NFL teams adopt cognitive AI to evaluate quarterbacks
- College statistics fail to predict NFL success because surrounding teammate quality distorts individual evaluation.
- iPad tests now measure spatial processing and decision-making speed under real-time stress.
- Combine data shows high cognitive scores directly correlate with higher passer ratings and fewer turnovers.
Evaluating an NFL quarterback remains sports management's most expensive guessing game.
Traditional college statistics fail to predict professional success because of what Scouting Academy director Dan Hatman calls football's N-body problem. A prospect's completion percentage reflects ten teammates, eleven defenders, and a coaching scheme as much as individual talent. Arm strength and height merely set a physical floor; they reveal nothing about how a player operates when the pocket collapses in two seconds.
Evaluators repeatedly fall for high-risk physical traits while missing elite processing speed. Hall of Fame quarterback Kurt Warner went undrafted because traditional scouts fixated on physical measurables and missed his ability to read defenses in milliseconds. Warner stocked grocery shelves before getting a chance in arena football, eventually leading the St. Louis Rams to a Super Bowl victory.
The NFL is finally abandoning paper intelligence assessments for real-time cognitive metrics. Sports psychologist Scott Goldman developed the Athletic Intelligence Quotient to measure visual-spatial processing, reaction time, and learning efficiency using spatial puzzles administered on iPads. The test isolates how fast a draft prospect processes dynamic visual input while managing competing audio distractions.
The empirical evidence backing these cognitive tools is growing. A peer-reviewed study analyzing AIQ data from 42 combine quarterbacks confirmed that higher subscale scores correlate directly with superior professional metrics, including higher passer ratings, increased passing yardage, and fewer turnovers.
Silicon Valley is pushing the predictive math further into machine learning. On Freakonomics Radio, Kitman Labs Chief Executive Officer Stephen Smith explained how his firm uses algorithmic models trained on 15 years of collegiate and combine performance data to forecast longevity. Smith noted that mental processing speed and emotional resilience remain the strongest statistical indicators for long-term success, a standard European soccer academies have tracked in youth players for decades.
Cognitive modeling will not eliminate draft busts, but it systematically narrows the margins.
Source Intelligence
- Deep dive into what was said in the episodes
686. Why Is It So Hard to Find an NFL Quarterback? • Sep 11
- A peer-reviewed study analyzing AIQ data of 42 combine quarterbacks confirmed that high test scores correlate with superior on-field metrics. These metrics include higher quarterback ratings, greater passing yardage, and fewer turnovers.
Also discussed on this episode: (9)
Sports (6)
- Dan Hatman and a 2011 academic paper agree that traditional college and combine statistics hold almost no predictive value for NFL quarterback success. This lack of correlation makes selecting a franchise quarterback highly unpredictable.
- Scout Dan Hatman argues that physical traits like height and arm strength merely establish an NFL prospect's developmental floor. A quarterback's career ceiling is entirely determined by cognitive attributes like processing speed and coachability.
- Pro Football Hall of Fame quarterback Kurt Warner challenges the conventional wisdom that great players elevate their game under pressure. Warner argues instead that elite athletes simply maintain their baseline execution while average players shrink under high-stakes circumstances.
- Kurt Warner asserts that because roughly 95 percent of an athlete's career is spent in practice, focusing on daily execution is the key to sustained professional growth. Warner used this mindset to transition from stocking grocery shelves to winning a Super Bowl.
- The NFL historically used the 1936 Wonderlic test to evaluate prospects, but teams found it non-predictive. Sports psychologist Scott Goldman developed the Athletic Intelligence Quotient to measure specific visual-spatial processing and decision-making speeds.
- Stephen Smith observes that European soccer programs track performance data from age nine, creating a connected development pathway. The NFL lacks this capability because American youth, high school, and college athletic data pipelines remain completely fragmented.
Psychology (1)
- Stephen Dubner took the 35-minute AIQ test and scored in the strong range on eight of 16 cognitive categories. The system's AI companion recommended Stephen Dubner use predefined solutions rather than trying to improvise when plays break down.
Models (1)
- Stephen Smith explains that Kitman Labs utilizes an AI model trained on 15 years of collegiate and combine data to reduce evaluation uncertainty. The model identified rookie Fernando Mendoza as having a 51 percent probability of NFL success.
Startups (1)
- Kitman Labs holds more than 4 million dollars in defense contracts to apply its athletic optimization software to the military. The U.S. Armed Forces use the platform to recruit, train, and maintain specialized personnel like fighter pilots and special forces.