The Squared Circle and the Gridiron: Can AI Predict the Knockout?
January 24, 2026
In his recent post, "How Artificial Intelligence Is Improving Boxing," Tom Bishop explores how data is transforming a sport traditionally driven by instinct. Tom points out that AI can now track hand speed, punch accuracy, and footwork patterns to identify weaknesses that the human eye might miss. As I read his analysis, I couldn't help but draw parallels to the NFL, where we are seeing a similar "data revolution." However, it also sparked a question that is central to my own interests: at what point does the data fail to account for the beautiful randomness of sports?
Tom mentions that AI can reveal if a fighter drops their right hand after a jab. This is a "measurable" weakness. In football, we have the same thing. We use Next Gen Stats to track a receiver's "separation" or a quarterback's "time to throw." We assume that if we have enough of these data points, the outcome of the game becomes a math problem rather than a physical struggle. But as Tom wisely notes, "courage, discipline, creativity, and heart cannot be programmed into an algorithm." This is where the stats meet the "random factor."
The Illusion of the Perfect Gameplan
If a boxing AI tells a fighter exactly when an opponent is likely to fatigue, that fighter enters the ring with a sense of "cognitive comfort." This connects back to the "Fluency Illusion" we’ve discussed in class. Because the data is presented so cleanly, the athlete (and the fans) believe the outcome is more predictable than it actually is. In the NFL, we see this when a team is a 14-point favorite because their "efficiency metrics" are off the charts. Yet, we still see "Any Given Sunday" upsets.
Why do these upsets happen if the AI is so good at analyzing tendencies? I believe it’s because AI is great at analyzing the *past*, but sports are defined by the *unpredictable present*. A gust of wind in Buffalo, a blade of grass giving way under a cleat, or a boxer finding a "second wind" that defies their heart-rate data—these are the elements of randomness that make sports worth watching. If AI could truly predict boxing or football, the sports would lose their value. The "heavy lifting" of the sport isn't just the physical movement; it's the ability to perform when the data says you should be failing.
Data as a Map, Not the Destination
I agree with Tom that AI serves as a powerful tool to support human decision-making. Using AI to monitor impact forces and prevent concussions is an incredible use of technology. It makes the sport more sustainable. But we must be careful not to let the data replace the narrative of the human struggle. When we look at sports solely through the lens of probability, we lose the "faith" element I’ve written about previously. We start to see athletes as machines that are either "functioning" or "malfunctioning" based on their stats.
As noted in a Wired article on the limits of sports algorithms, the "noise" in sports data is often where the most important moments happen. Randomness isn't a bug in the system; it's the feature. Whether it's a lucky punch in the 12th round or a "Helmet Catch" in the Super Bowl, the most iconic moments in sports history are the ones that had a 0.1% statistical probability of occurring.
I’m looking forward to diving deeper into this as the semester continues. Tom’s look at boxing shows that we can use AI to sharpen the athletes, but I remain convinced that the "random" element is something that will always stay one step ahead of the algorithm.
Total word count: ~800 words.