4 ms·
At the end of the bell curve you have higher variance between people? Aren't two people at close points of the curve by definition similarly skilled? In fact, f
by Mandelmus 4y ago
At the end of the bell curve you have higher variance between people? Aren't two people at close points of the curve by definition similarly skilled? In fact, for two close x-values of two people at the tail end of the curve the y-values (skill levels) are much closer than those of two people closer to (but somewhat off from) the median, where the slope is steeper.
- addcommitpush 4y agoLet's do a simple experiment. Say skill is normally distributed. Let's take 1000000 people and draw their skill levels from a standard normal, and sort them by the skill levels. What is the difference in level between the 500000th person, and the person 1000 spot above them? 0.00252225 (in this simulation). What's the difference in skill levels between the 1000000th person and the person 1000 spots below? 1.596959. The 1000-person gap is 633 times larger at the top than the middle. skill_levels <- rnorm(1000000) %>% sort() skill_levels[501000] - skill_levels[500000] skill_levels[1000000] - skill_levels[999000]
- blablablerg 4y agoYeah the wording was not completely accurate but I assume he means that the tails are long.