NFL NextGen Stats team unveils new tool to help fans better understand combine results

By Canadian Press on February 27, 2025.

The NFL scouting combine provides a blizzard of data with hundreds of prospects running through various drills ranging from the 40-yard dash to the vertical jump to a short shuttle.

Sifting through all that can be overwhelming for the most seasoned of NFL executives, much less the countless fans who will spend the next few months trying to become draft experts.

The NFL’s NextGen Stats team has partnered with Amazon QuickSight’s machine learning tool to put all those numbers in a more comprehensible form as part of Combine IQ, which was released to the public on Thursday on the NFL’s website.

“It’s in the name: Combine IQ. We came up with that because that is what we want people to take away, which is to make you smarter about the combine,” said Mike Band, the senior manager for research and analytics at NFL NextGen Stats. “We want fans to really understand the numbers that are being collected and the ways you can visualize and make this data digestible and yet insightful.”

The dashboard will post results on all of the drills, including tracking data from RFID sensors on each player, and put them into a player projection model that also takes into account college production, size and a consensus big board of 10 draft rankings to grade each player on a scale of 50 to 99 in athleticism, production and an overall draft score.

The NextGen stats team has modeled combine data back to 2003, allowing fans to compare players over a more than two-decade stretch with the rankings correlating well to NFL success based on measurements like becoming a starter or a Pro Bowler.

The NextGen stats team will validate all the data as its being collected to make sure it is accurate before being posted on the website, typically within 10 minutes or less of a player finishing a drill. The numbers and analysis will be avialable to fans on the website and used by NFL Network on the broadcast of the combine.

The data will be able to be sorted by the top performers in various aspects of each event, as well as having spider graphs that compare players at the same position in every drill and measurement, including the NextGen Stats’ production and athleticism scores.

“These are professional level analytics that were previously only available to NFL teams, and we’re making them accessible to fans everywhere, which is pretty cool,” said Ari Entin, the head of sports marketing at AWS. “This is a new way to capture, visualize and analyze athletic performance data in real time.”

For players who don’t test in the drills at the combine, the model will produce estimated athleticism scores based on college data that has been collected.

The numbers go deeper than just the straight times for an event like the 40 with players also ranked based on top speed, 10-yard split, speed at 10 yards and acceleration. The data shows that Xavier Worthy had the fastest speed at every 10-yard interval last year when he set a combine record in the 40 at 4.21 seconds.

The dashboard also has data on how combine tests correlate to NFL success at each position with the athleticism scores at edge rusher and cornerback being most important and least important for centers and safeties.

It also provides thresholds for players to reach at certain positions with running backs ranking as “good” with a time of 4.53 or better in the 40 if they are under 210 pounds and at 4.58 if heavier, and “elite” at 4.39 for under 210 pounds and 4.42 for heavier.

There are similar thresholds for other tests depending on the importance at each position with three-cone drills being important for running backs and edge rushers and vertical jump important for pass catchers, for example. Even the bigger players like offensive tackles have thresholds to meet in terms of 40 time, three-cone drill and arm length based on what the model has learned over drafts from the past two decades.

“The model does a really good job of finding cliffs,” Band said. “Where if you’re above this ranking or relative grade as we scale it. Then really the predictive nature goes to all of the analytics below that. It’s essentially like we want our big board to somewhat resemble the consensus big board with all of the ties broken between all of the advanced metrics across athleticism and production and size features.”

___

AP NFL: https://apnews.com/hub/nfl

Josh Dubow, The Associated Press

20
-19