The Referee's Make-Up Call in the Regular Season: The Numbers Behind the Table

Minute 89, Round 24 of the K League 1 regular season. I sat in the newsroom, the screen split into four panels: the main broadcast feed, the VAR feed, the goal-line camera, and the sideline camera. The center referee ran to the monitor at the edge of the pitch. Three seconds earlier, a striker had gone down inside the box after contact with a home defender. The stands erupted. The broadcaster had already shouted "penalty" before the referee made his decision.

I didn't watch the fall. I watched the referee's eyes as he reviewed it. The goal-line angle showed the defender touching the ball first, but his left knee had clipped the striker's thigh before the ball rolled away. The referee looked twice, then pointed to the spot. In my digital log, I wrote: "Minute 89, K League 1 Round 24, third penalty decision in four matches favoring the home side, 0.34-second deviation from the original frame."

That was when I realized something the table doesn't show: the make-up call is returning, and VAR is the camera that cannot be erased.

VAR entered world football in 2026, after FIFA approved trials at the World Cup in Russia. Since then, the technology has run for more than six seasons in the top domestic leagues. The question referees still argue over is not "does VAR work" but "does VAR make referees more consistent, or does it simply move the error from the pitch to a closed room".

To answer that, I went back to the dataset I began building in 2026. It started as 1,247 referee decisions from the 2026 World Cup and three K League 1 seasons. This regular season, I expanded it to 4,318 situations across five leagues: K League 1, J1 League, the Premier League, La Liga, and Serie A. Each situation is coded along four variables: decision type (penalty, card, offside), match timing, the benefiting team, and the match result that followed.

I don't watch the goal; I watch the camera that watches the goal. My working principle since the summer of 2026 in Kazan has stayed the same: every judgment starts with raw data, not with the emotion of the stands. When South Korea beat Germany in Kazan through Kim Young-gwon's VAR-approved goal, I did not celebrate. I downloaded all 64 matches and logged every reversed decision. That 47-page diary became the foundation for everything I have written since.

Six years later, I still keep that habit. Every matchday, I record the timing, the referee's position, the number of monitor reviews, and the gap between the contact and the ball going dead. These numbers never appear on the news ticker, but they decide who plays in the AFC Champions League next season and who gets relegated.

Of the 4,318 situations I coded, 1,092 are penalty decisions. The distribution by time and context matters more than the raw count.

Penalties awarded in the final 15 minutes account for 41%, more than double the actual share of that window (about 17%). In other words, referees award penalties from minute 75 onward at 2.4 times the rate of the first half. That figure is not new to analysts, but the explanation has changed. It used to be blamed on teams pushing higher for a winner. Yet when I isolate the "one-goal margin" and "two-goal margin or more" variables, the gap remains: even when a match is settled, late-penalty frequency is still 1.8 times higher than in the first half.

That led me to the make-up hypothesis. In 2026, I found a suspicious pattern: after a team suffered a wrong decision, referees tended to award them a "soft" penalty within the next two matches, at an 89% rate. I wrote a 30-page analysis then but never sent it, afraid I was missing some statistical detail.

This regular season, I re-tested the hypothesis on a larger set. The result: the make-up rate fell to 71%, but a new variant appeared. Instead of compensating with a penalty in the next match, referees compensate by going easier on cards: a team wronged by a VAR error has a 34% lower chance of receiving a yellow and a 52% lower chance of a red in its next match. The mechanism has not disappeared; it has shifted from heavy punishment to light punishment.

Why does this matter for the regular-season table? The regular season is a marathon, not a knockout series. A make-up penalty in Round 8 is not the same as one in Round 34. Late in the season, when relegation pressure and continental-qualification races peak, every point is worth many times more than in the opening weeks. If the make-up effect compresses into the run-in, it becomes a form of intervention in the race.

Take this K League 1 season. A club fighting for an AFC Champions League berth suffered two unfavourable VAR decisions in consecutive rounds. In the third round, it was awarded a penalty in minute 87, in a situation my model gives only a 38% chance of being called in the first half. The club scored, took three points, and climbed to third in the table. Without that penalty, it would still sit outside the top four.

I am not saying referees act deliberately. I am describing a psychological mechanism documented in decision-making literature: humans tend to compensate when they feel they owe something. Referees are human, even when trained to separate emotion from decision. And when VAR records every error, that sense of debt does not vanish — it is transformed.

Another variable I coded is "number of monitor reviews". On average, referees review 2.3 times for a penalty situation. That distribution shifts by round: over the first 10 rounds, the figure is 1.9; over the last 10 rounds, it is 2.8. Referees review more but call fewer — meaning they are more cautious as the season reaches its decisive stretch. That seems positive until you look at what follows: decisions made after multiple reviews have a 22% higher rate of formal complaint than decisions reached quickly. Reviewing more does not mean being more correct.

This is where VAR changed the nature of the job. Before 2026, a referee made a mistake and the match continued. After 2026, a referee still makes a mistake, but the whole stadium and millions of viewers watch him review the monitor before making it. The pressure does not drop; it changes form.

In my dataset, one pattern stands out in Serie A. Italian referees award the most penalties among the five leagues (0.31 per match) but also see the highest share of decisions overturned by VAR (14%). Where referees intervene most, technology intervenes most too. In K League 1, the penalty rate is lower (0.22 per match) and the overturn rate is lower (9%). The trend suggests that refereeing culture in each league shapes how VAR is used, not the technology alone.

Raheem Sterling at Euro 2026 remains my clearest example. In minute 104 of the England-Denmark semi-final at Wembley, Sterling went down in the box under contact from Joakim Maehle. Referee Danny Makkelie pointed to the spot. In my dataset, Makkelie had awarded four penalties for box contact in his previous five matches, always on the principle of "live ball". That decision was no surprise to a data reader; it was only a surprise to television viewers.

Data never makes mistakes; the writer is the one who gets carded. And I must admit a blind spot in my own analysis.

In 2026, I used the make-up model to predict referees would limit cards in the World Cup Qatar group stage to protect the flow of matches. The model was right in 26 of 36 matches. It collapsed entirely in Netherlands-Ecuador on 29 November, when the referee issued eight yellows and awarded two penalties. I had ignored the "pressure of an early home-nation exit" variable — a social-psychological factor that does not exist in a spreadsheet.

That lesson haunts me. Every referee model has an omitted assumption, and it usually sits in the human part. When I cite a 71% make-up rate, I am not claiming every referee behaves that way. I am saying that in my dataset, the pattern is large enough to be ignored at your peril.

A referee reads the match fastest; I only write one beat slower. Being one beat slower is the condition for seeing what the fast runner cannot. But it also means I might see a pattern where only coincidence exists.

Discipline is not punishment; discipline is a way of reading the match. If the make-up effect is real, the fix is not to remove human referees but to change the incentive structure. One concrete proposal: publish the full VAR log within 24 hours of each match, with written reasoning from the referee. When decisions are recorded and openly audited, the sense of debt becomes harder to sustain.

The direction of the coming regular season depends on whether federations dare open the VAR black box. And when they do, I will be the one reading one beat slower, log in hand.