For example, someone already made the point that many of these founders who could raise on just an idea were likely on their second go around or prior FANG employees. Most founders have historically been male (not saying this is good - just matter of fact) and only 17% of engineers at Google are female (and I think it goes without saying that YC has a bias towards engineers).
So if YC has an explicit positive bias towards second time founders and FAANG employees, you could easily see what appears to be anti-female bias, but in reality is just positive indexing towards other attributes that historically skewed male. This may or may not be the case, and there probably are instances where there is implicit bias towards non-positive male-associated traits (macho attitude? unwarranted overconfidence?) but this data in isolation doesn't tell us much.
I'm not saying that diversity is bad, or that we shouldn't be looking to increase female representation in the founder community, but we should be careful with understanding cause vs correlation and maintain intellectual honesty when looking at statistics.
There are obviously some confounding factors, particularly if the batch of idea-only founders includes a large number of people with previous success in VC-backed startups or FAANG companies.
The best analysis would be to divide up the group into cohorts and compare against expected populations for the groups. (ie, if only 17% of people with FAANG experience are women, I would expect to see >= 17% of FAANG founders who got in on idea-only to be female).
Somehow, I doubt that's the case. Or they'll use the really weaselly language of % of teams with a woman on them, which is what I've seen from most of the big VC firms up to this point.