Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Whole Number Multiplication: There's More to It Than Might Be Expected!

Author(s): Christy D. Graybeal, Al Cuoco and E. Paul Goldenberg


Source: The Mathematics Teacher , NOVEMBER 2007, Vol. 101, No. 4, Mathematical
Discourse (NOVEMBER 2007), pp. 312-316
Published by: National Council of Teachers of Mathematics

Stable URL: https://www.jstor.org/stable/20876117

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

National Council of Teachers of Mathematics is collaborating with JSTOR to digitize, preserve


and extend access to The Mathematics Teacher

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms
Christy D. Graybeal

Whole Number Multiplication:


There's More to It Than
Might Be Expected!
The problem: Arrange the digits 5, 6, 7, 8, and 9 different possibilities, I quickly realized that I did
into a three-digit number and a two-digit num not know what the largest product was or even an
ber so that.their product is as large as possible. elegant way of determining it. My approach to the
problem would have been very similar to my stu
efore you read any further, write down dent's. I began thinking that there must be a more
your prediction. Did you guess 987 x 65 or elegant method.
maybe 975 x 86? Do not feel embarrassed? Further, I was surprised by the answer given in
those were my first guesses as well! Although it is the back of the book: 875 x 96. As is often the case
not immediately obvious, there is more to this prob in mathematics, the answer generated more ques
lem than might at first be expected. tions: Why is it better to have the largest digit in
the tens place of the second number instead of in
MOTIVATION the hundreds place of the first number? What is the
My interest in this problem arose while I was tutor reasoning behind the placement of the digits? In the
ing a seventh grader. We were flipping through the spirit of Brown and Walter (1993), I used this prob
book Math Mind Stretchers (Fisher 1997) and came lem to pose new questions as well: What if the digits
across this and similar problems. My student began are not consecutive? What if all the digits are not dis
by making educated guesses and finding the prod tinct? Does this pattern hold for any five digits? The
ucts. Just as I would have, he placed the largest digit more I explored these questions, the more I learned
(9) in the hundreds place. As he worked though about whole-number multiplication. At the same
time, I also learned about the importance of repre
This department focuses on mathematics content that appeals to secondary school sentations. Different representations of this problem
teachers. It provides a forum that allows classroom teachers to share their mathemat led me to different insights and understandings.
ics from their work with students, their classroom investigations and projects, and their
other experiences. We encourage submissions that pose and solve a novel or interest SOLUTION STRATEGIES
ing mathematics problem, expand on connections among different mathematical topics, Because I was working with a prealgebra middle
present a general method for describing a mathematical notion or solving a class of prob school student when I became interested in this
lems, elaborate on new insights into familiar secondary school mathematics, or leave the
problem, I initially tried to solve the problem as
reader with a mathematical idea to expand. Send submissions to "Delving Deeper" by
accessing mt.msubmit.net. my student did, using the guess-and-check method.
"Delving Deeper" can accept manuscripts in ASCII or Word formats only. I selected five distinct digits and tried different
arrangements until I found what I believed was the
Edited by Al Cuoco, acuoco@edc.org largest product. The difficulty with this approach
Center for Mathematics Education, Education Development Center is knowing whether I had in fact obtained the larg
Newton, MA 02458
est product. Several times I had what I thought was
a winner, only to find a better arrangement later.
E. Paul Goldenberg, pgoldenberg@edc.org
Center for Mathematics Education, Education Development Center Without trying all 120 possible arrangements of
Newton, MA 02458 five distinct digits, could I guarantee that the prod
uct I obtained was the largest possible?

312 MATHEMATICS TEACHER | Vol. 101, No. 4 November 2007

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms
2nd 3rd 5th dor e
largest largest largest

largest 4th dor e


largest

Fig. 1 This arrangement seemed to produce the largest Fig. 2 The largest digits must be placed as shown.
product.

As I experimented with various sets of five digits, d or e b or c


a pattern emerged. For five digits, it seemed that the
arrangement in figure 1 resulted in the largest prod
uct. With conjecture in hand, I set out to prove this.
Along the way I gained new insights into the multipli
cation algorithm and multiplication itself. By examin
ing the algorithm, I noticed how different arrange
dor e b or c
ments affect the product. Switching the position of
two digits results in gains in some places and losses in
others; by maximizing gains and minimizing losses,
one can increase the product. Early in my work I
realized why my initial guesses did not result Fig.
in the
3 The next two largest digits must be placed as shown.
largest product; later, I found that the arrangement in
figure 1 does not always result in the largest Although
product. I worked in this approach informally, writ
Two representations of the problem (informal and
ing it up is easiest with formal symbolism.
formal) are presented, each offering different insights
and each having its own benefits and drawbacks. Consider the digits a, b, c, d, and e where 0 < a <
The informal approach helped me see what was b<c<d<e<9.
happening in the multiplication algorithm and why
my conjectured arrangement usually resulted in the
Although I originally thought of these digits
largest product. However, it did not satisfy my intrin
as distinct, one of my questions was about what
sic need for formal proof, nor did it, at first, would happen if some digits repeated. Thus, I did
help me
see when or why there is an exception to thenot arrange
use a strict inequality in the problem setup.
ment. (In fact, I did not realize that there was To
an solve the initial problem, we want to maxi
exception until I considered the more formalmize
proof.)the thousands, then the hundreds, and so on.
The more formal approach satisfies my need To maximize
for the number of thousands, the two
proof, and makes the special case more obvious but digits, d and e, should be in the leftmost
largest
does not provide as much insight into the multiplica
positions of both the three-digit number and the
tion algorithm. The symbol manipulations, at two-digit
times, number (see fig. 2). To maximize the
mask the quantities involved. For me, this proof
number is of hundreds, the next two largest digits, b
what Hanna (1989) would consider a "proofand
that
c, should be in the tens place of the three-digit
proves" rather than a "proof that explains." number and the ones place of the two-digit number
(see fig. 3).
INFORMAL APPROACH By default, the smallest digit, a, should be in the
For five distinct digits, there are 5! = 120 possible
ones place of the three-digit number. This results in
arrangements. But it is already clear, even tofour possibilities (see fig. 4).
a young
In summary:
student, that one does not need to investigate all 120
possibilities. Common sense (more precisely, place
value and even a beginner's understanding ofCasethe1: (lOOd + 101? + a){l0e + c) = lOOOd* +
multiplication algorithm) eliminates many of these.100{be + cd) + \0{ae + be) + ac
Case 2: (lOOd + 10c + a)(10c + b) = lOOOdc +
Visual inspection also helps suggest what compari
sons need to be made to refine the initial guesses. 100{bd + ce) + 10[ae + be) + ab

Vol. 101, No. 4 November 2007 | MATHEMATICS TEACHER 313

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms
Case 1 Case 2
c

Case 3 Case 4
c

Fig. 4 After placing a, there are four cases.

Case 3: (100* + 101? + a) (lOd + *) = lOOOd*


Given b < c+ implies ab < ac.
100(bd + **) + \0{ad + I?*) + ac So Case 1 and Case 3 have at least as many ones as
Case 2 and
Case 4: (100* + 10* + a)(10d + b) = 1000d* + Case 4.
100(1?* + cd) + \0{ad + &*) + ab
In summary:
All four cases have the same number (de) repre
senting the thousands. All four arrangements have the same number of
To compare the number of hundreds,thousands.
I needed
Case
to determine whether (bd + **) or {be + 2cd)
and is
Case 3 have at least as many hundreds
greater. as Case 1 and Case 4.
Case 1 and Case 2 have at least as many tens as
Given d<e,b<c^> Case 3 and Case 4.
(* - b)d < (* - b)e^ Case 1 and Case 3 have at least as many ones as
cd -bd< ce - be ^ Case 2 and Case 4.
be + cd<bd + ce.
So Case 2 and Case 3 have at least as many hun At first glance it seems that since Case 2 has
dreds as Case 1 and Case 4. the greatest number of hundreds and tens, it will
always result in the largest product. The exception
To compare the number of tens, I needed to to this did not become obvious until I developed a
determine whether {ad + be) or [ae + be) is greater. more formal proof.

Given d < e, FORMAL PROOF


ad<ae-> To determine with greater certainty which of these
ad + bc<ae + be. four cases results in the largest product, we will
compare them in pairs.
So Case 1 and Case 2 have at least as many tens as
Case 3 and Case 4. Comparison of Case 1 with Case 2
To compare the number of ones, I needed to Subtract Case 1 from Case 2 and consider the sign
determine whether ab or ac is greater. of the difference. Case 2 - Case 1:

314 MATHEMATICS TEACHER | Vol. 101, No. 4 November 2007

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms
[lOOOdc + 100(fod + ce) + lO(ae + be) + ab] [lOOOde + 100(bd + ce) + 10(W + be) + ab]
- [lOOOde + 100(foe + cd) + 10(W+ foe) + ac\ - [lOOOde? + 100(be + cd) + 10(ad + fc) + afc]
= 100(bd + cc) - 100(foe + cd) + ab - ac = 100(fe2 + ce) - 100(fe + cd) + 10(a<? + for) -
= 100(fod + ce - foe - cd) + a(fo - c) \0{ad + be)
= 100(b - c)(d - e) + a(b - c) = 100(bd + ce-be- cd) + 10(ae - ad)
= (fo-c)[100(d-e) + a] = 100(c - b)[e - d) + 10a(^ - d)
= 10(e-d)[10(c-b)+a]
The following lemma will be useful in determining
the sign of the difference. Lemma 3. [I0(c - b) + a] is nonnegative.

Lemma 1. If d < e, then [I00(d -e)+a] is negative. Proof. Given b<c,we have 0 < c - b. So 0 <
If d = e, then [100(d -e)+a] is 10(c - b). Because a is not negative, [I0(c - b) + a]
nonnegative. is nonnegative.
If d < e, then (^ - d) is positive. Thus, 10(e -
Proof. If d < e, then d - e < -1, and 100(d - e) < d) [10(c - b) + a] is nonnegative and Case 4 < Case 2.
-100, which makes [I00(d - e) + a] negative. If If d = ?, then Case 2 = Case 4.
d = e, then 0 = 100(d - e) and [100(d - e) + a] is
nonnegative. Summary of the three comparisons
\td<e, then Case 1 < Case 2, Case 3 < Case 2, and
Since fo < c, (fo - c) is nonpositive. Case 4 < Case 2. Thus, Case 2 results in the largest
If d < e, then Case 2 - Case 1 is nonnegative and product.
Case 1 < Case 2. If d - e, then Case 2 < Case 1, Case 2 < Case 3,
If d = e, then Case 2 - Case 1 is nonpositive and and Case 2 = Case 4. But d = e also makes Cases
Case 1 > Case 2. 1 and 3 indistinguishable. Thus, Case 1 (Case 3)
Comparison of Case 2 to Cases 3 and 4 follows a results in the largest product.
similar plan. An example shows the significance of this result.

Comparison of Case 2 with Case 3 Example


Subtract Case 3 from Case 2 and consider the sign Consider the digits 1, 3, 6, 7, and 9. Since the two
of the difference. Case 2 - Case 3: largest digits are not equal, Case 2 will result in the
largest product:
[lOOOde + 100(fod + ce) + 10(oe + be) + ab]
- [lOOOde + 100(fod + ce) + 10(ad + be) + ac] 761 x 93 = 70,773
= 10(ae + be) - 10(ad + be) +ab- ac
= 10(ae + bc-ad- be) + a(b - c) Now if the 7 is changed to another 9, one might
= lOGze - ad) + a(fo - c) think that the largest product would be
= a[\0{e - d) + (fo - c)]
961 x 93 = 89,373.
Lemma 2. If d < e, then [10(e -d) + (b- c)] is
positive. However, Case 1 = Case 3 results in a larger product:
Tfd = e, then [I0(e - d) + (fo - c)] is
nonpositive. 931 x96 = 89,376

Proof. If d < e, then e - d > 1 and 10(e - d) > 10. Originally, this surprised me. How could 3 more
Since 0<fo<c<9, -9<fo-c<0. Thus, [I0(e - d) + (96 - 93 = 3) of 30 less (961 - 931 = 30) result in a larger
(fo - c)] is positive. If d = e, then 10(e - d) = 0. Since product? My informal approach helped me see why.
fo < c, fo - c is nonpositive and [10(e - d) + (fo - c)] is
nonpositive. Call 961 x 93 = 89,373 Case 2.
Call 931 x 96 = 89,376 Case 1.
If d < e, then Case 2 - Case 3 is nonnegative and Both cases have 9 x 9 = 81 thousands.
Case 3 < Case 2. Both have 6x9 + 3x9 = 81 hundreds.
If d = e, then Case 2 - Case 3 is nonpositive and Both have 1x9 + 3x6 = 27 tens.
Case 3 > Case 2.
The difference is in the number of ones:
Comparison of Case 2 with Case 4 Case 1 has 1x6 = 6 ones.
Again, we subtract Case 4 from Case 2 and consider Case 2 has 1x3 = 3 ones.
the sign of the difference. Case 2 - Case 4: Thus, Case 1 results in a larger product than Case 2.

Vol. 101, No. 4 November 2007 | MATHEMATICS TEACHER 315

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms
CONCLUDING THOUGHTS of the work. Of course, what made the extensions
My favorite mathematics problems are ones that seem like an endless morass is that all the problem
are deceptively simple?easy to describe and extensions we had just listed?six digits, seven,
understand yet complex. The generalization of this thirty-three, three factors instead of two?were
problem to any five digits fits this description. As special cases, just like the original problem. We had
I worked on this problem, I talked to mathemati not really stopped to ask what they were special
cians, mathematics educators, and mathematics cases of.
teachers. All have been intrigued by it. Several The author has highlighted some insights into
have asked me how I think the author of the prob multiplication that she gained while analyzing this
lem expected middle school students to solve the discrete optimization problem. The generalization
original problem and what they would get out of that comes directly from the author's analysis is
it. Although I doubt that a middle school student this: Given n digits, how can we arrange them into
would approach the problem as I did, I think the m factors to produce the greatest (lowest) product?
problem is still a good one. It seems perfectly fair to fix n or m for the analysis,
For one thing, by considering place value and but fixing both would make it too much of a special
the multiplication algorithm, students can make case again. But the author's focus on optimization
educated guesses about the placement of the digits. in the context of elementary arithmetic reminded
Through logic, students will realize that they do not us of a problem in an old elementary textbook,
need to check all 120 possibilities. Moreover, stu Math Workshop, by Robert Wirtz, Morton Botel,
dents can gain insights from the informal and for Max Beberman, and W. W. Sawyer (Chicago:
mal approaches and appreciate the value in looking Encyclopaedia Britannica Press, 1964): Given
at a single problem in multiple ways. Generalizing some whole number, like 12, find a way to parti
the problem to any five digits provides a good les tion it?say (6, 6) or (3, 4, 5) or (1, 1, 1, 9)?that
son in the importance of paying attention to detail gives a set of numbers with the maximum product.
and being open to surprises. Without these disposi The products for the examples are 36, 60, and 9,
tions, it is easy to overlook the exception of what respectively, so (3, 4, 5) seems the best set, but
happens when the two largest digits are equal. perhaps there is one even better. However, finding
I encourage teachers to help their students the maximum product for partitions of 12 is not
explore this and similar problems. For example, interesting. What can we say in general about such
students can explore what happens if the goal is to a partitioning for n? What if we are not restricted
obtain the smallest product. (The pattern is slightly to integers as we look for numbers whose sum is n?
different.) How can seven digits be arranged into The ideas that arise from this investigation touch
a four-digit number and a three-digit number so mathematics that students in the final years of high
that their product is as large as possible? Is the pat school are just beginning to approach.
tern similar to the pattern for five digits? Are there
any surprises? By extending the problem, students REFERENCES
learn an important lesson in mathematics: There is Brown, Stephen L, and Marion I. Walter. Problem
always more to learn and explore. Posing: Reflections and Applications. Hillsdale, NJ:
Lawrence Erlbaum, 1993.
Editors' note: At first, the author's suggestion to Fisher, Ann. Math Mind Stretchers. Grand Rapids, MI:
explore the pattern for seven digits caught us by Instructional Fair, 1997.
surprise. Why not ask first about six digits? How Hanna, Gila. "Proofs That Prove and Proofs That
much more are we learning as we ask for more Explain." In Proceedings of the International Group
and more digits? Even after we solve the problem for the Psgchologg of Mathematics Education, Vol
for seven digits, we still do not know what might ume II, edited by G. Vergnaud, J. Rogalski, and M.
happen with nine?or thirty-three, for that mat Artigue, pp. 45-51. Paris, 1989. ~
ter! Moreover, with seven digits, do we want to
restrict ourselves to thinking about the product
of a three-digit number multiplied by a four-digit
one, or might a still greater (or smaller) product
be achieved through a different use of those seven
digits? With seven digits, we might, instead, con
CHRISTY D. GRAYBEAL, cgraybea@umd.edu, is ^^^H
a former middle school mathematics teacher. ^^^^H
struct the product of three numbers?one three
She is currently a doctoral candidate in curricu- ^^^^H
digit number and two two-digit numbers?and
lum and instruction with an emphasis on mathe- ^^^^H
that might produce the greatest product. We were
matics education at the University of Maryland, ^^^^H
about to lose interest in this initially very intrigu
ing problem because of the apparent endlessness
College Park, MD 20742. ^^^H

316 MATHEMATICS TEACHER | Vol. 101, No. 4 November 2007

This content downloaded from


146.200.146.246 on Mon, 01 Jul 2024 05:20:16 +00:00
All use subject to https://about.jstor.org/terms

You might also like