I've made lots of mistakes. If we take 'favorite' to mean the one that I regret the least, then I think getting out of the military in 2005 would be the one. That was a mistake only to the degree that it meant no more overseas relocations (at least for a while) and it meant having to think about what to wear every day (this is a much bigger issue than I had anticipated). The alternative to retiring would have been taking another assignment and then taking my turn on rotations to Iraq or Afghanistan (or both). In any case, I would have only been able to stay in another six years at the most, because I'd have had 30 years at that point and been forced to retire. On the other hand, as a result of retiring, I no longer have to do a large number of annoying things, and since retiring, I've finished a masters degree, bought an airplane, and have applied to Auburn to do a PhD (admission decision pending) starting this fall.

...in a group of 23 people, the odds are 50-50 that a pair of the 23 will share the same birthday.

Somewhere or other I read about the probability of two people sharing a birthday. For some reason, the answer is not very intuitive. In a class of, say, 23 students, what is the likelihood that at least two people will share the same birthday (not necessarily year, just the day and month)?

The typical person will guess a low probability, but the fact is that the chances are 50% that two people in a group of 27 will share the same birthday.

When the size of the group gets to 57, then probability is 99%!

I never believed this until, as part of my classes at a military school where I taught, I polled my groups of 23 to 25 students and discovered that over the course of 18 classes, 10 of them had a pair of shared-birthday students. That's a little over 50%, which is just what would be expected.

It's hard to accept, bu seems to be true, and I think it reflects how poorly equipped we are to estimate probabilities. After all, look how many people play the lottery!