I am trying to find an intuitive and non-rigorous explanation to this math question.

If you have 100 coin flips of a fair coin, P(X = H) = .5, the expected number of heads is 50. Likewise, if you have a weighted coin, where P(X = H) = .3, the expected number of heads would be 30. Now, if you were to win $1,000 dollars if the coin you picked had a true value of heads equal to the expected value after 100 flips, which coin would you choose?

I understand from a pure mathematical perspective that you should choose the weighted coin, as variance is measured by P(1-P), assuming only one flip, and thus having P equal to .5 will maximize your variance. But is there a geometric or simple way to interpret and show this phenomena?

Happy Holidays.


To make this easier to make sense of, try looking at a more extreme version of what's happening.

Start off by considering a two-tailed coin where $P(X\! =\! H) = 0$. In that case, the expected number of heads is $0$ and there is no variance at all. If this is an option, it's clearly the best choice: you're guaranteed to win the competition by picking it because the only possibility is for the outcome to match the expected value.

Once that makes sense, we can back off a bit and consider an extremely weighted coin where $P(X\! =\! H) = 0.01$. In that case, the expected number of heads is $1$. But because it's so heavily weighted, the chance of getting a head on any of the 100 flips is very low, so it's still a much better bet that we'll win picking this coin rather than a fair coin. From this, we can start to gain an intuitive grasp that low/high probabilities imply small variance which means that we're more likely to land "near" the expected value.


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.