Do we know how many horses are actually in the sample since 1937?

This is from the Beyer article:

* Since 1937, 20 horses have raced in the Derby with a single prep race at 3, and all have failed.

* Of horses with two prep races, only one has succeeded since 1948.

I would guess the 2 prep sample is a lot larger than the one prep sample, but Beyer didn't give the exact number.

The Davidowitz article from DRF Simulcast Weekly 3/7 - 3/13, references the same general stats and his rationale for believing the stat is relevant. Mostly it's just an analysis of the contenders he feels have or have not been prepped properly. No exact numbers there either.

I'll tell you one thing, it seems to be getting so much attention that "even if it is" relevant to getting a spring 3YO ready to run his best at 10F, it might be overcompensated for on the odds board this year. Everyone and his brother might totally throw the limited prep horses out instead of just insisting on a slightly larger margin of safety on the board before betting on one.