Before El Clásico: A Debut Referee, Two Small Samples, and the Illusion of 14 Percent

The appointment sheet carried a name, and the name came with a table. Javier Alberola Rojas, thirty-five years old, one name in Spain's admittedly long list of referees — and a first-ever El Clásico appointment. The table laid out his record: thirteen Barcelona matches, ten wins, two draws, one defeat, seventy-seven percent. Nineteen Real Madrid matches, twelve wins, three draws, four defeats, sixty-three percent. Two further numbers floated alongside: 3.6 yellow cards per match, 0.32 penalties per match.

The figure that travelled fastest was the fourteen-point gap. The figure that travelled slowest sat in small type at the edge of the table: thirteen and nineteen. A referee's biggest career night was being decided, and the weakest number in the preparation was the one being said loudest.

A 3 a.m. alarm, three-day-old news

The table came out of a news clip published on 23 October, which lists the fixture for 26 October, three in the morning on a Vietnamese clock. In Indian time that is the same alarm at roughly half past one, the one every European-football viewer on the subcontinent has been switching off for two decades. The news is three days old; the match is zero days old. Whatever gets written in that gap is necessarily inference — and when inference is dressed as a table, the reader starts treating it as evidence.

An El Clásico referee occupies a post where an entire career compresses into three or four decisions. One contested collision in the penalty area, one offside that VAR touches, one minute of added time, two sentences exchanged with the captains in the tunnel. The other ninety minutes are administrative work for him; those four moments are the coverage. That is why the appointment is itself a story, and why every number printed beside it circulates like a pre-verdict.

The structure of the source gives the same rush away. One lead item, four one-line items beneath it — the standard clip-journalism format, built for recency, not for analysis. Only the appointment itself is attributed, to a named journalist, Edu Garcia. But the four numbers that carry the actual controversy — thirteen matches, nineteen matches, 3.6 cards, 0.32 penalties — appear without any primary source. A table that cannot be defended if challenged.

In Pune, across fifteen years of switching off that night alarm, my notebook has often been open. The habit has not changed: after a match, write the decisions, not the score. Which card came in which minute, which appeal went unheard, which team collected two warnings inside the first ten minutes. One season fills the page, and then you begin seeing what the scoreboard does not show. When I later applied the same method to umpiring data in tennis, the result was identical — the number that looks most compelling is usually the least reliable.

Thirteen and nineteen: the arithmetic of the sample

The first thing that should accompany any such table, and does not appear here, is the uncertainty band on either side of the percentage. Ten wins from thirteen means one match shifts the number by roughly eight points. Nine wins would have printed sixty-nine, not seventy-seven. One penalty, one red card, one stoppage-time goal, and the headline changes.

Take an approximate ninety-five percent confidence interval. Ten from thirteen spreads across roughly forty-nine to ninety-two percent. Twelve from nineteen spreads across roughly forty-one to eighty-one percent. The two bands lie on top of each other — the entire stretch from forty-nine to eighty-one is shared. The fourteen-point gap that became the headline sits inside that shared zone, in the region where the two figures cannot be called different at all.

The gap between thirteen and nineteen matches is statistical noise; calling it evidence of a referee's temperament is a leap past the evidence. This is no technicality. It is the same error tennis repeats every season — "this player has a poor record with this umpire" — when the record is four or five matches and three of them involved an injured player.

The base rate: the variable missing from the table

The second and more serious omission is that the table carries no ordinary win percentage for Barcelona and Real Madrid. In La Liga, both clubs beat the other nineteen teams most of the time, whoever is officiating. If Barcelona's normal rate is around seventy, then seventy-seven tells us nothing about that referee — it tells us about Barcelona. And sixty-three, if Real Madrid's normal rate is higher than that, argues in the opposite direction: that Madrid has performed below its own average under this official.

A referee's win-loss record mostly describes the clubs, not the referee. Economists call this an omitted-variable problem, and it is the commonest flaw in sports journalism precisely because it looks like analysis. You take a dependent variable, choose an independent variable, and leave out the third variable — team strength — that was determining everything.

The third layer is selection bias. Referees are assigned matches; they are not dealt them at random. No official collects thirteen or nineteen fixtures from a lottery. He is posted to games that fall into a risk category, that fit a regional equation, that carry pressure from one team's recent form. The sample is pre-filtered, and the map of that filter is not in our hands. Without the map, calling any conclusion a "referee effect" is raising a building whose foundation nobody has seen.

Cards and penalties: the referee's own signature

The other side of the picture is that two numbers in that clip do measure the official's own behaviour rather than the clubs' performance: 3.6 yellow cards and 0.32 penalties per match. In principle these are useful indicators — they say how much he lets a game breathe and how readily he intervenes in the box.

But reading them requires a third figure that was not supplied: the La Liga average. If the league produces five cards a match — a commonly cited figure, which I am leaving unverified here — then 3.6 places him at the lenient end. If the league average is 3.8, the number is inert. Without a baseline, any claim of "lenient" or "strict" is a statement that cannot even be falsified, because no scale has been declared.

The two numbers that measure the referee's own behaviour are the two most weakly documented — no source, no baseline, no match-level detail. The irony is plain: the thirteen and nineteen were useless even if trustworthy, and 3.6 and 0.32 would be useful if they were complete.

Tennis holds up the same mirror

I have spent years assembling chair umpires' overrule rates and Hawk-Eye challenge success rates, because the question is the same there — does an official treat one player differently? After five seasons of data, the most honest finding was that an umpire's personal record was better explained by the player's court position and challenge habits than by the umpire's intent. The structure repeats here: the number that builds the most attractive story is usually the most confounded variable.

Tennis does one thing better. Nobody prints a statistical table before an umpire's first Grand Slam final, because the administrative record is kept differently. In football the record is printed first and explained afterwards — the order is reversed.

One line that tells you the reliability of the whole clip

Four one-line items sit inside the same clip. One says coach Fabian Hürzeler is not keen on taking charge of Manchester United as a replacement for Michael Carrick. The roles in that sentence do not line up with publicly known positions. Whether it is a translation failure or a syndication lapse, the message reaching the reader is the same: one of ten items here cannot be repeated as fact.

A clip's quality is measured by its weakest sentence, not its strongest. If the roles are scrambled in one item, no reason has been given to trust the other nine — a structural risk rather than editorial laziness, especially when the clip's most influential item is an unsourced statistical table.

One item in that same list is more durable than the entire file: Lionel Scaloni's commitment to Argentina until 2030. That is a long-horizon signal, still discussable three days later. The Mourinho item is its opposite — personality-driven, traffic-friendly, analysis-free. A single clip holds news of three different shelf lives and serves them at one speed.

The number that sells, and the question that matters

Read against the grain, the table yields one hard fact that is not a percentage: at thirty-five, after that many seasons, this official had never had an El Clásico. That is a design question about a system, not a data question. How does a federation build officials such that the heaviest fixture goes to a man with no experience of that level — and is that courage or carelessness? Answering it requires the training ladder, not a win-loss column.

There is a media effect nobody discusses: printing a referee's club-by-club record three days before a match changes the match's atmosphere. Every marginal decision now confirms a story already written. Wave away a Barcelona penalty appeal and the table is recalled; award a Real Madrid one and the same table is recalled. That effect is wholly independent of the official's actual performance.

And the crucial point: the clip contains no allegation. "Real Madrid is worried" belongs to the headline, not to any quoted person. No official, coach or club spokesperson expresses concern anywhere in the material. Drawing the conclusion of concern from a fourteen-point gap has been outsourced to the reader, and legitimised by the table's neutral appearance. Framing is not an accusation — but framing builds an atmosphere, and atmosphere presses on decisions.

What to actually watch on 26 October

If you are up at half past one, forget the table and count three things. How he handles the first penalty-box appeal inside twenty minutes — an immediate signal, or a deferral to VAR. The minute of the first yellow card, because that is the real reading of his tolerance, not a season average. And the tone of his exchange with the captains before half-time; an official who establishes dialogue changes fewer decisions in the last twenty minutes.

After counting those three, you will hold real data on that referee from one match — worth more than any table of thirteen and nineteen, because it was taken under the very pressure the table claimed to describe. Next season, when another clip prints another referee's record, the first question should be the sample size, the second the base rate, and the third — never yet asked — whether the number is about the referee at all or about the teams he happened to be assigned. The day those three questions appear beside the table, the fourteen-percent illusion ends on its own.

El Clásico runs through the night and is over by morning. Its referee will work for years afterwards, and his real record has not been written yet — it will be written on the night of 26 October, in the first twenty minutes, on the first appeal.