The NFL Scouting Combine has always been a treasure trove of data, with over a hundred prospective players participating in various drills such as the 40-yard dash, vertical jump, and short shuttle. The deluge of information can be daunting, even for the most experienced NFL executives, let alone fans who are keen on dissecting every detail in the quest to emerge as draft gurus.
To make sense of this data, the NFL’s NextGen Stats team has collaborated with Amazon QuickSight’s powerful machine learning tools to simplify the numbers through a feature called Combine IQ. This tool, unveiled on Thursday on the NFL’s official website, aims to enhance the public’s understanding of combine metrics by organizing the data into an accessible and insightful format.
“Combine IQ is aptly named to enhance awareness about the combine by providing an insightful understanding of the numbers and how they can be interpreted,” explained Mike Band, senior manager for research and analytics at NFL NextGen Stats. The tool aims to help fans visualize and make sense of the data by presenting results from all drills and tracking data from RFID sensors placed on each athlete. It also leverages a player projection model that considers collegiate performance, size, and a consensus of ten draft rankings to evaluate players on an athleticism, production, and overall draft scale ranging from 50 to 99.
The NextGen Stats team has meticulously compiled combine data dating back to 2003, offering fans a rare opportunity to measure players’ achievements over the past two decades. The rankings from this historical data have consistently correlated strongly with NFL achievements, measured by benchmarks like emerging as a starter or earning Pro Bowl honors.
During the combine, the NextGen Stats team ensures the precision of data, promptly validating it before making it available on the website, usually within 10 minutes of a player concluding a drill. This groundbreaking information and analysis are not only available to fans online but also featured in the NFL Network’s coverage of the combine.
The data provided by the dashboard can be sorted to showcase the top performers in various categories while also offering spider graphs for comparison between players at the same position across different drills and measurements. These measurements include the NextGen Stats’ assessments of production and athleticism.
“This represents a professional-level analysis that was traditionally exclusive to NFL teams, and now we’re making it accessible to fans everywhere,” noted Ari Entin, head of sports marketing at AWS. “It’s an innovative method of capturing, visualizing, and analyzing athletic performance data in real time.”
For prospects who may not participate in certain drills at the combine, the system estimates their athleticism scores based on collected college data. The comprehensiveness of the numbers extends beyond simplistic event times, offering rankings based on top speed, 10-yard splits, speed at various intervals, and acceleration. For instance, data from the previous year revealed that Xavier Worthy achieved the fastest speeds at each 10-yard segment when he set a record in the 40-yard dash at 4.21 seconds.
The dashboard further explores how combine performance correlates with NFL success for each position, noting that athleticism scores hold the most significance for edge rushers and cornerbacks, while being less critical for centers and safeties. It provides performance thresholds that players need to achieve at their respective positions, like running backs needing to clock 4.53 or better in the 40-yard dash under certain weight conditions for a “good” rating, and “elite” ratings requiring even faster times under the same conditions.
Thresholds apply to other assessments as well, based on their relevance to various positions, such as the importance of three-cone drills for running backs and edge rushers and vertical jumps for pass receivers. Even larger players like offensive tackles have targets to meet in drills like the 40-yard dash, three-cone drill, and measure of arm length, guided by insights gleaned from two decades of past drafts.
“The model is adept at identifying crucial performance thresholds,” remarked Band. “If you’re ranked above a certain point or relative grade, the predictive analytics kick in. Essentially, the big board we aim for mirrors a consensus big board but with distinctions made through advanced metrics of athleticism, production, and size.”
