It is inappropriate to ask "how likely was it for this server to have crashed", when there is nothing special about this server other than it is the one we already know has crashed. Once we carefully pick this crashed server from the set of all servers in the world, the "null hypothesis" is gone.

The correct question is "how likely was it for any server to have crashed", and I think the answer is that it was fairly likely.

It is unclear if being New Year's made a crash more likely, but I think we can agree it made it far more likely that we would hear about it.

Assuming your math is correct (I didn't bother to check it), that would be the odds of a random server failing in a random 20 second interval.

You didn't pick a random server, you picked one already known to have crashed. And you didn't pick a random 20 second interval either. The odds of that server crashing in that 20 second interval was 100%, because it was already known to have happened. This is a classic mis-application of statistics.

Admittedly, the interval right at New Year's is a bit suspicious, since there is some specific code to handle leap years, etc. But given that there wasn't a rash of outages reported, I am going along with the coincidence theory.

Read it again. He didn't say that at all. He said that the embarrassment of being publicly identified as a victim of sexual assault is less than the embarrassment of being publicly accused of being the assaulter. He was comparing the results of publication, not of the crime itself.

Actually, OTP does use a simple XOR cipher. Given a truly random key the same length as the message, that is perfectly adequate.