AWS for M&E Blog
NFL and AWS annual Big Data Bowl competition will predict player movement
The moment before an NFL quarterback releases the ball contains infinite possibilities. The receiver could break inside or fade outside, the safety might jump the route or hang back, the cornerback could close the gap or get beaten deep. For decades, these split-second decisions existed only as game film debates and coaching hypotheticals. Now, the NFL and Amazon Web Services (AWS) want to predict them.
In its eighth year, the is challenging participants to do something unprecedented: forecast exactly where players will move in the crucial seconds after the ball leaves the quarterback’s hand.
As this year’s competition marks a decade of collaboration between AWS and the NFL Next Gen Stats team, it represents a milestone in how data analytics has transformed professional football. Since 2015, when AWS first started working with the NFL to power Next Gen Stats, the technology has evolved from basic player tracking to sophisticated predictive modeling. It influences everything from in-game strategy to broadcast coverage.
Driving a new era of football analytics
What started in 2018 as an experimental competition has evolved into one of sports analytics’ most prestigious and influential events. “In year one, we just wanted enough good submissions to justify doing it again the next year,” admits Mike Lopez, senior director of football data and analytics for the NFL. “Now it’s become part of how we discover new metrics that end up in games, on broadcasts, and in front offices.”
The competition’s impact is visible every season. When you hear broadcasters discuss Tackle Probability or see real-time Pressure Probability graphics flash across your screen, you’re witnessing the direct results of past Big Data Bowl innovations. These metrics have transformed from academic exercises into essential tools for understanding the game.
Figure 1: A timeline infographic.
The 2025 challenge: Predicting the unpredictable
This year’s competition tackles one of football’s most complex puzzles: predicting player movement after the ball leaves the quarterback’s hand. Participants receive Next Gen Stats tracking data—the same information NFL teams use. The data shows player positions, speeds, and accelerations up to the moment of the throw. Their challenge? Map out where players will go next.
“We’re basically saying, ‘Here’s point A, tell us point B… and everything in between,'” Lopez explains. “It’s not just guessing the end location—it’s mapping the actual path, the micro-adjustments, the angles, the speed changes.”
This is a complex challenge due in part to the number of factors that could influence the outcome. A receiver might adjust to an underthrown ball, accelerate to split defenders, or deliberately slow down to create separation. Defenders must react in real-time, making split-second decisions that could mean the difference between an interception and a touchdown. Modeling these interactions requires understanding both the physics of human movement and the strategic elements of football.
This year’s focus on movement prediction represents a significant leap forward in football analytics. Previous metrics have largely focused on outcomes—yards gained, completion probability, and pressure rates. By attempting to model the actual paths players take, the 2025 Big Data Bowl is pushing into territory that has previously been challenging to measure and predict consistently.
“It’s getting at the what-ifs that players and coaches talk about all the time,” Lopez says. “If I’d gone here instead of there, what would have happened? That’s been almost impossible to model accurately. Until now.”
From notebook to broadcast
What sets the Big Data Bowl apart isn’t just the complexity of its challenges—it’s the speed at which winning solutions become real-world tools. Thanks to AWS infrastructure, the pipeline from competition submission to Next Gen Stats deployment is remarkably short. Take Rush Yards Over Expectation, developed by a pair of Austrian data scientists in the 2020 competition. Within months, it appeared as a live graphic during NFL playoff games.
“A lot of data science competitions stop at the winning slide deck,” Lopez says. “With AWS, we can take that notebook, productionize it, and get it into coaches’ hands or on a broadcast within a season.”
This rapid deployment cycle has already produced more than 75 machine learning-based stats, from Pressure Probability to Tackle Likelihood. Each metric is built on AWS infrastructure that helps process billions of data points each season without breaking stride.
A unique community of innovation
What truly sets the Big Data Bowl apart is its collaborative culture. This community openly shares code, exchanges ideas, and sometimes even combines teams mid-competition. This spirit of cooperation has helped create a talent pipeline that the NFL says has placed more than 70 participants in jobs with NFL teams, the league office, and other professional sports organizations.
“It’s the most positive, engaged sports analytics community I’ve ever seen,” Lopez notes. “People are here because they love football and they want to push the game forward.”
The competition offers two distinct tracks to accommodate different approaches to innovation:
- The Leaderboard track will challenge participants to build algorithms that minimize prediction error when tested against future game data. This will verify that winning models work on historical data, as well as genuinely predict future outcomes.
- The Data Visualization track encourages participants to create visual stories and interactive tools that reveal new insights about player movement. This track recognizes that sometimes the most valuable innovations come from helping people see patterns they’ve never noticed before.
This dual-track approach has proven invaluable in bridging the gap between data scientists and football professionals. While the Leaderboard track pushes the boundaries of predictive modeling, the Data Visualization track helps translate complex statistical concepts into intuitive visual representations that coaches and players can quickly understand and apply.
Rewriting the NFL analytics playbook
Registration for the 2025 competition opens September 25, 2025 and remains open until the end of December. Finalists will present at and the NFL Scouting Combine (Feb. 23–March 2, 2026 at Lucas Oil Stadium in Indianapolis) and winners will be determined. For developers and data scientists, it’s a rare opportunity to work with authentic NFL data and potentially influence the sport itself.
The challenge is daunting, but the impact is real—your model might help coaches make better decisions, scouts evaluate talent more effectively, or fans understand the game more deeply.
to be a part of the NFL and AWS annual Big Data Bowl competition to predict player movement.
Contact an AWS Representative to know how we can help accelerate your business.