I want to revisit a probability puzzle from a previous post: see Alternating coin toss game.
You play a game where you alternate tossing a coin with a friend, and the first person to toss heads wins. But let’s make it a little more interesting by making the coin weighted, or biased, so that it lands heads only 25% of the time. If you were to start, what would be your chances of winning?
Let’s call the probability of tossing heads and the probability of tossing tails .
On the starter’s first toss, his chances of tossing heads are . The chances he’ll toss heads on his second toss are . On his 4th toss it’s and so on.
So his chances of winning the game is the sum of the geometric progression …
Let’s use the shifting approach to solve this. The idea is to subtract out the first element, and then shift the rest of the elements to the left. So first subtract:
…
Now to shift the terms to the left, we need to divide by
…
So…
This gives us a general solution to the alternating coin toss problem. If we plug in the unweighted coin probabilities, we get which is what we’d worked out before. For the case where the probability of tossing a head is 1/4, then we get , which is the probability of winning for the starter.