"Honestly, I think the biggest factor in these longer distances is the smaller sample size. There just aren't as many elite athletes competing at the ultra distances as there are at marathon and shorter distances."

Yes, order statistics are definitely relevant here. The bigger the sample, the more random fluctuations even out. If men have any edge in running, in a race between 1 man and 1 woman, the winner will nevertheless often be the woman; if there's a race with 1 million men and 1 million women, the winner would almost always be a male.

The article points out (somewhat obtusely) how this works within a race (a smaller fraction of women means an even more disproportionate probability the winner will be male), but seems like it should also works across kinds of races, too. The more niche a sport...