The NFL’s Next Gen Stats system can now use AI and player-tracking data to identify what kind of running play an offense called, where the play was designed to go and which defender each blocker was responsible for—giving the league a machine-readable view of trench football that traditional box scores largely miss.
In its 2026 Next Gen Stats technical release, the NFL says it worked with AWS Professional Services on transformer-based models that interpret the spatial and temporal relationships among players during a running play. Player locations are captured 10 times per second, but the new step is teaching software what those movements mean inside the structure of an offense.
The result is a run-scheme classifier with 16 primary labels. It can distinguish broad man, zone and gap concepts, add secondary tags such as read option, split zone and pitch, and compare the intended rushing gap with the gap the ball carrier actually attacks.
The model reads assignments, not just yards
The companion run-blocking model goes deeper. It identifies each offensive player’s blocking assignment, the type of block being attempted and when engagement with the defender begins and ends. That creates data for double teams, block duration, defenders who disrupt a run without making the tackle and individual blocks that spring a runner into open space.
Independent reporting from Yahoo Sports describes the shift as an attempt to assign more context to outcomes produced by multiple interacting players. That is especially important for offensive linemen, whose work can determine the shape of a play while leaving almost no conventional counting statistic behind.
The same tracking ecosystem already produces matchup-level analysis during live games. In an official Next Gen Stats post on X, the system broke down individual pass-rush matchups and pressure rates, illustrating the kind of player-to-player context the league is increasingly extracting from tracking data.
Offensive-line play becomes measurable at the frame level
A running back can gain eight yards because five blockers execute perfectly, because one lineman creates a decisive lane or because the runner escapes a breakdown. Traditional statistics record the eight yards. The new models are designed to separate some of the interactions that produced them.
That extends a sports-data trend BitcoinVersus.Tech recently examined in the NFL’s use of shoulder-pad sensors to quantify contact. In both cases, the important shift is from describing the final outcome to instrumenting the physical process that created it.
The run models also illustrate why transformer architectures are useful beyond language. A guard’s movement does not mean much in isolation; its meaning depends on what the center, tackle, back and defenders are doing at the same time. The model is looking for relationships across a changing 22-player system rather than reading one coordinate stream independently.
Sports telemetry is moving closer to coaching language
Other sports are moving in the same direction. BitcoinVersus.Tech recently covered TDK’s sensor-equipped smart javelin, which converts a throw into measurable motion data, and markerless motion capture being used for rapid football player testing. The common thread is that events once judged mainly by film and expert observation are becoming structured data.
For the NFL, the next challenge is validation. A model can label an assignment, but teams and analysts will care about how consistently that label matches the actual playbook responsibility. If accuracy holds up across formations, motion, option concepts and defensive fronts, the system could give offensive-line and run-defense analysis a statistical vocabulary much closer to the one coaches already use.
The box score still tells who carried the ball and how far it went. Next Gen Stats is trying to explain who moved whom, where the play was supposed to go and why the lane existed in the first place.
BitcoinVersus.Tech
Advertisement
BitcoinVersus.Tech Editor’s Note:
We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support to help further secure the integrity of our research initiatives, please donate here: 3C9o19EH5HSiwEPyCTmEKzxhNCbo2X6TTb
BitcoinVersus.tech is not a financial advisor. This media platform reports on financial subjects purely for informational purposes.
Leave a comment