Scribd
Upload
115 views

CS70 ProblemSet12 Template

This document contains a homework assignment with 5 problems related to statistics and probability. It provides the names and student IDs of the students who worked together on the homework. The first problem asks students to perform a statistical hypothesis test on experimental data from the TV show Mythbusters. The second problem explores whether the Law of Large Numbers holds under different network routing protocols. The third problem analyzes vote counts from Palm Beach County in the 2000 US presidential election. The fourth problem defines and calculates properties of random variables. The fifth problem calculates the average time for a character to escape from a cell through different means.

Uploaded by

QFTW
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
115 views

CS70 ProblemSet12 Template

This document contains a homework assignment with 5 problems related to statistics and probability. It provides the names and student IDs of the students who worked together on the homework. The first problem asks students to perform a statistical hypothesis test on experimental data from the TV show Mythbusters. The second problem explores whether the Law of Large Numbers holds under different network routing protocols. The third problem analyzes vote counts from Palm Beach County in the 2000 US presidential election. The fourth problem defines and calculates properties of random variables. The fifth problem calculates the average time for a character to escape from a cell through different means.

Uploaded by

QFTW
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Problem Set 12

Name:

SID:

Spring 2016 GSI:

Sundry Points

Before you start your homework, write down your team. Who else did you work with on this homework? List
names and student IDs. (In case of hw party, you can also just describe the group.) How did you work on
this homework? Working in groups of 3-5 will earn credit for your Sundry grade.
#
1
2
3
4
5

Name(s)
Example Student

SID
00000000

Problem 1: Statistical Hypothesis Testing


On one of the Mythbusters episodes, the Mythbusters decided to run an experiment to test whether toast tends to
land buttered side down.
At the beginning of the episode, Adam and Jamie built a first attempt at a mechanical rig to drop toast in a controlled
fashion. When they tested it on 10 unbuttered pieces of toast as a sanity check, 7 pieces fell upside down and 3
pieces fell right-side up. Adam concluded based upon these numbers that this first rig was obviously biased, so he
threw it away in disgust and they built a new rig. Was Adam right, or is this just another case where he jumps to
conclusions too quickly?
Let p denote the probability that, if we drop 10 pieces of unbuttered toast from an unbiased rig (i.e., a rig where
each unbuttered piece of toast has a 50% chance of falling upside down and a 50% chance of falling right-side up),
7 or more of the pieces of toast land the same way. In other words, p is the probability of the event that at least 7
pieces land right-side up, or at least 7 pieces land upside down, when dropping from an unbiased rig.
a. As a warmup, compute the exact probability that if we flip a fair coin 10 times, we see 0, 1, 2, 3, 7, 8, 9, or 10
heads.
Solution
b. Now, back to the Mythbusters. With p defined as above, calculate p exactly.
Solution
c. Use p to decide whether the rig appears biased, using the following rules:
If p > 0.05, conclude that we cannot rule out the possibility that the rig is unbiased. The rig might be
perfectly good as it is. (The intuition is: Oh man, that totally couldve happened by chance.)
If p 6 0.05, with 95% confidence we can conclude that the rig appears to be biased. (Sure, its possible
that this rule could lead us astray. Even if our calculations show p 6 0.05, its in principle possible that
the rig is unbiased and the observations were just a big coincidence. However, this would require assuming
that an event of probability 0.05 or less happened, which is by definition pretty rare. Put another way, if
we conclude that the rig is biased whenever p 6 0.05, then well wrongly throw away a perfectly good rig
at most 5% of the time. This seems good enough.)
To put it another way, this decision rule gives us a way to test the hypothesis that the rig is unbiased: if
p 6 0.05, we reject the hypothesis (with 95% confidence level).
Using your value of p and this decision rule, decide whether Adam was right to conclude that his first rig was
biased, or whether he jumped to conclusions too quickly.
Solution

Problem 2: Law of Large Numbers


Recall that the Law of Large Numbers holds if, for every  > 0,




1

1
lim Pr Sn E
Sn >  = 0
n
n
n
In class, we saw that the Law of Large Numbers holds for Sn = X1 + . . . + Xn , where the Xi s are i.i.d. random
variables. This problem explores if the Law of Large Numbers holds under other circumstances.
Packets are sent from a source to a destination node over the Internet. Each packet is sent on a certain route, and
the routes are disjoint. Each route has a failure probability of p and different routes fail independently. If a route
fails, all packets sent along that route are lost. You can assume that the routing protocol has no knowledge of which
route fails.
For each of the following routing protocols, determine whether the Law of Large Numbers holds when Sn is defined
as the total number of received packets out of n packets sent. Answer Yes if the Law of Large Number holds, or No
if not, and give a brief justification of your answer. (Whenever convenient, you can assume that n is even.)
a. Yes or No: Each packet is sent on a completely different route.
Solution
b. Yes or No: The packets are split into n2 pairs of packets. Each pair is sent togehter on its own route (i.e.,
different pairs are sent on different routes).
Solution
c. Yes or No: The packets are split into 2 groups of n2 packets. All the packets in each group are sent on the
same route, and the two groups are sent on different routes.
Solution
d. Yes or No: All the packets are sent on one route.
Solution

Problem 3: Those 3407 Votes


In the aftermath of the 2000 US Presidential Election, many people have claimed that unusually large number of
votes cast for Pat Buchanan in Palm Beach County are statistically highly significant, and thus of dubious validity.
In this problem, we will examine this claim from a statistical viewpoint.
The total percentage votes cast for each presidential candidate in the entire state of Florida were as follows:
Gore
48.8%

Bush
48.9%

Buchanan
0.3%

Nader
1.6%

Browne
0.3%

Others
0.1%

In Palm Beach County, the actual votes cast (before the recounts began) were as follows:
Gore
268945

Bush
152846

Buchanan
3407

Nader
5564

Browne
743

Others
781

Total
432286

To model this situation probabilistically, we need to make some assumptions. Lets model the vote cast by each voter
in Palm Beach County as a random variable Xi , where Xi takes on each of the six possible values (five candidates
of Others) with the probabilities corresponding to the Florida percentages. (This, e.g., Pr[Xi = Gore] = 0.4888.)
There are a total of n = 432286 voters, and their votes are assumed to be mutually independent. Let the r.v.
B denote the total votes cast for Buchanan in Palm Beach County (i.e., the number of votes i for which Xi =
Buchanan).
a. Compute the expectation E[B] and the variance Var(B).
Solution
b. Use Chebyshevs inequality to compute an upper bound b on the probability that Buchanan receives at least
3407 votes, i.e., find a number b such that
Pr[B > 3407] 6 b
Based on this result, do you think Buchanans vote is significant?
Solution
c. Suppose that your bound b in part (b) is exactly accurate, i.e., assume that Pr[X > 3407] is exactly equal
to b. (In fact the true value of this probability is much smaller). Suppose also that all 67 counties in Florida
have the same number of voters as Palm Beach County, and that all behave independently according to the
same statistical model as Palm Beach County. What is the probability that in at least one of the counties,
Buchanan receives at least 3407 votes? How would this affect your judgment as to whether the Palm Beach
tally is significant?
Solution

Problem 4: Umbrella Store


Let = {1, 2, 3, 4, 5, 6} be a uniform probability space. Let also X() and Y (), for , be the random variables
defines as follows:
Table 1: The

X()
Y ()

random
1 2
0 0
0 2

a. Calculate V = L[Y | X]
Solution
b. Calculate W = E[Y | X]
Solution
c. Calculate E[(Y V )2 ]
Solution
d. Calculate E[(Y W )2 ]
Solution

variables X
3 4 5
1 1 2
3 5 2

and Y .
6
2
0

Problem 5: Casino Wins


James Bond is imprisoned in a cell from which there are three possible ways to escape: an air-conditioning duct, a
sewer pipe and the door (which is unlocked). The air-conditioning duct leads him on a two-hour trip whereupon he
falls through a trap door onto his head, much to the amusement of his captors. The sewer pipe is similar but takes
five hours to traverse. Each fall produces amnesia and he is returned to the cell immediately after each fall. Assume
that he always immediately chooses one of the three exits from the cell with probability 13 . On average, how long
does it take before he opens the unlocked door and escapes?
Solution

You might also like