I have n coins, where n-1 are fair and one is biased, showing tail on both sides. I want to ascertain whether the coin is fair by throwing k-times. If I get k-times tail, I decide it is biased. The question is :

What is the probability that this way of determination is false?

W( determination is false ) C(biased coin) T(throwing T k-times)

The solution is :

P(W) = P(¬ C,T) +P(C, ¬ T)

I am not sure why do we care for the probability of not getting tail with a biased coin. Why don’t we determine the probability of W like this:

P(W) = P(¬ C,T) +P(C, T)


You are wrong if you pick a fair coin and get tails $k$ times because you will declare it biased. You are also wrong if you pick the biased coin and do not get tails $k$ times because you will declare it to be fair. With your two tailed coin the second possibility cannot occur, but you might have a coin that shows tails $0.99$ of the time and use the same algorithm. The equation you were given correctly states that chance that your decision is wrong.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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