3 ms·
> the author doesn't make clear he makes the point across several points (4, 6, and 7 and maybe 1) which is that using a power law gets tail estimates very bad
by princeb 9y ago
> the author doesn't make clear
he makes the point across several points (4, 6, and 7 and maybe 1) which is that using a power law gets tail estimates very badly wrong (or they fit the tail and then get the distribution of the rest of the domain wrong).