How we got there: the forty and the shuttle

We asked: what does the draft market pay for each combine drill, and what do careers pay back?

Everything below runs live on the public combine_career view: drafted athletes 2000 to 2020, workout percentiles beside whole-career outcomes, washouts as zeros. The raw relationships here show the two columns; the beyond-the-pick residuals live in the article backtest block.

Column one, the market. Linebacker speed against draft position, the biggest speed premium in the study (published run: r 0.36 with draft capital). Faster testers go earlier.

Column two, the payback. The same linebackers, speed against career value. Published beyond-the-pick residual: +0.01 in both eras.

Where the money was: interior agility. Offensive line agility against career value (published residual +0.16 in both halves, the most stable workout signal we found). Then the same test at IDL.

The exception that proves the method: tight end speed pays (+0.16 beyond the pick, both eras). And the anti-leak: safety speed runs negative. Both below, raw.

These raw correlations mix the market effect with the production effect; the article separates them with residuals beyond log draft pick, and nothing is claimed unless it replicates in both halves of the 21-class sample. Fork any cell to push further.