This is a fantastic article and I am reading it from top to bottom, so it's possible my question is addressed further down. However, I am wondering if you explored other ways of measuring relative improvement after age 19. It seems like "percent improvement" is not ideal, since clearly there can't be 100% improvement and a 10% improvement is much much better than a 5% improvement. Although, it's tough to pick another measure since the athletes themselves are defining what's "physically possible." :) I don't have a good answer, just wondering if you considered something else. One idea would be to measure how many standard deviations their speed moves from the mean sprinting speed of males that age. But that's not good since (A) that information is probably not available and (B) these people are on the tails so a normal distribution is probably not applicable. Thanks!