Methodologically, why would you have one cutoff vs a different cutoff per group (as there are different qualifying times per group)?
I am not a marathoner, but I'd imagine that a 6 min decrease from the stated qualifying time cuts out a larger proportion of younger runners (i.e. decreasing the threshold from 2h55 to 2h49 for men 18-34 seems like a much sharper cut than decreasing 4h20 to 4h14 for women 60-64). I would have thought you'd want to pick the delta by looking at the distribution within each gender x age pool.
I am not a marathoner, but I'd imagine that a 6 min decrease from the stated qualifying time cuts out a larger proportion of younger runners (i.e. decreasing the threshold from 2h55 to 2h49 for men 18-34 seems like a much sharper cut than decreasing 4h20 to 4h14 for women 60-64). I would have thought you'd want to pick the delta by looking at the distribution within each gender x age pool.