Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Here's a question I got for homework:

In every single time unit, Jack and John are tossing two different coins with P1 and P2 chances for heads. They keep doing so until they get different results. Let X be the number of tosses. Find the pmf of X (in discrete time units). What kind of distribution is it?

Here's what I have so far: In every round (time unit) the possible results

HH - p1p2 
TT - q1q2
TH - q1p2
HT - q2p1

and so P(X=k) = ((p1p2 + q1q2)^(k-1))*(q1p2+q2p1) Which means we're dealing with a geometric distribution.

What doesn't feel right is that the question mentions 'discrete time units'. That makes me think about a Poisson distribution, BUT - Poisson is all about number of successes in a time unit, while here we only have one round in every time unit.

If I'm not too clear its only because I'm a little confused myself. Any hint would be perfect. Thanks in advance

share|cite|improve this question
Your solution is correct. "Discrete time units" probably refers that you count the number tosses, rather than the actual time when they were performed. – Sasha Dec 12 '11 at 16:39
up vote 1 down vote accepted

You may feel some discomfort, but you clearly understood the problem, and arrived at the correct answer using correct reasoning.

You should be explicit about what $k$ ranges over. Possibly you should say explicitly what you mean by $q_1$ and $q_2$. Maybe you should explicitly let $r=q_1p_2+q_2p_1$ and observe that $p_1p_2+q_1q_2=1-r$.

For the sake of completeness, you should separate out the cases $p_1=p_2=1$ and $p_1=p_2=0$, for which the problem (and answer) do not make sense.

share|cite|improve this answer

You can work out the probability that they get different results on the first toss, namely $p_1 (1-p_2)+ (1-p_1)p_2 = p_1+p_2 - 2p_1 p_2$.

If they have not had different results up to the $n$th toss, then the conditional probability they get different results on the next toss is the same; this is the memoryless property and so (since the number of tosses is a positive integer, i.e. discrete) you have a geometric distribution, as you spotted.

share|cite|improve this answer

The discrete time units refer to the series of flips. You can't have the process end after $1.5$ flips, only 1, 2, etc. Maybe that exponent $k-1$ will give you a hint to the name of the distribution.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.