A similar trough exists in general aviation—accident rates are markedly higher for somewhat experienced pilots than for beginners. The consensus among pilots seems to be that you are most dangerous between the time you think you know what you're doing, and the time you survive your first big scare.
There's probably a good lesson here for computer programmers...
I seem to remember hearing the same about drivers. Might have been for my country (Germany) specifically, where the minimum age for a license 18. If you get your license at 16, you're still in the upwards swing on risk-taking behaviour, which might drown out all other statistical signals.
Yes, the most dangerous drivers tend not to be the ones who just got their license, as they usually still adapt and don't drive too fast. It becomes dangerous after a few months or 1-2 years when a driver feels confident enough even though he still has very little experience.
I disagree here. You're dangerous until you know you know what you're doing. Having a bad scare might make you not repeat that situation, but you're still just as dumb after it as you were before (well, at least that's been true for me).
> when they leave their instructors behind and fly as pilot in command for the first time
This is very different. If an instructor is present the danger is clearly lower. The book deals with the phase where pilots first fly alone without external help.
Read the book and you'll see that the spike comes after the first few dozen flights without an instructor present (from 100 to 350 hours of flight time). I believe you're quoting the blurb, which doesn't reflect what's actually in the book.
There's probably a good lesson here for computer programmers...