Rewriting The Rules Of Fencing With Artificial Intelligence

PARIS, FRANCE – AUGUST 04: Tommaso Marini of Team Italy (L) competes against Iimura Kazuki of Team Japan (R) during the Fencing Men’s Foil Team Gold Medal match on day nine of the Olympic Games Paris 2024 at Grand Palais on August 04, 2024 in Paris, France. (Photo by Ezra Shaw/Getty Images)

Getty Images

In the months before the 2024 Olympic Games, USA Fencing suspended two referees for communicating with each other during an Olympic qualifying tournament. While the two men denied manipulating the results, they each received a nine-month suspension. The incident was part of a growing wave of scrutiny around the integrity of fencing officiating culminating in formal arbitration complaints that challenged the Olympic eligibility of two American fencers whose victories were called into question. Though both athletes ultimately retained their roster spots, the controversy underscored a deeper problem. Fencing, particularly sabre and foil fencing, require subjective judgments from referees. The judgements determine outcomes at the sport’s highest levels, and even the appearance of bias can erode trust.

That problem is exactly what Vance Wood, Jason Mo, and the team at Allez Go are working to solve. A former varsity fencing captain at Johns Hopkins with a degree in computer science, Wood is combining his expertise on the strip and in the lab to build AI-powered tools that use computer vision to assist referees and make the sport more transparent, consistent, and fun to watch.

Inherent Subjectivity In Fencing Scoring

Fencing is one of the oldest Olympic sports, with roots tracing back to Medieval and Renaissance Europe, when swordplay was not a game but a matter of life and death. The rules, techniques, and even modern scoring systems are echoes of those original duels. Today, fencing can be categorized into three disciplines: foil, epee, and sabre fencing. In epee, for example, a point is awarded to whichever fencer lands the first touch anywhere on the body, a rule derived from the idea that any wound from such a heavy blade could be fatal.

In foil and sabre, however, things get much more complicated. Both use the concept of “right of way” to determine who earns the point when simultaneous touches occur. The idea is meant to simulate who had the tactical advantage in a real swordfight. In other words, who attacked with intention, who parried effectively, and who had their arm extended first. In foil, only the torso counts as a target, and hits must be made with the tip. In sabre, the weapon is more of a slashing blade, and points can be scored with the edge or tip on any area above the belt, mimicking cavalry combat.

These nuances make fencing both thrilling and confusing. While scoring machines flash red and green lights to indicate contact, in foil and sabre that is just the beginning. It is up to the referee to decide whether the point actually counts, based on who had “right of way.” At the highest levels, those decisions often come down to split-second timing. Did the hand move before the foot landed? Was that a true parry or just a lucky block? Even with slow-motion replays, referees regularly disagree. When calls are inconsistent or seem biased, fencers and fans are left questioning whether the outcome was truly fair.

Turning Fencing Into Data

Enter AI. In a sport where the difference between victory and defeat can hinge on imperceptible movements, artificial intelligence offers a way to bring clarity to the chaos. Allez Go’s flagship tool is the Assistant Fencing Referee, or AFR. It uses computer vision and neural networks to analyze bouts in real time and offers referees a scoring recommendation based on objective movement data. The goal, Wood says, is not to replace referees, rather, to have “the assistant referee providing the scoring recommendations.”

In foil and sabre fencing, matches are often decided by movements that happen in a split second. Referees must judge who began their attack first, whether a parry was successful, or if a fencer pulled back their arm and gave up the right of way. Even with slow-motion video, these calls can be hard to make and are often inconsistent. Computer vision helps by analyzing video footage frame by frame and tracking the exact timing and position of a fencer’s hand, feet, and blade. It sees things that are easy to miss in the moment.

This kind of analysis allows the system to apply the rules more consistently. Instead of relying on human judgment alone, the Assistant Fencing Referee can compare each exchange to patterns it has learned from past bouts. It uses measurable information to determine who had priority in an exchange or whether a point should be awarded at all. For referees, it acts as a second set of eyes. For athletes and coaches, it makes outcomes easier to understand and trust.

This is part of a broader movement across sports to bring precision and consistency to moments of high-stakes judgment. In soccer, VAR has changed how goals, penalties, and offsides are called, giving referees a way to review critical decisions with more confidence. In the NBA, tracking technology now maps player and ball movement to support more accurate goaltending and out-of-bounds reviews. These systems do not replace officials, but they do give them tools to slow the game down, apply rules more consistently, and reduce the burden of human error. Fencing, with its lightning-fast exchanges and deeply subjective scoring, may be one of the sports most in need of this kind of help.

Tech As A Medium To Expand Fencing Viewership

The team behind Allez Go is not just focused on fixing inconsistent officiating. They are also trying to make fencing more accessible to people who do not already understand the sport. Wood thinks “that by highlighting the blades and displaying graphics, it’s more visually engaging for viewers who are not fencers.” Compared to other Olympic events, fencing can be difficult to follow. The action is fast, the rules are complex, and the reasons behind each scoring decision are not always obvious. That creates a barrier for casual viewers and limits the sport’s broader appeal.

To address this, Allez Go has developed a second tool called Atticus. Atticus is to fencing what the yellow first down line is to football. It is an invisible cue made visible for the audience. It uses visual overlays to highlight blades, track movement, and show directional cues during a bout. The idea is to help viewers see who is attacking, defending, or scoring in real time. It is currently being deployed in partnership with Sabre-X and Leon Paul.

An AI Era For The Ancient Sport Of Fencing

Whether these AI tools are fully embraced by fencing federations remains to be seen. But the need they address is clear. When athletes are losing Olympic qualifying points because two elite referees cannot agree on what just happened, the case for smarter, more transparent tools practically makes itself. If Allez Go’s vision succeeds, the next generation of fencers and fans might get to see the sport in a new era of radically transparent and consistent referees.