Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Intro to Probability and Statistics - Exercise 3

served by Mickey Frankel, ID 206335812

November 20, 2022

Question 1
a) We'll calculate the probability of Ace not being pulled for 13 pulls
in a row, and multiply by 52−13 4 4
= 39 . The odds for ace not being
picked are: 52 · 51 . . . 40 = 35! · 52! , so we have subtotal of 394 · 48!
48 47 36 48! 39! 39!
35! · 52! =
36·37·38
49·50·51·52 ·4

b) The odds of not getting an Ace in rst 13 cards is 1 − 48!·38! 52!·35! , the
odds to get a single ace in rst 13 drawings are 1 − ( 52!·36!
49!·38!
), the odds
52!·37! ) and the odds to
to get two aces in rst 13 drawings are 1 − ( 50!·38!
get three aces in rst 13 drawings are 1 − ( 52!·38! ). The subtotal is
49!·38!
49!·38! 4
(1 − 52!·38! ) · 39 + (1 − ( 49!·38!
52!·36! )) ·
3
39
50!·38!
+ (1 − ( 52!·37! )) · 2
39 + (1 − ( 49!·38!
52!·38! )) ·
1
39

Question 2
a) =⇒ Let us assume that a Tournament is transitive and we'll prove
there is a single arrangement where each participant defeated another
i he is premier in the arrangement. We will build such arrangement:
The Tournament is transitive so for each given triplet wα , wβ , wγ , if
wα defeated wβ and wβ defeated wγ , then wα defeated wγ . Let there
be some arrangement A, so we will traverse A's elements from 1 to
continue left defeated right
(
n with bubble sort algorithm, with λ :
swap right defeated left
as sorting function. We are guaranteed to get a sorted set of par-
ticipants, hence there exists such arrangement where a participant
is before the other one i he defeated him. Let us assume there
are several such arrangements, then besides the given arrangement
{w1 , w2 , w3 . . . wα , wβ , wγ . . . }, there is another arrangement which holds
the given criteria w1 , w2 , w3 . . . wβ , wα , wj . . .. But according to given ar-
rangement, wα defeated wβ , in contradiction to what other arrange-
ment represents, that wβ defeated wα which is impossible due to trichotomy.
contradiction! hence the assumption is wrong and there is only a sin-
gle such arrangement.

1
⇐= Let us assume there is a single arrangement where each partici-
pant defeated another i he is premier in the arrangement and we'll
prove that a Tournament is transitive. Let us denote this arrangement
as A. Let us assume the tournament is not transitive, meaning that
for some triplet, even though if wα defeated wβ and wβ defeated wγ ,
nevertheless wγ defeated wα . But wα defeated wβ , so wα appears in
A before wβ , and so for wβ and wγ . So wα appears in A before wγ ,
so according to A's deniton's, wα defeated wγ , in contradiction to the
assumption on the existence of such triplet. contradiction! hence
this assumption is wrong and necessarily there is no such triplet, e.g.
the tournament is transitive.
b) For a tournament to be transitive it is needed to exist a correlative
arrangement which we have shown in previous subquestion to be
single. However there are n! − 1 ways to present a non transitive
tournament, so the probability of having a transitive tournament is
n!−1 .
1

Question 3
We will use the Boole's inequation - the general probabilty for two
numbers to be chosen is m1 , the probability for two children to
2

choose the very same pair of numbers is ( m1 )2 = m1 . There are k


2 4

children in class hence the subtotal is k2 · m1 = k(k−1) 4 2m So P(∪i,j∈[mk] ) ≤


4

, where is the event where the 'th and j 'th kids


Pi<j≤m
i=1 P(E i,j ) Ei,j i
chose the same m. So that equals 2 = 2 = 2·m = m 2·m
α α α
k
 m
 k·(k−1) (m −1)
4 = 4

and clearly, lim inf m−>inf = inf ,


α−4 α 2α−4 α−4 2α−4 α−4
m (m −1) m −m m −m
2 = 2 2
which is clearly a bad boundary. So we'll use another approach:
for one of the√ 2
expressions we've got, k·(k−1)
2·m , it holds that for k close
4

enough to 2m we get a value which is close to one. But that's not


sucient as m approaches innity, so we'll try to use continuity of
probability function, notice that Ek,m is an ascending series of events
(if 2 kids in a class of k chose the same numbers, if another kid will
be added, this fact won't be changed), thus, the continuity holds. So
= 1 .
α α 2α α
−1)
P(∪i,j∈[mk] ) = limm→∞ P(Ek,m ) = k(k−1) 4
2m = m (m2m 4= m 2m −m
|{z}4

n→∞

Question 4
Let us examine dependency (U for Unfair coin, F for Fair coin): the
cases of A and B are {{F,F},{F,U},{U,F},{U,U}}, while P ({H, H}) =
0.5· 23 2
P (H|U ) + P (H|F ) = ( P (H∩U ) 2 P (H∩F ) 2
P (U ) ) + ( P (F ) ) = (
0.5·0.5 2
0.5 ) + ( 0.5 ) = ( 19 + 16
1

2
4= . The experiments will be A-choice of a coin, B-results, we'll
25
36
check if there is a dependency:
4 1
9 ·
 4 P ({H, H|U, U })
 29 · 1

P ({H, T |U, U }, P ({T, H|U, U })


 4
1 ·
 1
P ({T, T |U, U })
P (B|A) : 19 4
1
4 · P ({H, H|F, F }), P ({T, H|F, F }), P ({T, H|F, F }), P ({T, T |F, F })

 4

2 1 1
3 · · P (H, H|U, F ), P (H, H|F, U ), P (H, T |F, U ), P (T, H|F, U )


4 4



1 1 1
3 · 4 · 4 P (T, H|U, F ), P (T, T |U, F ), P (T, T |F, U ), P (H, T |F, U )

It is clear that choosing dierent coins guarantees independcy, hence


the experiments B and C are multiplication spaces. For experiment A
the odds to get twice Heads are 32 · 12 · = 13 , the odds to get Heads rst
time are 0.5·0.5+0.5· 23 = 127 , and odds for second time are the same. For
experiment B the odds to get twice Heads are ( 23 · 12 )· 12 +( 12 )· 14 +( 49 )· 14 =
72 , the odds to get Heads rst time are 0.5 · 0.5 + 0.5 · 3 = 12 , and odds
29 2 7

for second time are the same. For experiment B the odds to get twice
Heads are ( 94 ) · 21 + ( 14 ) · 21 = 25
72 , the odds to get Heads rst time are
7
0.5 · 0.5 + 0.5 · 23 = 12 , and odds for second time are the same.

You might also like