4 ms·
>that should be rarer than one in a trillion. You're assuming that the tail of the statistics behaves like in normal distribution. In practice this is incorrec
by Grue3 9y ago
>that should be rarer than one in a trillion.
You're assuming that the tail of the statistics behaves like in normal distribution. In practice this is incorrect (see "black swan theory").
Various sports that employ statistics (cricket, baseball, etc.) sometimes have a single player who strays from bell curve in a ridiculous way. For example Don Bradman's batting average in cricket [1].
[1] https://www.statslife.org.uk/sports/1989-did-don-bradman-s-cricketing-genius-make-him-a-statistical-outlier https://www.statslife.org.uk/sports/1989-did-don-bradman-s-c...
- scarmig 9y agoIQ is a synthetic measure that follows a normal distribution by construction. To the extent that measurements of IQ in a population via test deviate from a normal distribution (with, say, fat tails), that's a reason to doubt the validity of that particular test in that domain.