[glofish]

I think that your example shows an unexpected outcome, but not as shockingly so as the examples that it replies to.

The naive (and incorrect) deduction that one might make is that you start with 1/365 chance, so by adding twenty people you make it 20/365 that is 5%. So instead of the estimate of 5% the correct results that account for all possible comparison ends up being 50% - an order of magnitude higher.

The difference is substantial but it is not on the same orders of magnitude that this problem or the other examples are.