[brianwawok]

> Why 25th-40th percentiles

Why not? Someone wanted to quantify "slightly faster than average", so I made up some percentiles. I don't think there is a scientific definition of "slightly faster than average" so I think I can do this. You could I suppose counter with you think it means "35th-49th percentiles" which is valid but doesn't really change the argument too much.

> Do we know the distribution of first-and-only marathon runners to repeat runners or is this a pure intuition?

Just intuition. It would be a tricky thing to survey, as you would need to wait for all current people who have ran exactly 1 marathon to die, to confirm they do not indeed run more .