I did almost exactly that back in May: Towards a better method of slot allocation - #215 by latkin It was 15%, not 10% but same idea.
This was in response to a big report that went around claiming women were “more competitive” or had “deeper fields”, evidence by a bigger proportion of their AGs finishing within 15% of the AG winner than men.
As it turns out, proportion of AG finishing within X% of the AG winner is driven almost entirely by AG size and the dry statistics of sample sizes. Nothing to do with performance, really. This simply isn’t a valid measure of “competitiveness” or “depth”. Once you normalize for AG size, women and men look exactly the same.