risk Analysis
..
Characteristic Corporate Process Accepted economic measures, such as net present
for Exploration Risk Analysis value (calculated at the corporate discount rate), dis
counted cash-flow rate of return, investrnent effi
Most organizations have settled on using a consis ciency, and growth rate of return, are calculated for
tent process to assess all exploration prospects and each reserves case.
plays for management evaluation and decision. Proce The next step is to estimate the chances of success uti
dures, underlying concepts and principles, and helpful lizing a consistent set of well-defined geologic ch~nce
suggestions are carefully organized into handbooks or factors. Again, flow sheets or software takes the explo
manuals, which in turn relate to and explain computer rationigt(s) through the estimation of geologic ch;)l1Ce in
programs or flow sheets by which geotechnical staff in a step-by-step manner, with useful questions, reality
all office locations can readily, systematically, and con check!s, and questioned errors appearing throughout th~
sistently carry out risk analysis of all exploration ven process, thus encouraging iteration and reiteration of
tures. Some companies have developed their own risk estimated values before settling on final forecasts.
analysis software, using various forms of Lotus® or Once sound estimates of prospect reserves and
Excel® spreadsheets and risking software such as chance have been made, the next step is the generation
"@Risk"® or. "Crystal Ball."® Other companies have of risked prospect parameters, such as expected net
licensed customized software from several different present value, risked investment efficiency, and
vendors or consulting groups. A useful byproduct of prospect optimum working interest, by which prospects
such standardized approaches is that all geotechnical under consideration may be ranked for possible inclu
predictions and forecasts are preserved, thus facilitating sion in the company's current prospect portfolio. These
subsequent project review for purposes of performance values are used to help determine appropriate bid levels
analysis. Such data are rolled up and compiled by cen for acquisition, keyed to the anticipated methods for
tral staff to facilitate portfolio selection and analysis and offering and sale. .
to analyze overall company performance. Ordinarily, portfolio selection is carried out as a sepa
Prospects are identified by name, location, trend, rate computerized procedure that may employ the effi
reservoir,objectives, originator, and the like. Then cient frontier concept, using preselected corporate values
interactive risk analysis software guides the explo for acceptable portfolio risk (variance of economic
rationist(s) through the process of developing the results) versus portfolio rate of return. Some portfolio
prospect-reserves distribution, honoring the lognor management systems consider and balance growth with
mal distribution and expressing projected prospect timing of cash flows. Probabilistic expression of portfolio
ultimate recovery probabilistically, as PlO%, PSO%, results allows senior management to assess and evalua te
P90%, and mean reserves. Many firms have incorpo various portfolio tradeoffs before settling on a final selec
rated interactive graphics capabilities that allow the tion of prospects to constitute the final portfolio. HOil'
reserves distribu tions to be depicted in either fre ever, it Cnll/lOt be ovcrell/phasized that /IIeal/il/gflll portfolio
quency and/ or cumulative probability format and to al/alysis absolutely depl!l1ds 01/ respol/sible al/d III/biased esti
be shifted at will. Reality checks help common errors mates ofprospect resenJes al/d the chal/ce of sllccess.
to be identified throughout the estimating process. Another distinct computer module or standardized
Many programs contain analog field data, providing a activity is ordinarily devoted to performance analysis.
per~rective <!~;jinst which the prospect can be viewed. Predictions of prospect parameters are routinely ent~red
The software promotes trial fitting and testing of data and p·reserved, keyed to the parent play, location, objec
and early estimates, resulting in repeated iterations tive formation, and prospect name. Results of
and reiterations of data until best-fit distributions exploratory drilling are then entered into the inventorv
emerge and are finalized. to be used in compiling and analYZing the predicti\'~
In a detached procedure, cash-How models utilizing capabilities of the geotechnical staff and the overall
the differing probabilistic reserves level (and ancillary exploration performance. Frequent feedback and discus
well numbers, initial production rates and percentage sion of specific learnings and remedies between portfo
decline curves, costs, and contract terms), are used to lio management and exploration teams allows steadv
calculate the present value of different reserves out improvement of professional geoteclmical performanc~.
comes. Uncertainties such as initial production rates,
percentage decline rates, and wellhead prices are vari P;ay Analysis
ables within the cash-flow models. Based on these sev
eral reserves cases, (PIO%, PSO%, P90%, and the Some exploration companies have also formalized
mean), minimum commercial and ecol/omic field sizes procedures for conducting risk analysis and economic
are derived. These figures are used in connection with evaluations of new exploration plays, alo:l.g comparable
the prospect-reserves distribution to estimate the lines to the procedures outlined in Chapter 6's section
chances of rr:lnimum commercial and minimum eco on Play Analysis. Many firms have soitware or flow
nomic field sizes (given a discovery). sheets that systematize and standardize the process of
Chapter 7 121
exploration play analysis. A limited but useful literature 2. Exploration staff must thoroughly understand
outlines many of the common concepts, procedures, risk analysis concepts and their prope~ geo
and pitfalls in risk analysis of exploration plays: Baker logie/economic applications, and accept respon
(1988); Baker, Gehman, James, and White (1986); Brown sibility for unbiased predictive performance;
and Rose (2000); Jones and Smith (1983); Rose (1995, 3. There must be a consistent system by which risk
1996a, and 1996b); and White (1988, 1992, and 1993). analysis is carried out on all prospects and plays
considered by the company;
Implementation of Risk Analysis 4. There must be a centrally coordina ted prospect
ilwentory leading to objective ~election of t!~~ /
in Exploratcon Organizations annual drilling portfolio, whieh is keyed to meet ~
Introduction corporate goals; and I
optimum working interest for each included venture. should not be attended by manag(~rs who have direct
Alternatively, the ventures «111 be chosen tll optimize responsibility for salClry administration and promo
management's declared preference regarding risk ver tions, because their presence tends to impede honl'st
sus reward, folillwing the efficient frontier concept. In inqu iry ilnd discussion of perform'lnce. A postd ri II
this approach, risk a\'{~rsion considerations apply not review should take no more than one day. It should be
to each project but to the overall portfolio. In either facilitated by the 10cClI champion, who shOUld prepMe
method the process generall~' results in capital alloca a brief written sumlllilrY with lessons learned and COm
tion based on nwrit, consistent with the firm's risk parisonof all significant predrill prl'dictions and post
preferences.
Depending on the numbe.r of vehtures in the portfo
drill outcomes. Each tCilm member should receive a
copy. Copies should ~o to a very few concerned miln r
lio, their rescn'es varianct.i, and the il\'crilge chance of
success, it is possible to pred ict the probilbi Iistic resul ts
agement and high-level staff. NOTE: It is essential that
rccllI"dcd predrill predictions must be those on which
1
of drilling the portfolio. Naturally, the more wells in the the actu<ll decision to drill was based.
portfolio, the more accurate wi II be its prcd ictive power.
Performance Tracking and Communications
Most companies do not include plays and prospects
in the same portfolio. Results of ex pllll"l tory wells and of postl1rill reviews
are cOlllpiled, analyzed, and reported to top manilge
Portfolio Scrutiny by Educated Management
ment and to exploration mil nagel's and geotechnicill
The key to sllCcess in corporate risk analysis lies in stilff with the intent of identifying bias, recurring errors,
support and utilization by knowledgeable managers, and recommending remedies, and of recognizing note
often supported by high-Inel staff. To be effective, worthy and objective prospect performance.
management must understand the basis for risk cmaly
sis, accept its indications of appropriate wnstraints, and Prc,fessional Pride in Objective Estimating
use appropriate economic measures ((1nsistently to Companies that benefit by effective operationill risk
assign values to the constituent \'l'ntures. By reviewing analysis typiCillly undergo il change in some of their
portfolio rl'sults freLluently, managers (,1:1 reward geo orgilniz.,tional ethos, in which geotechnicill stilff recol>
C>
technical stuff whose predictiL)(1s cne unbiased and nize and accept the need for reasonable ilnd unbiased
objective, as well as correct those explorL1:illnists \\'hose estimates 'lnd develop professional pride in prOViding
forecasts are O\'l'r1y optimistic Lll' ll\'erly (,'l1servilti\·e. them. They begin to focus on adding value-milking
money for tilt' firm-rather thiln on just drilling wells.
Honoring Nature's Envelopes This results in impro\'ed corporate perfl'rmance and
Geoteehnic'll predictions are made b:: professional overall strength, with the stockholder being the chief
geoscientists and engineers \\'ho have I~'arned how to beneficia IT.
improve their estimating techniques. Thc~' employ the
expectation of Jognl.lrlllality, known ral1~es of geologic
parameters, cll1d re<llity c1wcks such as ficki-size distri
Persistent Problems
butions, edited success ratios. 'lIld eredil'le upper and in Implementation
lower \'alues. Characteristic, counterproductive piltterns of
implementing corporate risk analysis also recur
Independent Multiple Estimates 'lmong I1llxiern oil and gas companies.
Companies employ multidisciplinary exploration
teilms, thus promoting cOl1sider,ltion l't independent Maldng Risk Analysis More Important
multiple estim,1tes and multi~'le wllrkil1~ hypotheses. than Prospect Generation
The teilms generilte several n'r~illns of r1"h1ny geotechni The most important tilsk in petroleull1 exploration
cal maps. They. jll\'ite peer re\ie\\'s of eii~erging prnjects is the generation of new prospects and plays. That is
and utilize exploration committees thilt rro\'ide experi where the value is added, and it is the most creati\'e
enced counsel and mentoring, but not ,It the expensE.' of 'lspect of the business. Geotechnical stilff must be
project accountClbilit~· by the t'riginatin~ St,ltf. a\\'ilre of this bilsie fact. After the prospect or play hilS
been identified, gellscientists then carry out risk ancll\,
Positive Postdrill Reviews ~i~ of the rroposed venture. The:' ilttempt to meilsu~'e
We Il:arn best lw constru.:t!\"(, examil1,ltion l'i our it by estil1l,lting reserves, c11i1nce, 'llld profitability;
own errors. The learning pn)cess is 'Ttil11ized by a they de::.ign appwpriate bidle\'l'!:'i, select prospects for
spirit of open, objecti\'e inquiry, which 1.1JY be ditiicult ,111 optimum portfolio, and eXilllline predictions and
to sustain because dedicated professi'):l,lb otten find results to improve performance. Risk 111/17'.1/5;:=- iPill helll
their mistilkes painful and IllJy fcor criticism and loss 115 (Ol/dlll'l (1111' ['''::;lIc.;::._'';':·lJ"I' 11C.:';:,fI1/>1v I7l/d cll/dClltll/, [lilt
of status. It is essential that rllstdrilJ ;"t'\'icws Li,' not it dews 1I[lt. lilld oil t111t! g115. It pre~UPP[l;[':: e ,,;;t(";7dy SIIJ;P'.1f ll(
take on a punitivL' ,1S~)ect. h)r th,lt re,lSl"l ~u(h rL'\'iews .\l)(l"IJ/"l15,l\..-f,,;;, I7l/d fillb o(S(lod pr05/JI'L't[11'.';.
Chapter 7 123
The Desire for a Cookbook maps and sketches. Preparation of formal maps and
- Meaningful exploration risk analysis requires that
professional staff must understand and integrate (1)
diagrams is not cost-effective for such presentations.
. AlI.of t~le following methods are based upon a dis NOTE: Th~ statistical mean is theoretically the pre
tnbutlOn 111 which if == 9, and the constituent values ferred expression. However, I recommend ~gainst its
are: general use, for the follO\oving reasons:
NATIJRALLOGS
P90% == 0.6MM KEY VALUES (III)
1. For field-size distributions, we do not ordinarily
P80% == I.25MM P99% == O.1MM anticipate that a given trend or basin will have an
P70% == 2.IMM P90% == 0.6MM infinite number of fields; instead we generally
-. P60% == 3.3MM
P50% == 5.0MM
P84% == 0.95MM
P50% == 5.or...1M ].6094
find tens to perhaps a few hundreds of fields.
Accordingly, the continuous distribution, when n
P40% == 7.6MM PI6% == 26.5MM == 00, seems inappropriate, and leads to an inflated
P30'}'0 == I2.IMM PlO% == 43.0MM mean field size.
P20% == 20.0MM PI % == 250.0MM 2. For prospect parameters such as Area, Average
PlO% == 43.0MM Net Pay, Gross Rock Volume, HC-recovery in
Total == 95.95MM barrels per acre-foot (bbl / af) or thousand cubic
feet per acre-foot (mcf/ a f), Prospect Reserves,
'"'. Initial Production Rates, etc., the recommended
1. Arithmetic Menll (x)
x == 95.95 .;- 9 == 10.66 MM probabilistic estimating connotations of P99%,
....., 2. Swnnsoll's Menll (Msw): P90'lo, P50%, PIO%, and PI % treat those very
large outcomes greater than PI % as practically
0.3 (P90%) + 0.4 (P50%) .+ 0.3 (PlO%)
and geologically impossible. Because the statisti
Msw 0.3 (0.6) . + 0.4 (5.0) + 0.3 (43.0) ==
cal mean includes contributing values larger than
I5.08MM PI%, such untruncated means are unrealistically
3. Stntisticnl Menll (Mst): assumes a continuous dis large. The P99% and PI% estimates then become
tribution, (i.e., n == 00), with contribution from val very useful as lower and upper plausibility or
ues greater than PI 01., credibility checks that encourage iterations lead
==ell+a2/2
ing to greatly improved estimates.
'.
== e,,,(5.0) + 111(2;;.:;/3.0; 2/2
3. When such distributions are truncated at PI%
'Il
== C (5.0) + 111(5.3)2/2
Combining Probabilistic
Distributions by Multiplication
curves plotted on log probability uet, not 90%. However, mulliplying the three
paper; 50th percentile values docs indeed yield a
• Useless in calculating the mean. 50% reserves product!
~
127
128 Graphical Method for Combining Probabilistic Distributions by Multiplication
II
fit line through these three points: this line is X.I = In Y.I
I
the graphical reserves distribution. Pick off
the P90% and P10% reserves values (and the
PSO% reserves value) and employ them using
r
Wher Xi is normally distributed with mean ~L. and
variance a2• All the variances are equal. I
Swallson's Rille, to find mean reserves. (This assumption causes no trouble. Our tests
C. ALTERNATIVE APPROACH: What are the showed thilt we would have errors of no worse than
appropriate percentile values ~hat, when mul 4% using this approximation. In oil ann gas explo
tiplied, will generate 10% and 90% reserves ration, a 4'Yu error amounts to a direct hit.)
values? It turns out that multiplying the three
23% values gives a 10% reserves product, and Multiplying Ys is equivalent to adding Xs.
multiplying the three 77% values gives a 90%
reserves prodi.\ct (see Ed Capen's A Proof InR==ln(Y\xY2 xY 3)
following). We can put this to work: == Xl + X2 + X3
1. Find the 23% values for Area, Pay, and
HC-recovery from the plots you made on What is the variance of the sum of the Xs?
log probability paper. Multiply them to Assuming independence, the variance of a sum is
produce the 10% value for resen'es; plot the sum of the variances.
this value on log probability paper.
2. Similarly, find the 77% values for Area, aL2 == variance of the sum == 3 a2
...... (i'1""-l500)
,.---,--r7"=-, 0 I %
l---\---I-t----;10%
(PI =75') ~
01'«
10'k
P90% = 10'
40'
PSO% = 20' 90Ck
10'«
BAJi'.= barrels per acre-foot (also bal) (P99 = 60) C===::I::::~.uL.-_...J 99Ck
10 100 1000
He-Rec (bal)
00:
10 100 JOOOIO,OOO.J I. 10 lOll 10 100 1000
'--'4:T
__-' _ _ ~ _ _ 99'7,
MBbl = mil/ioll barrels .01M .IM I.OM 10M
M =milliOIl ill this figure
.3 (0.2) + .4 (0.8) +.3 (3.3) = 1.37 M M = Mz
130 Graphical Method for Combining Probabilistic Distributions by Multiplication
•
Multiplying Three Cumulative Probability Distributions
~===== 1500 9.6 P1
P23
- P10
0.8 P50 1
. P77 RESERVES, MM
60
0.2 P90
Log scale
P99
TWO IVIETHODS FOR DERIVING THE RESERVE DISTRIBUTIONS:
AREA AV NET PAY HC-REC. RESERVES
I. P90 x
P50 X
P90
P50
x
X
P90
P50
=
=
P9807}
P50
DETERMINE
P10 and P90,
P10 X P10 X P10 = P1.3 CALCULATE MEAN
P50 X
P23 X
P50
P23
X
X
X
P50
P23
= P50
= P10
} CALCULATE MEAN
Figure B-2
Appendix A Recurring Problem in Estimating
Prospect Reserves-lJetermi~5ng Reasonable
"Low-side" Values (1'99% and P900/0)
During the period 1935-1969, a field-size distribu which approaches an effective reservoir minimum).
tion of discoveries in the Texas Gulf Coast (Figure C~ 1) The resulting reserves distribution had P90% reserves
indica tes tha t the smallest field sizes (= P99%) con of about 96,000 bbl and P99% reserves of 27,000 bbl.
tained between 3,000 and 4,000 recoverable bbl. Dur
ing the next 19 years (1970-1988), the P99% discovery Example 2
... size was only a little smaller-about 2,000 bbl (Figure
C-2). Similarly, in the northern part of the Midland But consider a second situation, where consistent
Basin/Red River Arch province, the P99% field size and reliable regional log and rock data indicate that
was about tOoo bbl as of 1945, and about 650 bbl as of average net reservoir thickness is never less than 16
1957. In the Silurian (Niagaran) pinnacle~reef play of feet (5 m) and ranges systematically and regionally up
northern Mkhigan, P99% for disco';lered fields was to 80 feet (2-1 m). Even taking into account the geomet
.... about 5,000 bbl if) the period 1968-1974 (inclusive), ric adjustment factor, P99°1.. for A\'erage Net Pay prob
and about 1,000 bbl for the period 1975-1988. And yet ably shouldn't be less than about 12 feet; P90'X, should
all these very small, money-losing events qualify as be around 16 feet. Consider a prospect in this setting,
"discoveries"-:-mobile hydrocarbons were sensed, the where HC recoveries are consistently poor, ranging
wells were completed, and oil was produced, presum from 60 bblJaf to 150 bbl/af. Here, the productive area
ably for sale. The implications are depressing but pro necessary to provide around 10,000 bbl-about the
found: there is always a possibility that YOllr discovery is minimum amount of recoverable oil necessary to sus
going to tum Ollt to be n very small accu11Iulation! tain flow-\\'buld of necessity be around 14 acres (5.7
Most experienced exploration risk analysts know ha) if AveraO'eo Net Pay _ was 12 feet and HC-recovery
that the most common cause of overestimation of was 60 bbl/ af. However, only 6.5 acres would be
'.
prospect reserves is tha t the "low-side" estimates required to contain 10,000 bbl if Average Net Pay was
(P99% and P90%) are too large. But how to arrive at 16 feet and HC-recovery was 100 bbl/af.
appropriate low-side estimates? This section addresses
that problem using six different examples.
Example 3
Example 1 You must also take into account the effects of
'. geometries consistent with the anticipated trap type.
In our short courses, we use a simple prospect exer- For example, suppose your prospect is a fault trap as
cise called the RMAG prospect (for the Rocky Moun shown in Figure C-3, minimum reservoir quality is
tain Association of Geologists, where it was first used). projected to be 60 bbl/af (P99%), and average net pay
The RMAG exercise provided good examples thickness in such poor reservoir rock must be at le<lst
Whereby the low-side values for Area, Average Net 40 feet (1201).
Pay, and HC-recovery Factor were all qUite low Given that the dip-rate of the reservoir-zone is
because the existing data demonstrated that such mar- about 300 feet (90 m) per mile, this geometry imposes
ginal values were entirely possible. Thus the P90% minimum value upon possible area of the low-side
value for Area was about 60 acres (24 hal, and the case (Figure C-4).
P99% value was about 20 ilcres (8 l1a). The P90% villue In order to arrive at the requisite 40 feet of average
for Average Net Pay thickness was about 6 feet (2 m) net pay in such trap geometry, the height of the hydro-
and P99'X, was 4 feet. HC-recovery Factor was given carbon col umn must be abo\i t '6~-": -.:et. '~":ijs forces
(P90% = 100 bbl/af [barrels/acre-foot]; PlO'}'o = 300 the downdip position of the oil/water contilct to be
bbl/af, consequentially P99% was ilbout 60 bbl/af, positioned about 1,000 feet south of the trapping fault,
Ii 131
132 A Recurring Problem in Estimating Prospect Reserves
- - - - f--I - - f-- f - - I
v ;I'
I--
P1
P2
1
-
~.
1·
I
~
i..
'-- - - - f--I - - - - f-- f - - ' - , tI - - I- P5
.L~
_-t-+,_-H-H-H--_+--t-+-++++++--l-+--+~-+;+++H
1---_-+--_+-H--+++++--+ ...H-~-e-:an~· "'p~o,-+-t+ttt---+-I--+_-+-+++-I+l :::
:a '- - -
~ .-I--_--+--+
!== i=
1
=t : J...~
--+-f-++-I-!-"..l.-_+-+-"'-++++-t+_---7-----+-+-++-HJJ-~L~;
- - -~-
- -. - - - - ,
P30
:~~ i .
I
A1t
Q) - - -
=oi
-
:= := = ~ ::= =I i !
:-11
L.. I
:
' ! I
=±I : =-~ : ==: ±! ::~
1---1--+-t-+++~;-7-:--+--t~-:L;Y~rP901+--':":-+-+-+,-H-t+t---+-,-t--t-TI7 tttf---i:,-+-+-i-i-'i-Ti+--+--+-H-+++-LJ P90
~: !/~! :- : - + ~ tl : - -
t -+ =~
1 1
r-- - -
" "i,
-' ,'-.
~ _'---'--~--'-'-'.J....._
i . !
- - e- T - L....
I
I!
I
,- - - P98
'----'----JL.-..jI"-'--'-...u...._ I I .
_ _J.......C....:....:.-"-'-'-'---____'____'c......;....__"__'_..:....:.L.l....__.lL_JL_J......L_1..J..J...L_'____L...._L....l._..L.l..LLU P99
Figure C-l
measured along the north-south axis of the structure, bbl required to support sustained flow.
as shown by Figure C-4. Because the structural geome
try requires an elongated east-west closure, the length Example 4
of the minimum closure measured along the fault must
be about 1.85 miles, or 9,800 feet. The area of this mini Consider a prospect located on a large, gentle
mum closure (shaded in Figure C-3) is about 112 acres domal anticline (Figure C-S), in which the reservoir
(9,800 ft. x 1,000 ft...;- 2 x 43,560 ft 2 /acre).1l1e product of dip away from the apex of the closure is very gentle,
the three P99% values is 269,000 bbl (112 acres x -10 ft. say 30 feet per mile. Further, let us suppose that
ave. net pay x 60 bbll a£). Ob\'iously, the corresponding regional reservoir data indicate that P90°';, = 120 bbl/ af
P90% values for Ayerage Net Pay, Area, and Reseryes and P99°o = 60 bbl/af are appropriate estimates for
would have to be much larger. this prospect. Also, assume that appropriate estimates
For additional perspective, suppose that the required for net reservoir thickness are 15 feet (P90%) and 10
P99% value for average net pay is 20 feet, rather than 40 feet (P99°';,). The structural geometry leads to the fol
feet; still using 60 bbl/af, this places the oil/water con lowing possible P99% outcome, keyed to a 10,OOO-bbl
tact 500 feet south of the fault, and reduces the minimum minimum reserves case (Figure CoS):
area (P99%) by half, to 56 acres. Additionillly, prospec We now adjust net reservoir thickness to obtain
tors must remember that the reserves P99% value is a\"erage net pay: P99% = 8 feet and P90% = 12 feet.
the approximate product of the P90% \'alues for areil, Now, in order to achieve the requisite 8 feet of P99%
average net pay, and HC-recovery factor. For elongated average net pay thickness, the thickness of thC';0i~ cQl;.
structural traps, P99% reserves accordingly may be umn at the apex of the structure must be around 15
substantially greater thun the minimum 10,000 to SO,OOO feet, which then requires a linear distance of about 1.0
Appendix C 133
I-- !
""- ~
-+
-}
-
~I ~
~ -
•
P1
P2
1\
1\
~ l- I- - f- f- I I ~ - - - -
~
i
'
i.;' ;rri
P10
-- - I PS
P10
finI - J~
i I
f'
1\
....,
~
:0
III
1-
r-
I
I-
-
-
-
-
- -
- -
-
-
-
-
-
-
I-
f:;r
~;
"ii'
,vL . -'Mean
_ _ ~_
I
! +-t
\:1
,
1,-
P20
P30
I
,
.0 '-- l I- - - - - - - - --~- 1 ' P40
- ...
0
0
l"I'i j, : I pso
P50
~
+J u
Q)
>
+:i
~
:J
-
--
--
- -
-
-
-,
-
-
-
-
~
-
-
-
-
-
-
,; "-
~
-
--
-
!
ir--
-
-
-
-
-
-
-
-
-
-
-
I ..
+-:~ -
1-:, : i -
"'-'
J I-
-
I
r
P60
P70
P80
~
f-
E ~
-, :J
() ~
ttl'
i II I; : I
I
P90
i-
; , , I
-- - - ~- P90
I-- . - - -
I - - - }1 II j"""1 1 P95
~~'
~ - I
~
---1 ~ -- -J--I ; I
...-I~
I-- -I f- - - 1-1 f- - - - - - '.;.....L 1 P98
i i ,: .I P99
0.001 0.01 0.1 1.0 10.0 100.0
. I 1.B5 miles-pi
1'"""'<
"
I I
~ "'I iiffPid~t;:o;::;ttT:;;;;:;:
~I~~:: :a I6S: : :\G): i·/
b===i l::=::::;===:;::1
,_:!- ' / o
......... --..... ~oU..J..l~~_ ..~,\~G ,/
.2
' .......... - .4
- _- 2100
_I - .....- --- - .. ~~sG . / .6
____ - 2200 __ -
------t-
... 2300
.8
1.0 m
wI o. ..2 .4 .6 .8 1.0 mile
_ IowoI
~t
zCJ?1
Figure C-3
mile for the length of the elongated axis of the satu P99'X,. Obviously, P90% area would have to be much
rated area. The corresponding area of closure is: 11:r2 = larger, perhaps approaching 1,000 acres.
3.1416 x (2,640 ft,)2 -;- 43,560 ft. 2 /acre = 503 acres =
II
134 A Recurring Problem in Estimating Prospect Reserves
20
40
Figure C-4
I I
·1 ~ 1 mile (5280') )' I
Dip-rate ~
=30fVmile : :. . . .
~-'-----"
k J.t'i1. Hf HFHH1.lrS' \1 tiHi~L'!H .. t- - - - - ~
8' '-'~~:~~O'
~.,,~... 1 ~~
~
..... I I ····'····r~
. -'.'.
I I
1-' 1 mile (5280') ~I I
I I
Figure C-5
Example 5 Example 6
Suppose you are exploring an undrilled frontier Consider a prospect in the Gulf of Mexico that fea
trend containing perhaps 100 structural closures, some tures a fairly well-defined amplitude anomaly that is
large, some intermediate, some small. Given the usual congruent with structural contours, and a "flat spot"
exploration pattern, your prospectors call identify clo that may indicate a gas-water contact, also consistent
fiures too small to contain the 25 million barrels (MM with the structural picture. Your geotechnical staff
bbl) they think is the minimum required to de\ dop; . think there is ± 20% possible variation in the median
but what your prospectors CllllllOt consistently do in Area parameter of 400 acres because of imperfect reso
such a trend is to identify those large closures that con lution. So, in this case, P90% = 320 acres and P90% =
tain very sll1all accumulations (because of severe -180 acres.
underfilling, or very poor reservoir quality or thick What's the message? There's no "cookbook"-you call
ness, or because of faulty geophysical data or interpre not blindly assigll a low-side value to tlzereserves parame
tations). In any case, if you drill only prospects you ters. You must understand the range of geologic
think are large enough to contain at least 25 Mlvl bbl, values, the structural patterns, and the exploration
your low-side areas (P99%) could \\"ell be larger than maturity of the trend. Then you must integrate these
20 or 40 acres, and P90% areas could certainly be sev considerations and think probabilistically. But, in any
eral hundred ;Jcres or even more. As a result, P99% case, it's not a bad idea tc siart with very small
reserves could well be larger than 10-20,000 bbl-it assumptions for P90% and P99% values for Area,
could properly be 50-100,000 bbl, or maybe more. Cor Average Net Pay, and HC-recovery Factor-and thell
respondingly, P90% reserves might be from about adjust thelll upward as the geoteclmical evidellce justifies.
200,000 up to around one million b,,;;2L, 0; rnor~...But Final admol1ition:-RememrJer th.n~he n·•..,...;t corn
o/lly because yOli are /lot going ta drill llny but the largest man field size discovered in west Texas in the 1930s
closures. and 1940s .(during the "flush" period of discover .v,
Appendix C 135
In many prospects and proposed trends or contract areas, there may be multiple-zone, or multiple-play candi
dates. Some multiple ventures may be simple involving only two objecti\'es and ,·..,ith clear independent and lor
shared attributes. Other multiple ventures may be quite complex im'olving many objectives and with complex and
subtle issues of partial dependencies further complicated by issues of marginal commerciality.
Sometimes a prospective contract area may contain an existing trend ha\'ing some remaining exploration poten
tial, as well as a second, deeper play, or another play in a different part of the exploration area. Some plays may be
part of a single petroleum system involving partial dependencies among some or all geologic chance factors. In
SOme prospective contract areas there may also be one or more producing fields to consider as well.
Several practical principles of multiple ventures should be emphasized:
1. As the number of objectives in any multiple venture increases arithmetically, the complexity of combination
increases geometrically;
2. Monte Carlo or Latin Hypercube simulation is required for all but the simplest multiple-objective ventures;
3. Geologic dependency is common, especially involving the hydrocarbon charge and closure chance factors;
4. Many apparently complex ventures can be simplified and approximated by practicaL common sense methods
of considera tion;
5. Analysts are well advised to focus on objectives that are "stand-alone commercial or economic" and ignore
those that provide only incremental cash flows augmenting primary producing objectives; and
6. For comprehensive treatment of multiple-objective ventures, the reader is referred to James Murtha's excel
lent 1995 SPE paper.
FollOWing the principle that a picture (or example) is worth a thousand words, two examples are provided
herein, which should provide the careful reader with a good understanding of the concepts and procedures llsed in
eva.luating multiple-objective ventures. Part 1 describes several aspects of a two-objective multizone prospect, in
which Monte Carlo simulation is avoided by basing calculation on only the mean reserves case of each objective.
Part 2 describes a multiple venture involving two plays and a producing property, which employs a decision-tree
analysis as well as Monte Carlo simulation.
........'C" • •
137
138 Evaluating and Combining Multiple-objective Ventures a.
t
Part 1: Multiple-zone Prospect Example:
Geologic and Economic Assumptions •
•J
Description Prospect with two objectives, A & B. Dry I
CASES 1 & 2: hole cost = $1 MM. Note: In this example, only four geologic
chance factors are employed. Also, geologic chance of suc
I
cess (Pg) = commercial chance of success (Pc). ;1
._ Evap. seal
Prob. structure =0.5 (critical closure, seis resolu
~~~~~~~~.~-~-• - C03
Prob. seal/trap = 0.9 (preservation of evap. cap)
:J reservoir
Pc = .221 (Pf = .779)
EV = 0.221 ($10 MM) - 0.779 ($1.0 MM)
= 52.21 MM - $.779MM
= (+) $1.431 MM
Sandstone.: Objective B Low-relief anticlinal closure, sandstone
reservoir' reservoir o\'erlain by marine shale top seal, also mature (?)
source rock. Mean reserves est. 10MM BOE, equiv.
PV=$20MM.
._~ .
Appendix D 139
Case 1: All Geologic Chance Factors Are Independent; Dry-hole Cost =$1 MM
Productive"- .. .",- Venture
A B Prob. Value Risked Value
Ch ance
Both: 0.022 a)AandBbothproductive .221 x .098= .022 x (+)$30MM =(+)$0.660MM
Only one: 0.275 b) A productive, Bdry .221 x .902 = .199 x (+)$10MM = (+) $1.990MM
At least one: 0.287 c) A dry, Bproductive .779 x .098= .076 x (+)$20MM =(+)$1.520MM
-,d),-A_a_n...:...d_B--:.b_ot_h..:..d~ry .7....:,0_3_x----'(-~)....:..$~lM....:,M
...:....7...,...7..:....9_X_.9..:....0-=2:-.=_ _ =....><---'.).....:$_0_.7_0_3M_M_ I
1.000 EV = (+)B,467MM
Case 2: Dependent (Structure, HC-charge) and Independent (Reservoir, Seal/Trap) Chance Factors;
Dry-hole Cost = $1 MM
Step 1: Prospect A = Pres x Pseal =.7 x .9 = .63 = Pg (and .37 Pf) are independent factors
Prospect A = Pstr x PHC = .5 x.7 = .35 = Pg (and .65 Pf) are dependent factors
Prospect B = Pres x Pseal =.4x.7= .28 = Pg (and .72 Pf) are independent factors
=
Prospect B Pstr x PHC = .5 x .7:= ,l2. = Pg (and .65 Pf) are dependent factors
Step 2:
Productive Both Venture
Chance A B Prob. A and B Prob. Value Risked Value
Both: 0.062 a) A and Bboth productive .63 x .28= .176 35 .062 x (+)$30MM =(+)$1.86MM
Only one: 0.195 b)Aproductive,Bdry .63 x .72= .454. 35 .159 x (+)$10MM =(+)$1.59MM
At least one: 0.257 c)Adry,Bproductive .37 x .28= .104. 35 .036 x (+)$20MM =(+)$0.728MM
d).A and Bboth dry [1 - (.062 + .159 + .036)] = 1 - .257:= .743 x (-)$1 MM = (-) $0.743MM
EV = (+)$3.435MM
Case 3: Dependent and Independent Chance Factors, as in Case 2, but Now Including One Geologic
Chance Factor with Subfactors That Are Both Dependent and Independent; Dry-hole Cost = $1 MM.
Because of the position of the HC source rock as noted in Table D-1, there might be some reduced chance of hydrocar
bon emplacement in the sandstone reservoir. .
For Prospect A,we break down the HC-charge chance factor into three subfadors:
Probability or HC-charge = 0.7
Subfactors: HC-I) adequate organic richness = 0.9 }
HC-2) thermal maturity = 0.9 Product := 0.73
HC-3) migration/ emplacement efficiency = 0.9
S.ubfact?rs 1 and 2 are common to Prospect B as well as Prospect A, but 5ubfactor 3 (migration/emplacement effi
CIency) IS more problematical in Prospect B, and is therefore assigned a probability of 0.5. All other geologic chance
factors remain as shown preViously in Case 2.
Step 1: Prospect A (Independent) = Pres x Pseal x PHC3 = .7 x .9 x .9 = .56 = Pc (and .44 Pf)
Prospect A (Dependent) =Pstr x PHC1, 2 = .5 x .8 = .40 = Pc (and .60 Pf)
Prospect B (Independent) := Pres x Pseal x PHC3 = .4 x .7 x .5 = .14 =Pc (and .86 Pf)
Prospect B (Dependent) = Pstr x PHC1, 2 = .5 x .8 = .40 == Pc (and .60 Pf).
Step 2:
Productive Both Venture
Chance A B Prob. A and B Prob. Value Risked Value
Both: 0.031 a) A and B both productive .56x.14:=.078 .40 .031x (+)BOMM =(+)$.936MM
Only one: 0.218 b).A, productive, B dry .56x .86=.482 .40 .193x (+)$lOMM :=(+)$1.93MM
At least one: 0.249 c) A dry, B productive .44 x .14=.062 .40 .025x (+)$20MM =(+)$.500MM
d) A and B both dry [1 - (.031 + .193 + .025)] == 1 - .249 == .751 x (-)$1 MM == (-) $.751 MM
EV == (+)$2.615MM
140 Evaluating and Combining Multiple-objective Ventures
•
Combining Reserves Distributions
1. In this example, the reserves distributions of the t'v'{o prospects have not been combined, because the P10%
P50%, and P90% values were not provided, only the means). So, for the "both A & B" productive cases:
reserves were combined simply by adding the two means, which is not strictly accurate.
2. Properly, reserves distributions of the two prospects, represented by PlO%, P50%, and P90% values, can be com
bined by Monte Carlo simulation. However, this requires combining chance-weighted distributions through
ndditiol/, not IIlJlltiplicntioll, as we did graphically in Appendix B, with the three components of prospect reserves
(Area, A\'erage Net Pay, and HC-recovery Factor). Unfortunately, adding probabilistic distributions is not prac
tical to perform graphically, so we should rely on Monte Carlo simulation to carry it out. .. - :
,
Part 2: Combining Multiple Types of Ventures*
This multiple-venture example illustrates the procedural steps involved in combining three objectives-a pro
ducing property, a mediulll-Iisk ne'v\' play, L:~d a high-risk new play (Section A).** It employs a probability tree
for the partially dependent plays, as well as a Monte Carlo simulation. The probability tree solution (Sections B.1
and B.2) uses only mean reserves outcomes for the different cases, yielding a total expected mean outcome for
reserves of 36.10 MMBOE, hadng a total expected net present value (ENPV) of $261.7 MM. Section C, the Monte
Carlo simulation (50,000 trials), shows the probability range of possible reserves outcomes. The distribution of
reserves outcomes for all possible cases is shown as a frequency distribution (or probability-density graph). The
P90% outcome is 11.19 MMBOE, the P50% outcome is 35.49 MMBOE, the P10% outcome is 153.22 MMBOE, and
the mean outcome is 36.72 ~1~1BOE, which agrees well with the 36.10 MMBOE figure derived using the probabil
ity tree and mean values (Section B). Thinking about the overall venture in a simplistic way, the P90'Yo outcome
expresses a scenario in which the producing property low-side case turns out to be the only project of the three
that is successful, e\"en marginally. The P50% case (35.49 MMBOE) represents an outcome in which the producing
property turns out to be the P50% to P \1£.-\:" case, and one of the two plays emerges successfully, providing "mid
range" resel"\"es (=P50% to P\IE.-\N if Play "Beta," or P80% - P60% reserves if Play "Gamma"). The PIO'X. outcome
for the overall venture would either require all three ventures to succeed moderately, or the producing property
and one of the h,,'o plays to succeed at relatively high probability levels.
Similarly, part 0 shows i\lonte Carlo simulation of probabilistic ENPV for the multiple venture outlined in part
A, employing the appropriate ~PV /BOE for the different subventures ("Alpha," "Beta," and "Gamma"). In other
words, part 0 utilizes all the probabilistic outcomes of the entire venture in terms of ENPV, rather than reserves,
and allows the ENPV of all possible resen'es outcomes to be expressed.
"The assistance oj Robert V. Clarr i~ ackno\\"lt'ciged with gratitude; Clapp prepared the probability tree (Section B) and carried out the
Monte Carlo simulations (Sections C and D). ;-':l't.~ that the distributions were not truncated above PI% and that Clapp utilized the "PIO =
small" convention.
""Chance of geologic success bast'.-l upon fin~ .-h.mce factors; also, ch'l1lce of geologic success = chance of commercial success in these
examples.
._~ .
Appendix D 141
A moderate-risk, moderate-potential, relati\'ely shallow play ha\'ing chance of play success = 0.72, chance of
average economic prospect success = 0.34, four tests committed, so chance of at least one economic prospect suc
cess is 1 - (l -.34) 4 = 0.81, so chance of economic play success =0.72 x 0.81 = 0.58: Cost of play failure is $20MM.
Economic Play Reserves: Geologic Chance Factors:
PI % = 1.0 MMBOE Shared = 0.72
P10% = 3.4 MMBOE HC source rocks = .80
PSO% = 14.0 MMBOE HC generation/migration = 1.00
P90% = 60.0 MMBOE Independent = 0.34
P99% 200.0 MMBOE Resel'\'oir = .60
Mean = 26.5 MMBOE Closure = .80
Containment = .90
Pmefs= .70
A high-risk, high-potential, deeper play, having chance of play success = 0.4 or 0.8, chance of average economic
prospect success = 0.17, three tests committed, so chance of at least one prospect success is 1 - (l - 0.17)3 = 0.42, so
chance of cumulative economic play success = 0.8 or 0.4 x.42 = 0.336 or 0.168. There is p",·tial depnndency between
Play Gamma and Play Beta, involving the shared chance factor HC source rocks: if Play Beta proves that HC source
rocks are present, then the chance of HC source rocks will increase for Gamma to 0.8; if HC source rocks are absent
in Beta, chance of HC source rocks will decline to 0.4 at Gamma. Cost of play failure is $30MM.
Economic Play Reserves: Geolog.ic Chance Factors:
PI % = 2.0 MMBOE Shared = 0.80
P10o;" = 8.0 MMBOE HC source rocks = .80
P50% = 44.0 MMBOE HC generation/migration = 1.00
P90'Yo = 240.0 MMBOE Independent = 0.17
P99''l'o = 1,000.0 MMBOE Reservoir = .60
Mean = 105.0 MMBOE Closure = .80
Containment = .50
Pmefs= .70
---
Appendix D 143
Probability Tree Showing All Possible Combinations for Beta and Gamma Plays
,---------------------------------\---------------------------1
BETA ::
,I
GAMMA : Beta
1_ Gamma Probability
:: At least 1discovery :
:: (3 tries) : Success Success 0.196
II
r
Source rock 0.42 ,
Atleasl1 discovery J
. I
0.8 . ,
I
Nodiscove Success Failure 0.271
0.81 I 0.58
I
I
1LU.>L.>ol"-""""-J~ _
Success Failure 0.117
II 0.2
,I,I
II
0.9 I
: 0.8 ,
I
I
No discovery
I
I 0.58 : Failure Failure 0.063
0.19 ,
I
If-.U.IL-"""'lj,&J"""'- _
Failure Failure 0.027
:'II 0.2
,I
0.42 Failure Success 0.027
:: Source rock I
: 0.8 !
No Containment I No discovery Failure Failure 0.037
0.1 0.58
. : II so~r~e roCk~
:. No discovery Failure Failure 0.046
'""-":'~":""'-Ll.ll.<lL-~ --";'[ No Souffie ~k 0.58 :.
::'I .05
""Tl
+oJ -,
.
.0
<0 .03
(l)
.0
C
(l)
.0 :3
0
~
n
a.. .02 u:::
735
.00 ~JlliUunuIl1UJJilullIunlq\llw.tutJupw.u.w1..............~ a
0.00 37.50 75.00 112.50 150.00
MMB
100% 11,948.83
._~
....... t::' • •
End of Forecast
Figure 0-2
Appendix D 145
.05 -r----;------------------~2.467
~h~rn~~,rnw~-----~-:17 j
.03
.0
o
L
a.. m .-.
.oo~
0.00 250.00 500.00 750.00 1,000.00
MMS
End or Forecast
Figure D-3
Appendix Steps in Geotech~i~alProcedure for
Prioritizing Petroleum'P'rospectivity
1. Map trends or fairways of different trap a. Before deciding to enter area? Cost?
types Confidence in results? Time?
2. What reservoirs would they affect? b. Before deciding to drill? Cost? Confi
3. Analog examples dence in results? Time?
4. Did traps (closures) fonn before generation/
migration? Evidence? V. PROVISIONAL PLAY IDENTIFICA nON,
5. Are traps favorably located with respect DELINEATION, GRADING
to migration? A. List possible plays (include proven plays
6. Reliability. of geological!geophysical da ta showing remaining prospective area of play).
J
(show areas on maps) B. Overlay maps showing kitchen, reservoir,
~. Repeatability-identification and traps, and containment for each play, grade
delineation by independent workers sectors of coincidence.
b. Reliability of geological/geophysical
data (show areas on maps) VI. CRITICAL PLAY UNCERTAINTIES
(i) acquisItion A. What are they? List and map.
(ii) time-depth conversions B. Can they be economically investigated/
(iii) processing resolved in time to influence decision to get I
(iv) variable reliability among differ into area? Cost? Confidence in resolution? I
ent trap types Time required?
(v) map your confidence in seismic C. Can they be economically investigated/
I
.' interpretation resolved in time to influence-decisions to drill? ,I
(vi) 2-D vs. 3-D-resolution vs. cost
7. Critical trap uncertainties-What are
Cost? Confidence in resolution? Time required? I
they? Can they be resolved or reduced? By VII. ANALOG BASINS AND PLAYS I
seismic? Other geophysical methods? A. Location, geologic succession, tectonic situa I
Geologic methods? tion, documentation
a. Before deciding to enter area: Cost? B. Elements that are very similar and significant
Confidence in results? Time? C. Elements that are dissimilar and significant
b. Before deciding to drill: Cost? Confi
dence in results? Time? VIII. EXPECTED FIELD-SIZE DISTRIBUTION
A. Source of data and confidence
IV. CONTAINMENT B. For analog basin?
A. Is containment already demonstrated by C. For analog play?
ability barriers?) INhere? Map them. IX. DETERMINING MINIMUM ECONOMIC FIELD
I
B. Where containment is not prO\'en and must be SIZE (MEFS)
predicted: A. For the proposed area, estimate the cost of I
1. What are potential top seals, seat seals, and exploration (drilling, seismic, oiher G&G,
lateral seals? Where are they developed? land, overhead).
Map them. B. At your estimated $PVX% per BOE (for example,
2. Preservation-make HC-gra\'ity and HC $3.50 PV12% per MMBOE), how much total
type maps oil must you find to cover the above explo
a. Thermal-map areas of thermal ration costs?
immaturity, oil-prone, gas-prone, C. What is the smallest field (MMBOE) that
overcooked (by stratigraphic or struc could be profitable on a stand-alone basis,
turallevel). Evidence? assuming several similar producing fields
b. Flushing-salinity and piezometric exist in this area?
maps showing grOlmd-water movement D. Divide B by C-Is it reasonable to suppose
c. Biodegradation-e\'idence? manifes that many fields exist in the area? Reiterate
tations? Map. until you have a reasonab!e number of eco
d. Post-accumulation structural adjust nomic fields that together would provide the
ments-tilt-and-spiIl, breaching; loca total oil to cover only exploration costs. The
tions? Evidence? average field size is MEFS.
"'1. CrifIcal containment uncertainties-What
4
149
150 Reconstructing Parent Field-size Distributions from Offshore FSDs
d. The natural or parent distribution can be recon If, with additional technology and infrastructure,
structed as follows: fields down to 2 MMBOE could be produced, the
1. Assume P99% = ""10,000 BOE-estimated reconstructed FSD suggests that perhaps four of the
minimum reserves required to flow (i.e., to be show-holes may represent such fields. .
detected); Also, if one estimates the likelihood of finding a
2. Total number of accumulations (14 shows + 10 field larger than, say, 100 MMBOE (given that the test
fields) = 24= n; well indeed makes a discovery in the first place!),
3. Incremental fractile % assigned to each accu using the observed offshore FSD, such an event has a
mulation = 100% + (n + 1) = 25 = 4%; chan~e of about 18%. However, if one employs the
"correct" FSD, reflecting the natural or parent distrib
.
,
4. Reconstructed parent distriblltion (Figure 38b):
ution, the chance of finding the same 100 MMBGE or
EUR FRACTILE larger accumulation is greatly reduced, to only about
FIELD (MMBOE) CUMULATIVE %
8% (Figure 38b).
The important things for the offshore explorationist
14 small -40,000 to 96% ~ 44%
to grasp are tha t:
accumulations 4.2 MMBOE 1. Any offshore FSD has already been naturally
38b);
H 9.5 32%
3. Marginally economic opportunities are repre
Appendix F 151
1% 1%
L §l
-4--+
(tj
, •
""
•
,.
,$
.~ en::2: .'
10% Q)
I
Xc) -_.
cuI.{)
• 10%
EW
--
~.
c
<ll
EO ~ II
/,.
10 OJ en
£
Qj
00
en O
(1)0
S-S [,./
iii
~
ctl°
xc)
(1)0
- 0
V"
C.9 50% ~~
CU._
EE
/' 50% /
..... 0 V
?f<.
~
enr-
s ....
Q) II
'- 0
~c
'-0 V
:0
<ll
(I) ....
..... 0
0.0
~~ V
.0
0
ctl..!..
E:J , ,. ~'
Q: 90% 'x j
.
. .
0
0 , " 90%
'
0. ,
0.
~ ,
99%
.' . ,
-' 99%
10,OOOBOE 50k 100k O.2MM 0.5MM 1.0MM 2.0MM 5MM 10MM 20MM 50MM 100MM 200MM 500MM 1.0BBOE
Figure 38 a
1%
1%
t
9
:5
()
10%
c
<ll
u
'E
-- A
10%
£ 0
c
Qj 0
iii u
CIl W
(5 2
0~
50%
50%
.£
:0
<ll
.0
e
0...
90%
90%
99%
10,OOOBOE 50k 100k 0.2MM 0.5MM 1.0MM 2.0M M 10MM 20MM 50MM 1OOMM 200MM 500M M 1.0BBOE
5MM
MM = million; k = thousands; B = billion
Figure 38 b
--
App~ndi~ ~-.
Matrix for Comparing, Ranking, and
Planning New Exploration Plays
I
.'
Play Name
i
Location, Basin, etc.
Analog Plays
Working Interest Effective WI
MEFS
P90%
Reserves P50%
Geotechnical Pmean
Parameters P1O%
Shared Chance
Independent Chance
Chance of Success Po
Pmefs
Pse
Critical Geo-Risks
Unrisked NRI Actual $Value (Net/Gross)
Value Present $Value (Net/Gross)
Explor. Failure Cost (Net/Gross)
Costs Total Success Cost (Net/Gross)
Risked Finding Costs (Net/Gross)
Max. NeG. Cum. Net Cash Flow (Net/Gross)
Economic Ranking Measures EPV ($) (Net/Gross)
Parameters Risked Inv. Eft. At 100% WI
OWl (%) At Corp. Risk Tolerance
Risked Inv. Eff. At OWl
Total Failure
$PV, Net Economic Worst Case
Reserves' Outcomes P900f0
($MM) P50%
Pmean
P1O%
Start Date
Timing Max. NeG. Net Cash Flow mate)
Pavout Date
Proiect Life
Political Risks & Primary
Consequences Secondarv
Business Operating Risks & Primary
Criteria Consequences Secondary
Method of
Acquisition
Contractual Needs
Business Needs &
Sensitivities Contractual Concessions
NeaotiatinG Strateqies
Personnel & New Staff
Service Needs Tech Services
Comments & References .
153
References Cited
Alexander, J.A., and J.R Lohr, 1998, Risk Analysis-Lessons Breale\·. RA., and S.c. ~'Iyers, 1988, Principles of Corporate
Learned: SPE 49030: Society of Petroleum Engineers Finance: ~1cGr'I\\'-Hill, New York, 889 p.
ArU1ual Meeting, New Orleans. Brown. P.}., and P.R Rose, 2000, The "Gray Area" between
Andreatta, G., and G.M. Kaufman, 1986, Estimation of Prospects and Plays-Assessing Volumes, Value, and
Finite Population Properties when Sampling is without Chance: EAGE 62nd Conference, Glasgow, Scotland, p.
Replacement and Proportional to Magnitude: Journal A-33.
of the American Statistical Association, Vol. 81, No. Capen. E.e., 1976, The Difficulty of Assessing Uncertainty:
395, p. 657-666. Journal of Petroleulll Technology, August, p. 843-850.
Arps, J.K., and T.G. Roberts, 1958, Economics of Drilling for Capen. E.e., 1984, \\-h;; Lognormal?, in Capen, E.e., R.E.
Cretaceous Oil and Gas on the East Flank of the l\legiIJ, and P.R. Rose, eds., Course notes for AAPG
Denver-Julesberg Basin: AAPG Bulletin, Vol. 42, No. School "Managing and E\'aluating Petroleum Risk":
11, p. 2549-2566. AAPG Education Dept., Tulsa, Okla., 350 p.
Arthur, M.A., and S.o. Schlanger, 1979, Cretaceous "Ocean Capen. E.e., 1991. Rethinking Sunk Costs: SPE 22017,
Anoxic Events" as Causal Factors Development of Reef H\'cirocarbon Economics & Evaluation Symposium:
Reservoired Giant Oil Fields: AAPG Bulletin, Vol. 63, S0ciety of Petroleum Engineers, Dallas, p. 1--{,.
p.870-885. Capen. E.e., 19')2, Dealing with Exploration Uncertainties,
Attanasi, E.D., and D.H. Root, 1994, The Enigma of Oil in Steinmetz, R., ed., The Business of Petroleum
and Gas Field Growth: AAPG Bulletin, Vol. 78, No.3, hFloration: A.-\PG Treatise of Petroleum Geology
p.321-332. Handbook of Petroleum Geology, Chapter 5, p. 29-61
Baker, RA., 1988, When is a Prospect or Play Played Out?: (plus 4 errata pages for some printings).
Oil and Gas Journal, January 11, p. 77--80. Capen, Le., 1995, A Problem with Experts: SPE Hydrocarbon
Baker, RA., H.M. Gehman, W.R. James, and D.A. White, Economics dnd E\'aluation Symposium Proceedings:
1986, Geologic Field Number and Size Assessments of Society of Petroleum Engineers, Dallas, p. 255-266.
Oil and Gas Plays, 'in Rice, D.O., cd., Oil and Gas Capen, Le., 1996, A Consistent Probabilistic Definition of
Assessment-Methods and Applications: AAPG Reserves: SPE Reservoir Engineering: Society of
Studies in Geology, No. 21, p. 25-31. Petroleum Engineers, Dallas, p. 23-28.
Bally, W., and S. Snelson, 1980, Realms of Subsidence, in Capen, Le., R.V. Clapp, and W.M. Campbell, 1971,
Miali, A.D., ed., Facts and Principles of World Competitive Bidding in High-Risk Situations: Journal
Petroleum Occurrence: Canadian Society of Petroleum of Petroleum Technology (SPE-AIME), June, p. 24-36.
Geologists, Memoir 6, p. 9-94. Capen, E.e., R.V. Clapp, and W.W. Phelps, 1976, Growlh
Bayes, T., 1764, An Essay Toward Solving a Problem in the Rate-A Rate-of-Retum Measure of Investment Efficiency:
Doctrine of Chances, in Bernstein, P.L., 1996, Against Journal of Petroleum Technology, May, p. 531-543.
the Gods, the Remarkable Story of Risk: John Wiley & Cardona, G., 1545, Ars Magna, in Bernstein, P.L., 1996,
Sons, New York, 383 p. Against the Gods, the Remarkable Story of Risk: John
Bernoulli, D., 1738, Specimen Theoriae Novae de Mensura Wilev & Sons, l\:e\\' York, 383 p.
Sortis (Exposition of a New Theory on the Measure Chambe;lin, T.e., 1931, The Method of Multiple Working
ment of Risk), in Bernstein, P.L., 1996, Against the Hypotheses: Joumal of Geology, Vol. 39, No.2, p. 155-165.
Gods, the Remarkable Story of Risk: John Wiley & Chorn, L.G., 1999, Matching the Analysis Technique to the
Sons, New York, 383 p. Im"estment Problem: Traditional DCF to Real Options:
Bernoulli, J., 1713, Ars Conjcctandi, in Bernstein, P.L., 1996, Society of Petroleum Engineers #52950, presented at the
Against the Gods, the Remarkable Story of Risk: John SPE Hydrocarbon Economics and Evaluation Symposium,
Wiley & Sons, New York, 383 p. Dallas, March 1999.
_<".Bprn.:tein, ~.l-" 1996, Against the Gods, the Remarkable Clapp, R.V., 1995, An Alternative Concept of Investment for
~lury or J(lsk John Wiley & Sons, New York, 383 p.
Imprm"ed Profitability Measures: SPE 30051, Hydrocarbon
BOCCIa, A., 1996, The Trouble with High Risk Exploration:
Economics & E\'aluation Symposium: Society of
Carlisle Associates, Cambridg(>, Mass., May, p. 1-4.
Petroleum Engineers, Dallas, March 26--28, p. 179-185.
II
155
156 References Cited
cr"pp, R.V., "nd R.D. Stibolt, 1991, Useful Metlsures of Grnunt, J., and W. Petty, 1662, Nntural and Politicnl
Explor"tion Perform"nce: Journal of Petroleum Observa lions Mnde upon the Bills of Mort"lity, in
Technology, October, p. 1252-1257. Bernstein, P.L., 1996, Agninst the Gods, the Remarkilble
Coleman, J.M., and D.B. Prior, 1982, Deltaic Environments, Story of Risk: John Wiley & Sons, New York, 383 p.
in Scholle, P.A., "nd D. Spc"ring, cds., S"ndstone Deposi Grayson, c.J.,. Jr., 1960, Decisions Under Uncertninty,
tional Environments: AAPG Memoir 31, p. 139-178. Drilling Decisions by Oil nnd Gas Operators: Hnrvard
Cozzolino, J.M., 1977, A New Method for Risk Analysis: University Graduate School of Business Administra
Massachusetts Institute of Technology, Sloan tion, Boston, 402 p.
Management Review, Vol. 20, No.3, p. 53-66. Hillley, E., 1693, Trnnsnctions, in Bernstein, P.L., 1996,
Cozzolino, J.M., 1978, A New Method for Meilsurement & Against the Gods, the Remarkable Story of Risk: John /!
C"mtrol' of Exploration Risk: Society of Petroleum Wiley & Sons, New York, 383 p. .
(Engineers of AIME, SPE 6632, M"rch, 37 p. Hnrper, F.G., 1999, BP Prediction Accuracy in Prospect
Cronquist, Chnpmnn, 1997, SPE/WPC Reserves Definitions Assessment: A 15-Yenr Retrospective: preprint of pnper
Approved: Journal of Petroleum Technology, May, p. to be presented ilt AAPG lnternntionnl Conference,
527-528. Birminghnm, Engltlnd.
DilVid, F.N., 19(,'7., Correspondence between Blaise Prlscal Horner, D., 1990, On the Theory of Inevitable Disappoint
and Pierre de Fermat, as reported by David, F. N., in ment, Decision Making Under Risk: Theoretical Note
Games, Gods, and Gnmbling: Hafner Publishing Co., No. 21 (unpublished), 2 p.
New York, p. 229-253. Howell, J.r. Ill, R.N. Anderson, A. Boulanger, nnd B. Bentz,
De Moivre, A., 1733, Doctrine of Chnnces, in Bernstein, P.L., 1998, Mnnaging E&P Assets from a Portfolio
1996, Against the Gods, the Remarbble Story of Risk: Perspective: Oil nnd Gas JournnJ, November 30, p. 54-57.
John Wiley & Sons, New York, 383 p. Hunt, J.M., M.D. Lewen, and R.J.C. Henuet, 1991, Modeling
Demaison, G., 1984, The Generative Basin Concept, in Oil Generation with Time and Tempernture Index
Demaison, G., nnd R.J. Murris, eds., Petrol cum Grilphs Bend on the Arrhenius Equntion: AAPG
Geochemistry nnd Bnsin Evnluntion: AAPG Memoir 35, Bulletin, Vol. 75, No.4, p. 7%-807.
p.I-14. Johns, D.R., S.G. Squire, nnd M.J. Ryans, 1998, Measuring
Demnison, G., nnd B.J. Huizinga, 1994, Genetic Clnssificntion Exploration Performnnce ilnd Improving Exploration from
of Petrolcum Systems Using Three Factors: Chnrge, Predictions-with Exnmples from Santos' Exploration
Migration, nnd Entmpment, in M"goon, L.B., and W.G. Program: Procecdings of 1998 Australinn Petroleum
Dow, eds., The Petroleum System-from Source to Producing nnd Exploration Association, Cilnberra.
Trnp: AAPG Memoir 60, p. 73-89. Johnson, B., 1992, PoJnrity Management: HRD Press,
Dickinson, W.R., 1974, Plate Tectonics nnd Sedimentntion, Amherst, Mass.
in Dickinson, W.R., ed., Tectonics "nd Sediment"tion: Johnston, D., 1997a, Oil Company Finnncial Analvsis in
Society of Economic Paleontologists and ),'lineralogists Nontechnical Language,3rd printing: PennWell'Books,
(SEPM) Specinl Publicntion 22, p. 1-27. TulsCl, akIn., 362 p.
Dixit, A.K., "nd R.S. Pindyck, 1994, Im'estment Under Johnston, D., 1997b, Intemntional Petroleum Fiscill Systems
Uncertainty: Princeton University Press, Princeton, <lnd Production Sharing Contmcts, 3rd pri'nting:
N.J., 468 p. PennWell Books, Tulsn, Okla., 325 p.
Dott, R.H., Sr., and M.J. Reynolds, 1969, Sourcebook for Jones, DR, nnd M. Smith, 1983, Ranking South LouisinnCl
Petroleum Geology: AAPG Memoir 5, ·Fl p. Trends by Probability of Economic Success: Oil and
Downey, M.W., 1992, It All Begins with People, in Steinmetz, Gas Journal, October 24, p. 149-156.
R., ed., TIle Business of Petroleum Explor<ltion: AAPG Kaufman, G.M., 1962, Statistical DE'cision and Related
Treatise of Petroleum Geology-Handbook of Techniques in Oil and Gas Exploration>Prentice-Hn\l,
Petroleum Geology, Ch"pter 16, p. 201-203. New York, 307;,.
Drew, L.J., 1990, Oil and Gas Forecnsting, Ret1ections of " Kaufman, G.M., 1983, Oil and Gns-Estimation of Widespread
Petroleum Geologist: Oxford Unh'ersity Press, New Resources, in Adelman, et <II., eds., Energy Resources in an
York, Oxford, 252 p. UncertClin Future: Bnllinger, Cambridge, Mass., 43-1 p.
Edwards, W., 1982, Conservatism in Human lnfonnation Klemme, H.D., 1975, Giant Oil Fields Related to TIleir Geo
Processing, in Kahnemann, D., P. Slo\'ic, and A. T\'ersky, logic Setting-a Possible Guide to Explomtion: Bulletin
eds., Judgment Under Uncert"inty: Heuristics and of Cilnndian Petroleum Geology, Vol. 23, p. 3(},-66.
Biases: Cambridge University Press, New York, p. Klemme, H.D., 1980, The Geology of Future Petroleum
359-369. Provinces: Revue de l'Institute Francais de Petrole, Vol.
Fisk, H.N., 1954, Sedimentary Framework of the )'lodern 35, No.2, Specinl issue for 25th lnternationtll Geologic<ll
Mississippi Delta: Journal of Sedimentnry Petrology, Congress, Paris, p. 337-349.
Vol. 24, p. 76--99. Klemlile, H.D., and G.F. Ulmishek, 1991, Effective Petro
Formnn, D.J., and A.L. Hinde, 1'186, Exclmination of the ie'lm Source Rocks of the World: Stratigraphic
Creaming Methods of Assessment Applied to the Distribution and Controlling Depositional Fnctors:
Gippsland Basin, Offshore Australia, in Rice, D.D., ed., AAPG Bulletin, Vol. 75, p. 1809-1851.
Oil <lnd G<lS Assessment-Methods and AppJiciltions: Landes, K.K., 1951, Petroleum Geology, 1st edition: John
AAPG Studies in Geology No. 21, p. 101-110. -c. w:......... & Sons, New York, 660 p.
Gehmann, H.N., 1970, Graphs to Deri\'l? Geometric Lehman: .1.,1989, Valuing Oilfield Investments Using
Correction Factor: Exxon Training ),Iaterials (unpub Option Pricing Theory: Society of Petroleum Engineers,
lished), Houston. SPE 18923, p. 125-136. .
References Cited 157
Lerche, r., and J.A. MacKay, ]999, Economic Risk in Moran, S., ] 992, Analysis of International Oil and Gas
Hydrocarbon Exploration: Academic Press, 20 p. Contracts, in Steinmetz, R., ed., The Busitless of
Levorson, A.I., ]967, Geology of Petroleum: W.H. Freeman Petroleum Exploration: AAPG Treatise of Petroleum
& Co., San Francisco, 724 p. Geology-Handbook of Petroleum Geology, Chapter
Lohrenz, J., ]988, Profitabilities on Federal Offshore Oil and 25, p. 307-318.
Gas Leases: A Review: Journal of Petroleum Teclmology, Murris, R.J., 1984, Introduction, in Demaison, G., and RJ.
jWle, p. 760-764. t-,·Iurris, eds., Petroleum Geochemistry and Basin
MacKay, J.A., 1995, Utilizing Risk Tolerance to Optimize haluation: AAPG Memoir 35, p. x-xii.
Working Interest: SPE 30043, Hydrocarbon Economics Murtha, R.J., 1995, Estimating Reserves and Success for a
& Evaluation Symposium: Society of Petro.!p.J.lm .. Prospect with Geologically Dependent Layers(
Engineers, Dallas, p. 103-110. Proceedings of 1995 SPE Hydrocarbon Economics an<!l
MacKay, J.A., 1996, Risk Management in')nternationaJ E\'aluation Symposium, Dallas, p. 73-84.
Petroleum Ventures; Ideas from a Hedberg Conference: Nederlof, M.H., 1979, The Use of Habitat of Oil Models in
AAPG Bulletin, Vol. 80, No. ]2, p. 1845-]849. Exploration Prospect Appraisal: Proceedings from the
Magoon, L.B., 1988, The Petroleum System, A Classification 10th World Petroleum Congress, Panel Discussion ],
Scheme for Research, Exploration and Resource Pnper 2, p. 13-24.
Assessment, in Mngoon, L.B., ed., Petroleum Systems Ncwell, N.D., J.K. Rigby, A.G. Fischer, A.J. Whiteman, J.E.
of the United States: USGS Bulletin 1870, p. 2-]5. Hickox, and J.5. Bradley, 1953, The Permian Reef
Mag?on, L.13., and W.G. Dow, 1994, The Petroleum System, Complex of the Guadalupe Mountains Region, Texas and
1ll Magoon, L.B., and W.G. Dow, eds., The Petroleum New Mexico: W.H. Freeman & Co., San Francisco, 230 p.
System-from Source to Trap: AAPG Memoir 60, Newendorp, P.O., ]975, Decision Analysis for Petroleum
p.3-24. Exploration: Petroleum Publishing Company, Tulsa,
Mann, D., G. Goobie, and L. MacMillian, 1992, Options Okla., 668 p.
Theory and Strategic Investment Decisions: Journal Otis, R.M., and N. Schneidennann, 1997, A Process for
of Canadian Petroleum Technology, Vol. 3], No.5, E\'aluating Exploration Prospects: AAPG Bulletin, Vol.
p.52-55. 8], No.7, p. 1087-1109.
Markowitz, H.M., ] 952, Portfolio Selection: Journal of Owen, E.W., 1975, Trek of the Oil Finders: A History of
Finance, Vol. VII, No. ], March, p. 77-91. Exploration for Petroleum: AAPG Memoir 6, ] 647 p.
McCrossan, R.G., ed., ] 973, The Future Petroleum Paddock, J.L., D.R. Siegel, and J.L. Smith, ]983, Option
Provinces of Canada-Their Geology and Potential: Valuation of Claims on Physical Assets: The Case of
Canadian Society of Petroleum Geologists, Memoir Oiishore Petroleum Leases: MIT Working Paper No.
No.1, 720 p. MIT-EL 83-005WP, p. ]-to.
McMnster, G.E., ] 998, New Trends in Exploration Risk Pascal, 13., and P. de Fermat, ]654, as re!'orted by David, EN.,
Analysis: privately printed paper for limited circulation in Games, Gods, and Gambling: Hafner Publishing Co.,
by Amoco Corporation. New York, p. 229-253.
McMaster, G.E., and P.O. Carragher, ]996, Risk Assessment Pees, S.T., ed., ] 989, History of the Petroleum Industry
and Portfolio Analysis: The Key to'Exploration Success: Symposium~uidebook:AAPG, Tulsa, Okla., 84 p.
privately printed paper for limited circulation by Pickles, E., and J.L. Smith, ]993, Petroleum Propedy
Amoco Corporation. Valuations: A Binomial Lattice Implementation of
Megill, R.E., ]984, 1977, An Introduction to Exploration Risk Option Pricing Theory: The Energy Journal, Vol. 14,
Analysis, 2nd and 1st editions: PennWell Publishing No.2, p. ]-26.
Co., Tulsa, Okla., 273 p. Pray, L.c., 1988, The Western Escnrpment of the Guadalupe
Megill, R.E., 1985, Long-range Exploration Planning: Mountains, Texas, and Day Two of the Field Seminar,
Penn Well Publishing Co., Tulsa, Okla., 90 p. in Tomlinson Reid, S., R.O. Bass, and P. Vvclch, cds.,
Megill, RE., ]988, ]979, ]971, An Introduction to Exp!oration Guadalupe Mountains Revisited-Texas and New
Economics, 3rd, 2nd, and ]st c";i,ions: PennWel! Mexico: West Tex<ls Geologic;}1 Society Publication
Publishing Co., Tulsa, Okla. 88-84, p. 23-31.
Megill, R.E., 1992, Exploration Economics, in Steinmetz, R, Rose, P.R., 1984, Checklist for n Pragmatist, in Cnpen, E.e,
cd., The 13usiness of Petroleum Exploration: AAPG R.E. Megill, and P.R Rose, eds., Notes for AAPG short
Treatise of Petroleum Geology-Handbook of Petroleum course "Prospect Evaluation," Chapter 5: AAPG
Geology, Chapter 10, p. ]07-115. Education Dept., Tulsa, Okla., 2 p.
Megill, R.~., and R.B. Wightman, ]984, The Ubiguitous Rose, P.R., 1987, Dealing with Risk and Uncertainty in
. Overbid: AAPG 13ulletin, Vol. 78, No.4, p. 4] 7-425. Exploration: How Can We Improve?: AAPG Bulletin,
MIall, A.D., ]984, Principles of Sedimentary Basin Analysis: Vol. 7], No. ], p. 1-]6.
. Springer-Verlag, New York, 490 p. Rose, P,R., ] 992a, Chance of Success and Jts Use in
MIlicI', 13.M., ] 986, Resource Appraisal Methods: Choice and Petroleum Exploration, in Steinmetz, R., cd., The
Outcome, in Rice, D.O., cd., Oil and Gas Assessment Business of Petroleum Explomtion: AAPG Treatise of
Methods and Applications: AAPG Studies in Geology Petroleum Geology-Handbook (If Petroleum Geology,
No.2], p. 1-23. Chapter 7, p. 7]-86. -C'. .
Moody, J.D., J.W. Mooney, and J. Spivack, 1970, Giant Oil Rose, P.R, ] 992b, Risk l3eh<lvior in PetroletTlll Ex~iiOmtion,
Fields of North America, in Halbouty, M.T., cd., in Steinmetz, R., cd., The Business of Petroleum
Geology of Giant Petroleum Fields: AAPG Memoir 14, Exploration: AAPG Treatise of Petroleum Geology
p.8-]8.
Handbook of Petroleuill Geology, Chapter 9, p. 95-]04.
158 References Cited
Rose, P.R., 1994, Implementing Risk Analysis in Exploration Steinmetz, R., ed., 1992, The Business of Petroleum
Organizations: What Works? What Doesn't?, Program: Exploration: Treatise of Petroleum Geology-Handbook
AAPG Annual Convention Abstracts, Denver, June of Petroleum Geology: AAPG, Tulsa, Okla., 382 p.
12-15,p.245-246. Stermole, F.J., and J.M. Stermole, 1990, Economic Evaluation
Rose, P.R., 1995, Getting from Geological Chance-factors to and Investment Decisions, 7th edition: Investment
Chance of Economic Success: What Are We Risking?: Evaluations Corp., Golden, CO, 587 p.
AAPG Annual Convention Abstracts, Houston, March Thaler, RH., 1992, The Winner's Curse: Paradoxes and
5-8,p.83A. Anomalies of Economic Life: Free Press, New York,
Rose, P.R., 1996a, Exploration Plays: Risk Analysis and 230 p.
Economic Assessment, Course notes for AAPG short Thompson, R.5., and J.D. Wright, 1985, Oil Pto,P..erty /
course: AAPG Education Dept., Tulsa, Okla., 400 p. Evaluation, 2nd edition: Thompson-Wrigh~ Associates,
Rose, P.R., 1996b, Risk Analysis of Exploration Plays, Golden, Colo., 243 p. i
Abstract: AAPG Bulletin, p. 1331. Tversky, A., and D. Kahnemann, 1974, Judgment Under
Rose, P.R., 1996c, Making Better Estimates of Prospect Uncertainty: Heuristics and Biases: Science, Vol. 185,
Reserves: AAPG Annual Convention Abstracts, San p.1124-1131.
Diego, May 19-22, p. 122A. Uman. M., W.R. James, and H.R. Tomlinson, 1979, Oil and
Rose, P.R., 1997, Risk Analysis of Petroleum Exploration Gas in Offshore Tracts: Before and After Drilling:
Ventures by Modem Corporations: A Report Commissioned Science, Vol. 205, p. 489-491.
by Japan Nntional Oil Corporation, March, 102 p. Vail, P.R., R.M. Mitchum, Jr., and S. Thompson III, 1977a,
Rose, P.R, 1998, Exploration Plays: Risk Analysis and Seismic Stratigraphy and Global Changes of Sea Level,
Economic Assessment: A Report Commissioned by Part Three: Relative Changes of Sea Level from Coastal
Japan National Oil Corporation, February, 103 p. Onlap, in Payton, C.E., ed., Seismic Stratigraphy
Rose, P.R., 1999, Common Methods for Acquiring Applications to Hydrocarbon Exploration: AAPG
Petroleum Rights: Advantages and Disadvantages: Memoir 26, p. 63-82.
Society of Professional Earth Scientists Newsletter, Vail, P.R., R.M. Mitchum, Jr., and S. Thompson III, 1977b,
May, Vol. 37, p. 1-7. Seismic Stratigra'phy and Global Changes of Sea Level,
Rose, P.R., and R.S. Thompson, 1992, Economics and Risk Part Four: Global Cycles of Relative Changes of Sea
Assessment in Development Geology, in Morton Level, in Payton, C.E., ed., Seismic Stratigraphy
Thompson, D., :md A.H. Woods, eds., Development Applications to Hydrocarbon Exploration: AAPG
Geology Reference Manual: AAPG :YIethods in Memoir 26, p. 83-98.
Exploration Series No. 10, p. 1-57. Waples, D.W., 1980, Time and Temperature in Petroleum
Roy, K.J., 1975, Hydrocarbon Assessment Using Subjective Formation: Applications of Lopatin's Methods
Probability and Monte Carlo Methods, in Grenon, M., to Petroleum Exploration: AAPG Bulletin, Vol. 64,
ed., Methods and Models for Assessing Energy p.916-926.
Resources: Proceedings of the First IIASA Conference Waples, D.W., 1994, Maturity Modeling: Thermal
on Energy Resources: International Institute for Indicators-Hydrocarbon Generation and Oil Cracking,
Applied Systems Analysis, Luxemburge, Austria, in Magoon, L.B., and W.G. Dow, eds., The Petroleum
p.279-290. System-From Source to Trap: AAPG Memoir 60, p.
Schuyler, J.R, 1989, Applying Expected I\'Ionetary Value 285-306.
Concept: How Many Prospects is Enough?: Oil and Gas White, D.A., 1980, Assessing Oil and Gas Plays in Facies
Journal, December 11, p. 87-90. . cycle Wedges: AAPG Bulletin, Vol. 64, No.8, p.
Sloss, L.L., 1963, Sequences in the Cratonic Interior of North 1158-1178.
America: Geological Society of America Bulletin, Vol. White, D.A., 1988, Oil and Gas Play Maps in Exploration
74, p. 93-114. and Assessment: AAPG Bulletin, Vol. 72, No.8,
Sluijk, D., and M.H. Nederlof, 1984, Worldwide Geological p.944-949.
Experience as a Systematic Basis for Prospect White, D.A., 1992, Selecting and Assessing Plays,
Appraisal, in Demaison, G., and RJ. :-'Iurris, eds., in Steinmetz, R., ed., The Business of Petroleum
Petroleum Geochemistry and Basin E\'aluation: AAPG Exploration: Tre.-atise of Petroleum Geology-Handbook
Memoir 35, p. 15-26. of Petroleum Geology, Chapter 8, p. 87-94.
Sluijk, D., and J.R. Parker, 1986, Comparison of PredriIling White, D.A., 1993, Geologic Risking Guide for Prospects and
Predictions with PostdriIling Outcomes, Using Shell's Plays: AAPG Bulletin, Vol. 77, No. 12, p. 2048-2061.
Prospect Appraisal System, in Rice, D.D., ed., Oil and Wood, J.E., et aI., 1990, The Domestic Oil and Gas
Gas Assessment-Methods and Applications: AAPG Recoverable Resource Base: Supporting Analysis for
Studie1: in Geology #21, p. 55-58, the National Energy Strategy: Energy Information
Squire, S.(~., 1996, Reservoir and Pool Parameter Administration, Office of Oil and Gas, US Dept. of
Distribu tions from the Cooper / Eromanga Basin: Energy, SR/NES/90-05, 56 p.
Society of Petroleum Engineers, SPE 37365, 9 p. Yergin, D., 1991, The Prize, The Epic Quest for Oil, Money,
St. John, B., 1992, Managing an Exploration Program for and Power: Simon & Schuster, New York, 877 p.
Success, in Steinmetz, R., ed., The Business of
Petroleum Exploration: AAPG Treatise of Petroleum
Geology-Handbook of Petroleum Geology, Chapter
15, p. 187-199.
Index
Amerada, 95
Cash-flow models
Arco/Vastar,95
exploration, 49-50
field; 11,22,69, 70
segmented,50
productive, 17-19
Cash-flow rate of return, discounted (DCFROR), 53-54
Auctions, 98
Chance, 84-85
Bias
Chevron Oil, 58, 95
conservatism, 8
Closure,3,35, 65, 147-148
motivational,28
CNCF. See cumulative after-tax net cash flow
overconfidence, 8
Combining \'entures, 137-145
overoptimism, 8
Commercial success (Pc), 24-25, 32-33, 38-42
prevalent optimistic, 6, 7
Completion success, 24, 33
Black-Scholes model, 52
sanctity of, 99
cost of, 49
industry practices, 120
risk reduction, 93
prospect evaluation, 31-32
159
160 Index
Cost
Estimated ultimate recovery (EUR), 4
capital,49
Estimates, 9-15
dry-hole, 54,138-139
barrels of oil equivalent (BOE), 7
exploration, 85-86, 88
biases under uncertainty, 8
"sunk-cost concept," 40
commercial success (Pc), 39-40
Crystal Ball, 50
economic success (Pe), 39-40
geo.technical parameters, 5 .
,
DCFROR. Sl'l' Discounted cash-flow rate of return
geotechnical uncertainty, 6-8
Discounting, 49-51, 95
uncertainty and, 6-8
Discovery, economic, 40
EV. Sce Expected value
Drilling technology, 3
Expected net present value (ENPV), 54
acquisition strategies, 4
analysis
contract problems, 57
Black-Scholes model, 52
drilling prospect identification, 3
of capital, 49
drilling technology, 3
reinvestment rate, 49
implementing, 121-123
"sensitivity analysis," 50
play analysis, 120-121
software, 50
"prospector myth" versus systematic exploration, 117-11 ')
Economic discovery, 40
Index 161
Field size, 4
prospect resen'es distribution, 24, 25
86,89
GRR. See Growth rate of return
global, 72
Hydrocarbon-recovery factor, 23, 24,127-130
71-77
IE, 5,,<' Investment efficiency
Flowability,33
historical changes in trend success rates, 46-47
Geochemistry, 3
patterns of predictive bias, 45-46
coincidence of, 80
!In'estment efficiency (IE), 54
containment, 36
migration, 35
Latin Hypercube simulation, 24, 127, 137
primary, 35-36
area, 19
recommended, 34-36
a\'erage netpay thickness, 19,20,21,22
virtues of, 38
graphing, 9-11, 118
Geologic reserves, 4
importance of, 11
Geophysics, 3
Low-side estimates, 30,131-135
Estimates; Uncertilinty
acquiring petroleum rights, 93-99
improving, 13-15
biases affecting risk decisions, 91, 92
Jognormality,9-11
criteria indicating competence, 114-115
Global discoveries, 2, 45
optimum working interest, 92-93
Gr<1phiCillmethods,9-11
prospect and play portfolios, 99-107
127-130
risk and risk"'::, crsiuil;"'::11-93
162 Index
Markets, 110
Petroleum rights acquisition, 93-99
Mean
corporate acquisitions, 99
arithmetic, 125
overbidding, 95,97-98
lognormal distribution, 12-13, 125
performance contracts, 98
1)
Migration, 33
Petroleum systems, 58-60, 61
86,89,148
Pl<lY risk analysis, 57-90, 115-116. See nlso Economic
Mobil Oil, 95
chance of play success, 8D-83
Modeling
economic play success, 82-83
Black-Scholes model, 52
geologic chance factors, 34-36, 38, 41-42, 80-82
cash-flow, 49-50,120
criteria indicating competence, 114-115
exploration, 49-50
field-size distributions, 67-80, 86, 89
geotechnical analog, 13
geotechnical data, 84-87
hydrocarbon charge, 65
Spy per BOE of discovered reserves, 86
Motivational bias, 28
secondary (derived) parameters, 86-87
dry holes, 88
Perform<lnce
play models and maps, 87
exploration, 113-115
techniques, field number estima tes, 69-70, of)
forec<lsting, 100
total geology integration, 6-!
Perform<lnce contr<lcts, 98
worldwide versus provincial source r(lcks, 6-1
Index 163
data integration, 62
commercial, 38-42
exploration contracts, 57
corporate system and, 31-32
managing, 107-116
industry experience, 43-48
Portfolios, 99-107
Prospectivity checklist, 3, 65, 147-148
benefits, 99-100
Prospectors, 2, 117-119
estimates, 100-102
Prospect and play portfolios. See Portfolios
management, 1-2,4,99-107
Prospect-resen'es forecasting, 6
model,101 '
Prospect risk analysis, 4. See also Prospect chance of success;
prospect, 99-107
distribution, 24-26
requirements, 99
estimating reserves, 6-8, 17-31,131-135
selection, 120
gross rock volume, 19, 21, 23
Potential reserves, 4
"low-side" values, 131-135
Probability
structural geometry, 131-132, 133, 134
producible reserves, 3
productive area, 17-19
subjective estimates, 36
uncertainty and, 26
RAV. See Risk-adjusted value
Producible reserves, 3
Reality checks, 120
Profitability distribution, 24
Record keeping, 43
Prospect
Reinvestment nite, 49
definition, 57
I~eserves, 4
evaluation, 31-32,137-145
calcula ling, 127-130
identification, 3
Resen'oirrock,3,35,147
\I
164 Index
Reservoir technology, 3
completion, 33
exploration, 117-123
by well class, 45
implementation, 121-123
wildcat, 45, 46
Risk reduction, 93
"Theory of Inevitable Disappointment," 106--107
prospect, 131
and exploration performance, 44-45
Scheduling, 100
Traps (closure), 35, 147-148
Sea-level stands, 64
Truncation, 12, 40, 78-80
"Sensitivity an<llysis," 50
Ubiquitous Overbid, 95, 97-98
Shows, 33
biases affecting judgment under, 8-9, 28-31
Single-value forecilsts, 5
expressing, 6--8
Software
geotechnical, 6-8
Basin-Mod,64
magnitude of, 6
substantial, 6
..
St<ltistics
Unocal,95
bell-shaped curve, 9
U.s. Geological Survey, 58
Stratigraphy, 3, 63-64
success rates by well class, 45