Additionally, the % I reported seem to fit a smooth distribution. Judging from the slope of the right tail of said distribution, I would say that the real 5-ball % may even be lower...