The Rounding Error on MMA Scorecards: When Three Seats Create Three Different Fights

On the night of December 10, 2026, in Las Vegas, the three scorecards for Paddy Pimblett versus Jared Gordon at UFC 282 were read out with the same sequence: 29-28, 29-28, 29-28. No card differed by a single point. No card read 30-27. Three people seated at three different positions around the cage, watching three different frames, produced the same result.

The crowd reacted before the announcer could lower the microphone. Over the following twenty-four hours, thousands of people re-scored the fight using phone clips, screen grabs, and three-second cuts. I did something else. I reopened the Nevada State Athletic Commission's scoring rules, read section 467, and counted how many words in it are actually measurable.

I do not watch the punch; I watch the scorecard that records the punch.

That habit formed in the summer of 2026. I was seventeen, sitting in Seoul, watching South Korea play Germany in Kazan, and saw a goal disallowed by an assistant referee's flag and then restored after video intervention. Kazan erased a goal, but it opened an eye. From that night, I stopped watching sport as a spectator. I watched it as someone auditing a record.

The record of an MMA fight sits on three small sheets of paper, filled in with pencil, in roughly fifteen seconds between rounds. That is where everything is decided, and also where the fewest people look.

The ruleset: a system borrowed from boxing

The ten-point must system was not invented by MMA. It emerged in boxing in the early twentieth century, was standardised in the United States by mid-century, and was carried almost intact into MMA when the sport needed a scale that could be printed on a paper slip. When the New Jersey State Athletic Control Board drafted the unified rules in 2026, it kept the ten-point structure and only rewrote the criteria descriptions. The round became the unit of currency. Three rounds, three coins. Five rounds, five coins. No fractions.

In 2026, the Association of Boxing Commissions revised the order of criteria. Before that revision, three criteria — effective striking and grappling, aggression, and control — sat roughly level, allowing a fighter to win a round by holding an opponent for most of it. The 2026 revision, effective from the start of 2026, pushed aggression and control down to tiebreaker status, used only when effective output on both sides is equivalent.

This was the biggest change to MMA rules in nearly two decades, and it came from a very specific observation: too many fights were being scored for the wrestler rather than the striker who did damage.

Alongside it, the 10-8 round was rewritten. A 10-8 round requires clear domination, beyond merely winning the round. The common interpretation in judging circles calls this the three Ds: damage, dominance, duration. All three are qualitative. There is no quantitative threshold for the word damage.

That is where every controversy begins. It is not that judges are biased, but that the system requires them to convert a continuous quantity into an integer.

The round is the unit of currency, and the smallest coin is one point

Try a simple calculation. A round ends with damage split 55-45 in favour of fighter A. In any sport using a continuous scorecard, that is a narrow margin. In MMA, it rounds to 10-9. If that round were 90-10, it would still be 10-9. Two rounds that are worlds apart in nature receive the same figure.

The first consequence is aesthetic: a fight in which fighter A wins three razor-thin rounds is recorded as 30-27, identical to a fight in which fighter A lands without being touched across three rounds. On paper, two performances a world apart become the same result.

The second consequence is the part worth pausing on. In a close three-round fight, it takes only two rounds leaning one way and one leaning the other for the result to be 29-28. But if a judge, through a small difference in viewing angle, flips one of the first two rounds, the result becomes 29-28 for the other man. The same fight, the same volume of events, two opposite outcomes. No intermediate threshold exists to say the fight was really a draw.

I often recall a story from another sport here. In the summer of 2026, at a semi-final at Wembley, a single collision inside the penalty area became the penalty kick that decided both the match and the tournament. One moment of contact, lasting less than half a second, was elevated into an event as weighty as the ninety minutes before it. In MMA, the smallest unit elevated to equivalent status is not half a second of contact, but a fifteen-minute round compressed into a single digit.

Three seats, three frames

Every MMA fight has three judges. They sit at three positions around the cage. One watches from the side that matches the broadcast camera angle. One watches from the opposite side, where the cage post, cabling and the bodies of officials block part of the view when fighters work against the fence. The third often sits lower, near the floor, where body shots and leg kicks are seen more clearly than head strikes.

A judge watching from a low angle will see the force of a kick to the ribs. A judge watching from above will see the head snap back after a right hand. The same kick, two levels of damage in the eyes of two people sitting four metres apart.

Across years of tracking disputed decisions, I always ask one question before reaching a conclusion: where was that judge sitting? How many times was their line of sight blocked in that round? In how many rounds was the broadcast camera on the opposite side of their seat?

This is why a three-second clip circulating online can be right in image and wrong in score. Fans see the broadcast director's frame, chosen to tell a story. Judges see the labour contract's frame, assigned by seat.

How a scorecard is actually filled in

The mechanics of filling in a card are rarely discussed, yet they determine how error forms. Judges do not record a running total mid-fight. They write each round into a separate box. When a round ends, they have roughly fifteen seconds to write three digits for the rounds already completed, usually after holding a conclusion in their head for the preceding five minutes.

There is no mechanism for judges to compare notes during the fight. No draft sheets are collected. No explanation is filed for why a round was scored 10-9 rather than 10-8. The card is folded, handed to a commission official, and opened only minutes after the final horn.

This creates a motivational paradox. Because no one is required to explain, no one has an incentive to write an unusual score. Within judging circles, issuing a 10-8 is treated as an attention-seeking act, while issuing 10-9 for every clear round is treated as the safe act. Safe always wins.

I once sat beside a retired combat sports judge at a university talk. He said that across his career he wrote a 10-8 only a handful of times, and every one of those times led to a phone call of explanation with the promoter the next morning.

Jones versus Reyes: one round, three readings

On February 8, 2026, at UFC 247 in Houston, Jon Jones defended his belt against Dominick Reyes. The three cards read 48-47, 48-47 and 49-46. The first two caused little argument. The third did.

For a judge to score 49-46, that person must give Jones four of the five rounds. Across the first four rounds, Reyes threw more and controlled the distance for most of the time. Jones produced the heaviest strikes of the fight, including a kick to the knee that buckled Reyes, and a submission attempt in the fifth round that nearly ended it.

Four digits. 48-47 says the fight was narrow. 49-46 says the fight was clear. One fight, two messages, and both sit within the margin of a single rounding operation.

Holloway versus Volkanovski: split by one round

On July 12, 2026, at UFC 251 in Abu Dhabi, Max Holloway and Alexander Volkanovski went five rounds. The result was a split decision: 47-48, 48-47, 48-47. Volkanovski won.

All three cards agreed on four rounds. The second round was the only point of difference. Two judges gave it to Volkanovski. One gave it to Holloway. One round, one point, one championship belt.

The notable thing is structural. The controversy did not come from three judges seeing the whole fight differently. It came from one round, where Volkanovski's low kicks landed on the thigh and Holloway's hooks landed on the midsection. Two different kinds of damage, with no conversion unit.

Here I often hear fans say that you can tell who won by looking at who is more beaten up. The problem is that damage does not display symmetrically. Leg kicks accumulate across four rounds and only surface in the fifth as a step half a beat slower. A punch to the head displays immediately as a shake of the head. What displays immediately is remembered more than what displays late, regardless of true value.

Two months in late 2026: two opposite directions

Back to the fight from the opening. On December 10, 2026, Paddy Pimblett met Jared Gordon. All three cards read 29-28 for Pimblett. Across three rounds, Gordon controlled the pressure and threw more in two of the three rounds according to publicly available records from specialist MMA statistics sites. Pimblett produced the more visibly concussive strikes, particularly in the second round.

Two months earlier, on October 22, 2026, at UFC 280, Sean O'Malley met Petr Yan. The result was a split decision, 29-28 for O'Malley. Yan threw more and took more control time. O'Malley produced the heaviest strike of the fight, a kick to the face in the second round that made Yan pause.

Two fights, two opposite directions. In the Pimblett fight, damage was prioritised over volume. In the O'Malley fight, volume was roughly prioritised. People use these two fights to argue that judges have no standard. I read them differently: both sit within the error margin of a system in which two very different levels of damage can carry the same 10-9 value.

Who appoints the judges

A detail few fans notice: promotions such as UFC, ONE Championship and PFL do not appoint judges. State or national athletic commissions do. The promotion rents the venue, pays for the commission's presence, and signs a contract with the arena. The person holding the pencil is appointed by the state.

This creates two opposing consequences. On one hand, it separates scoring power from the promoter's wallet, an important safeguard many other sports lack. On the other, it places scoring in the hands of a very small, frequently reused group working a dense schedule, with no obligation to publish the reasoning behind any card.

In California, since 2026, the state athletic commission has applied training and periodic assessment requirements for MMA judges, including reviews of disputed decisions. It is one of the few states with a formal review mechanism, and one of the few places where judge performance data is stored in a searchable way.

In most other markets, a judge can work hundreds of fights across a career without any public performance summary. Fans know their names, know their disputed cards, but no formal mechanism exists to compare their accuracy rates across events.

A different system in Asia

ONE Championship, headquartered in Singapore and heavily active in Asian markets, has repeatedly adjusted its scoring system to emphasise damage and the completeness of finishes, while reducing the weight of control time.

In principle, this is a commendable effort to reduce wins by holding. But structurally it does not solve the rounding problem. It only changes the weights of the variables before they enter the same rounding function. The round remains the unit of currency. Three rounds remain a three-round fight. No fraction appears anywhere.

This explains why scoring controversies persist under every ruleset. A system favouring damage creates a class of unjust losers among control-based fighters. A system favouring control creates a class of unjust losers among counter-strikers. The problem is not which weight to choose, but that every weight gets compressed at the final step.

The lesson from the season without crowds

In 2026, when arenas closed and the UFC ran events in Jacksonville and then at the Apex in Las Vegas with empty stands, something unexpected happened. The 2026 season had no crowd, but it had a very large ear.

There was no crowd noise to react on the judges' behalf. No collective intake of breath when a heavy shot landed. Judges had to decide for themselves, without social feedback to guide them. And in that near-silence, what they could hear was the impact of strikes — a damage signal that noise normally conceals.

During that period, fights at the Apex tended to be scored in favour of whoever produced the heavier sound. It is a small observational sample, insufficient for conclusions, but enough to show one thing: scoring criteria are not read in a vacuum. They travel through ears, through eyes, through seating, through surrounding noise.

This is also why I distrust analysis based only on clips. A clip has no crowd audio, no judge seating position, no waiting interval between rounds. It is an edited version of the fight, not the fight.

Betting markets and the lag of information

Another variable often ignored in scoring debates: betting odds. Pre-fight lines reflect the collective expectation of people putting money down, and in many cases the line moves during the fight in ways that reflect the majority reading of each round.

When the result is read out and diverges from the live line, the market reacts instantly. Sportsbooks adjust, pundits publish, and within hours a new narrative is built. I have tracked these lines across multiple seasons, and what I noticed is that they tend to reflect the same bias as judges: a preference for what displays clearly and immediately.

In combat sports, bookmakers adjust odds based largely on broadcast imagery and publicly available statistics. Neither source measures the force of a strike. They count contacts, classify by body region, and in some systems distinguish heavy from light strikes based on a human recorder's estimate. No device measures impact force.

The result is that judges, markets and fans are all scoring through the same class of proxy — something that measures an approximate aspect of damage, then gets used as though it were damage itself.

Why this is a technical problem, not a moral one

When a disputed decision appears, the default reaction is suspicion of the judge: bias, corruption, incompetence. I think most common cases are far more technical.

There is a single number that can be adjusted without changing the structure of the rules: the frequency of 10-8 rounds. If judges were willing to issue 10-8 when a round is genuinely clear, the 55-45 problem would solve itself. A round in which one fighter dominates entirely and causes clear damage should be marked differently. But most judges avoid the 10-8 for psychological reasons: it looks too harsh, it makes the losing fighter feel humiliated, and it has never been strongly encouraged by commissions.

The result is that the entire spread of a fight is compressed into two digits: 10 and 9. Every clear round becomes 10-9. Every close fight becomes 29-28. And once compressed like that, a one-round error — nearly unavoidable when three people watch from three angles — immediately becomes an error for the whole fight.

Another line of analysis deserves consideration: adding criteria does not make scoring better. The 2026 revision, with the three Ds for 10-8 rounds, made the system more coherent on paper while widening the interpretive gap. The more qualitative criteria there are, the more degrees of freedom exist before the digit is written.

There is a line I use often in talks with students: discipline is not punishment; discipline is a way of reading a fight. And in this sport, the only verifiable discipline sits on the scorecard — the one thing whose reasoning we are still not shown.

What I want to see next season

What I want to see next season sits with transparency, specifically a small procedural change: publishing one line of reasoning per card, along with the judge's seat number. Three lines explaining a decision cannot change a result, but they turn every controversy into checkable data.

Suppose that in the Holloway versus Volkanovski fight, the second-round score came with a short note: the judge in seat two saw three right-leg kicks land with force. The argument then stops being an argument about one judge's credibility. It becomes an argument about whether three low kicks are worth two hooks to the body. That is a technical question, and it can be answered with data.

A second change could apply more broadly: open scoring round by round. This has been trialled by some commissions and at some international events, with results recorded on two sides. The upside is that fighters know where they stand and can adjust tactics. The downside is that it can reduce the drama of the final announcement and, in some cases, create new pressure on judges who know their scores will be published immediately.

Neither change solves the rounding problem at its root. They only turn an anonymous error into a debatable one. But in a sport where the final decision is usually made by three people sitting in three different corners, turning error into something debatable is the most important step available.

The law is the only thing that never enters stoppage time, but the law is also the only thing that can be reread after the final bell has stopped.

The assumption I left out

The analysis above assumes scoring criteria are the primary variable. I ignored three things.

The first is the quality of publicly available data on scorecards. Many smaller events, particularly in regional Asian and European promotions, are not fully recorded. This skews any scoring analysis toward major promotions, where there are more cameras, more recorders and more media pressure.

The second is judges' scheduling load. A judge may work four events in a month in four different cities, with significant time-zone shifts and travel time. The cumulative effect of this on decision quality has not been adequately studied in any public documentation I have read.

The third is commercial influence. A fighter being built into a public image often benefits from close cards. Proving a causal link between commercial pressure and any individual card is impossible with existing data, but ruling it out is equally impossible.

Data never commits a foul; the writer is the one who receives the card. The referee reads the fight fastest; I only write one beat slower.