The 40-yard dash has stopped predicting NFL careers
For fifty years the 40-yard dash has been the headline number of the NFL combine. It opens every draft broadcast, it moves prospects up and down boards by a round, it gets credited as the reason teams reach for fast receivers and demoted as the reason teams overlook slow ones. It is also, by every modern measurement, a worse predictor of NFL careers than it has ever been. The same stat that used to do real work in player evaluation has been steadily losing its predictive edge for twenty years, and the reasons aren't mysterious. The dash has stopped predicting because the league it was measuring for has changed faster than the test did.
What the 40 was built to measure
The 40 was originally chosen as a proxy for game speed because forty yards was the average distance a punt traveled in the air in the 1960s and a coverage gunner needed to cover that distance before the return man fielded the ball. The measurement made sense for an era in which the game was played mostly between the twenties, special teams accounted for a meaningful share of total game value, and the dominant offensive scheme produced relatively few long pass routes per game. The skills required to run a fast 40 — raw straight-line acceleration over a medium distance — correlated reasonably with the skills required to play certain positions, especially cornerback and outside receiver, in the league of that era.
Modern offenses do not look like that. The pre-snap motion, the route adjustments at the line, the option routes, the condensed splits, the run-pass options — the game has become a series of short bursts and direction changes within a few yards of the line of scrimmage. The number of actual in-game sprints that look like a 40-yard dash, in any meaningful sense, has compressed. A receiver runs a true straight-line 40 maybe two or three times per game, if that. Most snaps demand a shorter, more directional movement that the dash doesn't test.
What the modern research has found
Studies of combine-to-career data going back through the cap-era draft classes have produced a consistent finding: the 40-yard dash's correlation with NFL success peaked in the 1990s and has been steadily declining since the mid-2000s. The correlation between 40 time and total snaps played in the NFL across a career is now, depending on position, somewhere between 0.08 and 0.18 — weak by any reasonable statistical standard. For comparison, the correlation between college production and NFL success is around 0.30-0.40 at most skill positions. The dash is doing far less work than the broadcast coverage of the combine implies.
The drivers of the decline are knowable. Stronger college S&C programs have compressed the spread of 40 times among elite prospects. Modern pre-combine training is explicitly optimized to shave fractions of a second off the time, which means the score increasingly measures who committed to specialized training in February rather than who is fastest in football pads. And the in-game movement skills that distinguish good players from great ones at the professional level have very little to do with maximum straight-line speed at any distance.
The split-time problem
The 40 is timed as a single number, but it is actually three separate measurements squashed together: the 10-yard split, the 20-yard split, and the full 40-yard finish. The three components measure different physical traits. The 10-yard split is mostly start-up acceleration. The 10-to-20 component measures continued acceleration under load. The 20-to-40 component measures maximum velocity. NFL game speed correlates most closely with the 10-yard split, then the 20-yard split, then much more weakly with the full 40 time. The headline number averages across all three, which means it dilutes the most predictive part of the test with a top-end speed component that almost never matters in actual games.
Smarter front offices have, for over a decade, focused internal evaluation on the 10-yard split and treated the full 40 as a secondary number. The public broadcast still leads with the 40. The market still adjusts to the 40. The teams that have rebuilt their evaluation around the split times are quietly outperforming the teams that haven't, and the players who get drafted on the strength of a great 40 but a mediocre split increasingly do not become NFL players. The market signal and the evaluation signal have drifted apart.
Position-specific decay
The 40 still does some work for cornerbacks and outside receivers, where straight-line speed correlates with the ability to keep up with deep routes. For those positions, a slow 40 is genuinely a red flag that often shows up in NFL production. But the trend is even visible there. Modern cornerback play depends much more on hip flexibility, change of direction, and ball skills than on top-end speed. The small handful of corners drafted in the last decade who couldn't break 4.6 in the 40 but produced at All-Pro levels in the league were not exceptions. They were the visible signal that the testing battery is missing the relevant skills.
At every other position, the 40 has essentially no predictive value. Linebackers, interior offensive linemen, defensive tackles, tight ends, quarterbacks — none of them play a position in which straight-line forty-yard acceleration is the rate-limiting physical skill. The 40 time gets recorded, broadcast, and discussed for all of them because the combine produces the number, but the correlation with their NFL careers is statistically zero. The broadcasting convention of caring about every prospect's 40 time is a holdover from an era when the position-specific decay hadn't yet set in.
The tracking-data replacement
Every modern NFL front office now has access to tracking data from college games, which measures actual in-game speed at every snap. The data answers the question the combine was trying to answer, with better fidelity and against actual defenders. A receiver's average separation at the top of his route, his max speed across all snaps, his acceleration profile in the first three steps, his change-of-direction time on comeback routes — all of this is measurable directly from college film. The 40 time is a single proxy observation in a clean environment. The tracking data is thousands of observations in the actual game environment.
Front offices that rely on the tracking-data layer have moved the 40 from "headline number" to "confirming check." It is consulted to flag obvious physical deficits the tracking data might have missed because of small sample, and otherwise ignored. The public conversation has not followed. Mock drafts still cite 40 times as if they were primary evaluations. Player rankings still get reshuffled on combine day. The market is using a measurement that the actual buyers of football talent have stopped treating as primary.
How to read the combine numbers
The honest version of the combine is that the 10-yard split, the broad jump, and the three-cone drill are doing most of the predictive work for the skill positions where movement matters. The 40-yard dash gets the headline because the broadcast is built around it and the time is easy to remember. Within any reasonable cluster of prospects at the same position, the 40 differentials are usually small enough that they fall well inside the noise band of single-trial measurement. A prospect who runs 4.42 and a prospect who runs 4.49 at the same position are essentially identical on that measurement. The way the discourse treats those two times as meaningfully different is the part that has stopped being supported by the data.
The 40 will not disappear from the combine, because the broadcast value is real and the historical continuity matters to fans. But the part of the stat that used to do work in player evaluation has been steadily disappearing for twenty years. The teams winning in the late rounds of the modern draft have figured this out. The teams still reaching in the early rounds for the burner with the headline 40 time keep getting outperformed by the teams that drafted the prospect with the worse 40 and the better tracking-data profile. The same stat that once moved boards by an entire round has quietly stopped earning the airtime it still gets.