Three-Two: The Split Vote That Exposed the Handball Law's Fault Line

There is a moment in every match when the sugar rush ends and the truth begins. Last Saturday in Manchester, that moment arrived about an hour after the final whistle.

Saturday afternoon. The sofa in my living room, a cup of tea beside me, two screens on the laptop — the match on one, and on the other that old Zoom call, eight Manchester pub regulars and two friends from Dhaka. The Zoom party started in 2026, when the stadiums were empty, and the 70% crowd theory grew out of it. The crowds are back now, the stands are full, but I never closed the call. Habit.

The match was Manchester City against Sunderland. The score was 5-3. Eight goals. An eight-goal game normally keeps people talking about attacking play for a week. Not this time. The most discussed subject of an eight-goal match became a forearm.

The incident was simple. Inside the City penalty area, Elliot Anderson was contesting the ball with Enzo Le Fée. The ball touched Le Fée's leg, deflected, and struck Anderson's arm. Referee Rob Jones watched, then waved it away — nothing doing. Sunderland's players surrounded him. From the bench, manager Regis Le Bris stood up, visibly furious. At Stockley Park, VAR Michael Salisbury looked at the replays and stayed silent.

That silence came back as the biggest question of the week.

After the match, the pub chat, the Zoom call, the WhatsApp groups — everyone said the same thing: "It hit his arm, didn't it?" I have been watching football since before the backpass rule, and this still felt new. It felt new because the question is no longer a football question. It is a question about video evidence. We all watched the same replay, the same angle, the same speed — and we still could not agree.

I didn't unsee it.

Then, midweek, came the thing this piece is actually about. The Premier League's Key Match Incidents panel — the body that reviews controversial decisions and publishes its votes — announced that Jones's decision was supported 3-2. And that Salisbury's non-intervention was supported 4-1. In the same week, in Everton against Ipswich, VAR had disallowed a goal; that intervention was also supported 3-2.

Three-two. Four-one. Three-two. Small numbers, but they are now the loudest numbers in English football.

What the panel is, and what it is not

Let me clear one thing up, because there is a lot of confusion about this panel. The Premier League Key Match Incidents panel has met regularly since the 2026-23 season, reviewing controversial decisions matchday by matchday and publishing the voting outcomes. It contains three former players or coaches, one Premier League representative and one PGMOL referee representative. Five people. Hence 3-2, 4-1 or 5-0.

This panel is not a court. It cannot change a result, dock a point, or even sanction a referee. It is an accounting ritual — an open ledger in which, every week, someone writes down who thought what.

But there is a structural subtlety here that most discussion misses. The panel does not answer one question per incident. It answers two. First: was the referee's on-field decision correct? Second: was VAR right to intervene, or right not to intervene?

There is no obligation for those two answers to match. And that is exactly what makes the Anderson incident interesting.

Two questions, two votes, and the gap between them

On the field, Jones's decision — no penalty — was supported 3-2. Two of the five believed a penalty should have been given. The panel's explanation noted that the ball deflected off Le Fée's leg at very close proximity before striking Anderson's arm, and that two panel members felt Anderson made a deliberate secondary arm movement.

Now look at the second vote. VAR's non-intervention was supported 4-1.

Same incident, same panel, two questions, two different numbers — 3-2 and 4-1. What does that gap say?

My reading is this: the panel is not saying Jones was perfect. It is saying his decision fell inside the range of acceptable interpretations, and that VAR's silence was clearly correct. In the panel's language, the referee's call was contested but legitimate; the VAR's call was uncontested.

That two-tier judgement is the real story. Because the VAR protocol's threshold for intervention is a "clear and obvious error". That phrase is far easier to say than to apply. It is a judgement about a judgement. It does not ask "what do I think"; it asks "what did the referee think, and was that thought so wrong that staying silent was impossible".

That second question has no metric. No number, no threshold line, no calibration table. Only the memory and taste of five people.

Le Fée's leg, Anderson's arm, two pairs of eyes

The history of the handball law matters here, because this is the knot. In 2026-20, IFAB made a major change: contact with the arm at or above shoulder level would be an offence almost automatically. In 2026-21 came the clarification on "unnatural arm movement" and the idea of the arm making the body "unnaturally bigger". In 2026-22 came another amendment: accidental handball by a teammate in the build-up to a goal would no longer be penalised.

Read those amendments together and one thing becomes clear: the lawmakers keep trying to draw a line between intent and accident. In the Anderson incident, that exact line was tested.

The ball struck Anderson's arm — nobody disputes that. The dispute is whether the arm was there for balance, or whether it was a conscious second movement. Three of five said balance. Two said conscious.

Same video, same frames, same speed — and two experienced football people saw intent while three saw physics. That is the entire handball crisis in one sentence.

The law writes the word "deliberate". A camera does not measure deliberation. A camera measures movement. And the distance between movement and intent has to be closed inside a referee's head, where no VAR screen can reach.

You cannot get inside a goalkeeper's head

Look at the second incident of the same week. Everton against Ipswich. Everton had a goal disallowed because Johnson was in an offside position and, from close range, moved in a way that affected goalkeeper Christian Walton's ability to play the ball. The panel supported the intervention 3-2. Two panel members believed Johnson's movement did not affect Walton at all.

Law 11 says a player in an offside position is penalised if he challenges an opponent for the ball, or clearly attempts to play a ball, or makes an obvious action that clearly impacts an opponent's ability to play the ball.

Note that the word "clearly" appears twice. And clarity is precisely what was missing.

Goalkeeper interference is the most subjective corner of offside law, because deciding it requires you to model a goalkeeper's reaction time, reach and body angle. There is no data for that. Nobody has ever measured, from which angle and at what speed a specific goalkeeper could have reached a specific shot in a specific millisecond.

So what is the panel measuring? It is measuring its own imagination. That is a legitimate method. It is not evidence.

I didn't unsee it.

The control group: where there is no video, there is no history

Now my old habit. When I am uneasy about a decision in a big match, I look downwards — to the lower leagues, to county grounds, where there is no VAR.

This week I thought about a semi-professional match I have sat through many times, where the same scene plays out. The ball deflects onto a defender's arm. The referee gives a corner. There is a protest, and two minutes later everyone has forgotten. By Tuesday nobody remembers the incident at all.

Did no error occur there? It may well have. But it was never written down. No record, so no argument. No argument, so no crisis.

To me this is like floodlighting. VAR did not create error. VAR created a permanent archive of error. Ambiguity used to be instantaneous; now it is permanent. Once, the whistle ended the argument. Now the argument starts after the whistle, because the evidence hangs on the screen.

I bring in Bangladesh, because that is my own experience. Watching from Dhaka, I saw national team matches and domestic league games where the referee's decision was final because there was no camera. The crowd's anger had nowhere to go; it left with the match.

Now anger is archived. It becomes a highlight clip, a thread, an entry in a panel's file. That is the biggest structural change in how football feels — not nostalgia, just an observation.

A public ledger: a permanent record of ambiguity

I have used one metaphor about this panel for a long time, and people laugh it off: it is an open book, a public ledger. Every week a few entries are written. Each entry is timestamped. Each entry is permanent, without correction. Nobody can delete it or edit it. At the end of a season you can look back and see which way the panel leaned on close-range deflections — consistently, or by mood.

That is the real value. Before 2026 there was no way to answer that question. Now there is. That is genuine progress, and I concede it.

But an open ledger has a problem. A ledger tells the truth, and a ledger cannot make a decision. The panel can write "3-2"; the panel cannot return a point. So a sporting grievance becomes a media grievance — and media grievances last much longer.

This is where live data's dark side enters. Within seconds of the Anderson non-call, in-play markets moved, because live feeds are wired into bookmakers' models, and models work on probability rather than decision. A referee waving his arms is simultaneously three things: a refereeing decision, a broadcast product and a trading event. When those three happen at once, transparency does not win easily, because transparency has a market price too.

Where I could be wrong

Now the part hot-take people usually skip.

Suppose I am wrong. Suppose the 3-2 split is not a crisis but a sign of a healthy system.

That argument is not worthless. Imagine the panel returned 5-0 every week. It would become a rubber stamp. Whatever the referees said, the panel would endorse — and then what would review mean? Publishing dissent means admitting dissent exists. And no credible system survives without admitting that.

I am a 69-year-old who learned the game when the whistle ended the argument. To me, dissent feels like weakness. In today's football, dissent may mean the process is working. That is the second truth, and I cannot shrink it.

But a question remains. If transparency is not tied to a decision rule, transparency means only publishing disagreement. The panel says each week: we do not agree. Fine — but who settles what the law says? Not the panel. IFAB does, once a year, and even then through language rather than measurement.

So we sit in an uncomfortable place: the law is written in words, the protocol is written in words, and application runs on taste. In that system the panel is a mirror, not a wall. A mirror shows a face; it does not change it.

One word on the 5-3

I could sit here and write that the 5-3 proves a City defensive-transition weakness. I could, and I will not. Eight goals in a match is a mood, not a finding. Without xG, shot maps or possession data, tactical analysis means dressing guesswork as evidence. I do not do that.

What I can say: if the centre of the conversation after an eight-goal match is a forearm, that says less about football and more about our attention economy. The bigger the game, the smaller the decision that generates the most noise.

What to watch next

Here is a testable prediction, because opinions without measurement are cheap.

Watch the next three close-range deflections inside the penalty area. If even two of them come back with split votes like 3-2 or 4-1, then the problem is not the law — it is the threshold. The phrase "clear and obvious error" needs a numerical or at least documented definition, or the same story returns every week.

And if you see several similar incidents suddenly supported 5-0, then the panel is not reading the law; it is reading public opinion. That is worse news.

I am also watching for a PGMOL audio release. If the few seconds of conversation between Jones and Salisbury are published, we will learn in what language Salisbury decided not to intervene — "there is contact but it is natural", or "there is contact but the evidence is not enough". Those two sentences are miles apart. The first interprets the law; the second interprets the protocol. And the gap between the panel's two votes on the Anderson incident is exactly the gap between those two sentences.

Whether the FA acts on Le Bris's anger is worth monitoring too, though it is not the main story. The main story is a system that publishes its own doubts every week and still cannot resolve them — because the people with the power to resolve them do not cast votes.

I didn't unsee it.

Before the match I wrote in my Zoom group: if there is a handball controversy today, we will talk for three hours and reach no conclusion. That is exactly what happened. The only difference is that we could not agree in three hours, and the panel could not agree with five people. Who is more honest, I do not know.

I know one thing. In that moment on the pitch, when the ball struck Anderson's arm, forty-seven thousand people around him took a breath at the same time. That breath was the anticipation of a decision. Today that breath has become an archive, locked in a file. That is exactly where football stands — emotion instantaneous, judgement permanent, and a gap between them that no panel can fill.