The 40-Yard Dash Is Moving Draft Capital and Almost Nothing Else

27 of 34 running backs who went on to finish as fantasy RB1s posted elite Broad Jump scores at the combine. That number sounds like a revelation. In certain circles right now, it is getting treated like one. People are reading it and immediately recalibrating their entire draft boards based on a standing long jump.

Slow down. That number is actually a perfect illustration of what is wrong with how everyone, teams included, uses combine data: we find a signal that works in one quadrant, generalize it everywhere, and act like we have cracked the code on a problem that has been humbling NFL front offices for fifty years.

What the Research Actually Says

There is a lot of it, and it is not flattering for the combine-as-oracle crowd. A landmark study published in the Journal of Strength and Conditioning Research found no consistent statistical relationship between combine tests and professional football performance, with the sole notable exception being sprint tests for running backs. A separate peer-reviewed study analyzed combine participants from 2005 to 2010 across six positions and found that combine measures could predict anywhere from 4% to 62% of variance in individual on-field variables. That 62% figure sounds impressive until you read the fine print: it applies specifically to QB rushing yards, which is one of the least predictive stats for actual quarterback value.

Sumer Sports ran correlation coefficients on combine tests versus NFL success from the 2008 through 2022 draft classes. In statistical circles, a coefficient of 0.60 is the minimum threshold to actually believe in a relationship. Only 5 of 11 position groups had a maximum coefficient anywhere near 0.30. For wide receivers, a position where the media loses its mind over every 40 time, there were no tests that even cleared the 0.30 threshold. None.

And yet, the forty-yard dash remains the single strongest predictor of draft order. Teams are using a number that predicts where players get picked, not whether they actually succeed, and calling it evaluation.

The Market Is Pricing the Wrong Thing

This is where it gets interesting to me as someone who thinks about mispriced markets. Kitman Labs ran an AI model on historical combine and college production data and found that first-round picks reach success just 56 to 69% of the time depending on position. Their model, applied to the 2022 draft class, flagged Brock Purdy as the quarterback with the highest probability of NFL success at 51%. He was the last pick in the entire draft. The combine did not see that. Draft capital did not see it either. What the model caught was college production efficiency and athleticism working together, not forty times in isolation.

RotoWire analyzed 800 first-round selections across 25 draft classes. Even at pick five, historically the most productive slot in round one, the All-Pro rate sits at 36%. That means nearly two-thirds of the supposedly elite picks at the best spot in the draft never make a single All-Pro team. Cornerback, one of the most heavily drafted positions in round one, carries a 64% miss rate. Teams have been drafting fast corners based on forty times for decades and burning first-round capital at a clip that should embarrass any analytically literate front office.

Meanwhile, college performance keeps showing up as a better predictor than combine testing in study after study. A study of 764 college players across positions found that college production was more strongly correlated with future NFL success than raw physical ability tests at the combine. That is a finding worth sitting with. The thing scouts fly around the country to watch in person, the actual football, is better at predicting football than the workout event every media outlet treats as a holy text.

Where the Real Edge Lives

Not all combine data is noise. The Broad Jump for running backs, the three-cone drill for wide receivers and tight ends, the shuttle for linebackers: these are genuine signals, but they require position-specific context to mean anything. Sharp analysts know this. The 40 time for wide receivers correlates more with draft position than with actual production. That mismatch is a market inefficiency, and markets that consistently misprice the same variable are markets you can exploit.

For fantasy and dynasty purposes, the practical takeaway is simple. A bad combine should make you look harder at a prospect, not cross him off. College production against competitive defenses remains one of the strongest available signals. Combine data is most useful as a secondary filter, not a primary ranking tool.

The combine is evolving, slowly. The 2026 combine just added an isometric bench press hold to better measure force production. Last year, only 25% of prospects even participated in the standard bench press, and over the past four years the overall participation rate has been just 30%. Teams are already voting with their feet on which drills actually matter.

The public hammers certain combine performances every year and reprices players based on forty times that will be forgotten in three seasons. You know what that means. The consensus board, built partly on media hype around measurables, is leaving value on the table for anyone willing to weight college production correctly and treat combine data the way it deserves to be treated: as one input in a larger model, not the model itself.

Tom Brady ran a 5.28 forty. The combine thought he was a tight end who couldn't run. He turned out fine.