UCCS Faculty Algorithm Reveals Undervalued And Overvalued Quarterbacks In NFL Draft (Photo) -04/23/25
As the 2025 NFL draft approaches, we use an algorithm (https://www.sciencedirect.com/science/article/pii/S0377221721002034?via%3Dihub) to evaluate how prospect quarterbacks will perform at the professional level, challenging traditional scouting rankings. The algorithm reveals undervalued players like Jackson Dart and Jalen Milroe, while questioning the hype around Shedeur Sanders. By emphasising rushing ability, the algorithm provides a new way to evaluate quarterbacks and their potential success in the NFL.
Pregame huddle: How the algorithm sizes up quarterbacks
The algorithm, developed by Joe Craig at University of Colorado Colorado Springs and Niven Winchester at Auckland University of technology, uses data from college games to predict the performance of rookie quarterbacks in the NFL. Quarterback performance is measured using ESPN’s Total Quarterback Rating (QBR), a numerical measure of quarterback performance that ranges from zero to 100, with higher numbers indicating better performances. The algorithm determined that to succeed in the NFL quarterbacks must be good passers but rushing ability also matters.
The algorithm estimates each quarterback’s Max5-QBR, which represents the projected maximum QBR value a quarterback is likely to achieve in their first five years in the NFL.
Evaluating 2025 draft quarterbacks
Max5-QBR values for the top-ranked quarterbacks in the 2025 NFL draft and ESPN scout scores are displayed in the table below.
Jackson Dart has the highest expected Max5-QBR value (56.6), and is consequently the algorithm’s top-rated draft pick, followed by Cam Ward (56.3), and Jalen Milroe (53.8). Quarterbacks with the highest scout scores are Ward (90), Shedeur Sanders (90), and Dart (87).
Differences between the algorithm’s rankings and scout scores can be used to identify undervalued and overvalued prospects. Dart is undervalued – he is the top draft prospect according to the algorithm, but scouts rate him only as the third-best contender.
Other quarterbacks that are undervalued by scouts include Jalen Milroe (ranked third by the algorithm but fifth by scouts), and Dillon Gabriel (ranked fifth by the algorithm but eighth by scouts).
The most overvalued top quarterback according to the algorithm is Shedeur Sanders, who is ranked equal first by scouts but only sixth by the algorithm.
The difference in rankings arises from the algorithm’s emphasis on rushing ability, which, based on historical data, scouts tend to undervalue when assessing quarterback prospects. Specifically, when scout scores and a measure of rushing ability are included in the algorithm, rushing skill has a positive and statistically significant effect on future NFL performance, indicating that it is not fully accounted for in scout evaluations. The algorithm finds that strong rushing skills will enhance these quarterbacks’ performance in the NFL, possibly because effective rushing at opportune times enables quarterbacks to gain the additional yards required to be awarded a new set of downs.
Among the top quarterback draft prospects, Dart has the best college rushing statistics. Milroe and Gabriel, two other undervalued quarterbacks, also posted strong rushing statistics during their college careers.
At the other end of the scale, Sanders averaged –0.5 yards per rush in the 2024 season and has the weakest college rushing statistics among the top quarterback prospects.
How has the algorithm performed in the past?
A systemic analysis (see Figure 3 of the article describing the algorithm) reveals that the algorithm’s predicted NFL performance of college quarterbacks is more closely correlated with actual output than a prediction model based solely on scout scores.
Although the algorithm’s highly rated draft picks don’t always work out in the NFL, two recent examples are noteworthy. First, the algorithm ranked Jayden Daniels the top 2024 draft quarterback by a large margin, despite scouts ranking him the second-best quarterback. In 2025, Daniels won NFL Rookie of the Year and was selected for the Pro Bowl.
Second, Lamar Jackson, who was the NFL’s MVP in 2019 and 2023, was considered the top quarterback in the 2018 NFL draft by the algorithm but only the fifth best prospect by scouts.
Both Daniels’ and Jackson’s college rushing statistics were considerably better than other quarterbacks in their respective draft years.
Not a vintage year
The algorithm also supports the general impression that the 2025 quarterback draft lacks standout talent. In the 2024 draft, Daniels’ predicted Max5-QBR was 66.7, and all drafts since 2021 have at least one quarterback with a predicted Max5-QBR greater than 60. In contrast, the top prospect for the 2025 draft (Dart) is expected record a Max5-QBR of just 56.6.
The final play
As the NFL draft unfolds, the algorithm's highlights undervalued prospects like Dart and Milroe. With rushing ability proving key to success, it will be interesting to see if teams recognise this or rely on traditional scouting rankings.