I think you could make a very strong case for setting the soft caps based on normal distributions as well using the NFL combine data for each position.
Only Bort knows what 40-yd dash time equates to what speed rating and what bench/squat max equates to what strength rating.
Say HBs at the combine ran an average (μ ) of 4.42 with a standard deviation (σ ) of .08 in the 40. Whatever number equals a 4.42 is the first soft cap (average), whatever equals a 4.38 is the second soft-cap (μ-.5σ ), whatever equals 4.34 is the next soft-cap (84.1 percentile, μ-σ ), and so on with each step being the average plus one-half the standard deviation.
Only Bort knows what 40-yd dash time equates to what speed rating and what bench/squat max equates to what strength rating.
Say HBs at the combine ran an average (μ ) of 4.42 with a standard deviation (σ ) of .08 in the 40. Whatever number equals a 4.42 is the first soft cap (average), whatever equals a 4.38 is the second soft-cap (μ-.5σ ), whatever equals 4.34 is the next soft-cap (84.1 percentile, μ-σ ), and so on with each step being the average plus one-half the standard deviation.
Last edited Oct 27, 2008 22:07:13