Math 308/361: A Sampler of Problems

http://www.math.uiuc.edu/~hildebr/361/sampler.html

Here is a preview of some of the problems that can be treated with the methods developed in this course. Some of the problems will be discussed in class, while others may show up as homework problems. The goal of the course is to develop tools to solve problems of this sort theoretically. Many of these problems lend themselves to computer simulations, and is of interest to come up with (approximate) answers via such simulations and compare these answers with those derived by theoretical methods. In some cases ready-made java applets are available to carry out these simulations online; I have provided links to such applets. If you are curious, try one of these simulations, or write your own programs.

Simulating the Florida election

In the Florida election, out of some 6 million votes, the number of votes for Bush and Gore were within a few thousand of each other. This raises the following interesting probability question: If each of the 6 million votes had been cast by tossing a coin, would it be reasonable to expect the number of heads and tails to differ by no more than a few thousand? For example, what is the probability that in 6 million coin tosses the number of heads and tails differ by at most 3,000? What are the corresponding probabilities if 3,000 is replaced by 300, or by 30? What about the extreme case, when the difference between heads and tails is zero? In other words, what is the probability that out of 6,000,000 coin tosses exactly half (i.e., 3,000,000) end up heads?

Computer simulation: It is very easy to write a program to simulate repeated coin tosses. If you are so inclined, do that to get a feel of what to expect.

Polling

Continuing with the election theme, here is problem related to polling: A polling firm wants to determine the proportion of likely voters who favor the incumbent candidate in an upcoming election. How many voters does it have to poll so that, with 90 % probability, the polling error is at most 2 percentage points?

De Mere's paradox

Which is more likely, getting at least one six in 4 rolls of a die (Game I), or getting at least one double six in 24 rolls of 2 dice (Game II)? The Chevalier de Mere (1610--1685) argued (incorrectly, as it turned out) that both probabilities are the same (and equal to 2/3), since 4 times 1/6 is 2/3 and 24 times 1/36 is also 2/3. What are the correct probabilities for winning in these two games? Which (if any) of the two games gives you a better than 50-50 chance of winning?

Computer simulations: Java applets with computer simulations of both games are available, for both games, Game I, and Game II (courtesy of the Chance Website) Try to determine the probabilities in question experimentally by playing these games a sufficient number of times, and try to determine which of the two probabilities (if any) is greater. Alternatively, write your own program to simulate the two games.

The Birthday Paradox

A rather well-known (apparent) paradox is that a relatively small group of people is sufficient to ensure, with very high probability, that there are at least two members of the group who have the same birthday. For example, it takes only 23 people to have a better than 50-50 chance of finding a matching birthday. A java applet with nice graphics simulating this birthday paradox is available here. (This is part of the Probability by Surprise collection of java applets.) A different java applet, which computes the probabilities for finding a matching birthday in a group of n people, can be found here (again from the Chance Website).

Related to that paradox is the following problem, suggested by a reader question in the ``Ask Marilyn'' column in Parade Magazine, January 11, 1998. [From the "Chance News" section of the Chance Website; "Chance News" is a monthly compilation of newspaper and magazine articles that raise interesting probabilistic or statistical questions.

I started counting from the first person who came came into the office. I counted until I found a matching birthday in the group. Then I started a new survey with the next person. In eight surveys, the smallest number of people it took before I found a matching birthday was 12. The largest number was only 54.
Why is a minimum of 12 not surprising? Specifically, what is the probability that there is no match in the first 12 people in 8 surveys?

Baseball World Series

In the baseball world series, two teams play against each other. The first team to win 4 games is declared the winner. Thus, a world series lasts at most 7 games and at least 4 games. Suppose A is the better team and has a 2/3 chance of winning a game against B. What is the probability that the series is over in 6 or fewer games?

Airline overbooking

Suppose an airline accepted 12 reservations for a small commuter plane with 10 seats. They know that 7 reservations went to regular commuters who will show up for sure. The other 5 passengers will show up with a 50 % chance, independently of each other. (a) Find the probability that the flight will be overbooked; (b) Find the probability that there will be empty seats; (c) Find the average number of passengers turned away.

Records

A die is rolled repeatedly. Saying that a record occurs at roll n means that the number appearing on the n-th roll is strictly greater than the number appearing on all previous rolls. The number appearing on the first roll is, by default, considered to be a record. (a) What is the (numerical) probability that a record occurs at the 4th roll? (b) Find a general formula (which may involve a summation) for the probability that a record occurs at the n-th roll, for n=1,2,...
Math 361 Course Web Page
Math 308 Course Web Page

Last modified Tue 21 Jan 2003 09:11:16 AM CST