Error at the Edge of the Frame: VAR Does Not Lack Technology, It Lacks Angles

In the 35th minute of the 2026 World Cup final between France and Croatia, referee Nestor Pitana stopped play, walked to the pitchside monitor and awarded France a penalty after the ball struck Ivan Perišić's hand. That night I sat in front of a screen in Chengdu, reopened the six main broadcast angles of the match and measured frame by frame. Only one of the six angles showed Perišić's arm opened in the way the law calls unnatural. And that was precisely the only angle the VAR team gave Pitana to watch, for one minute and forty-seven seconds. I found the mistake not at the centre of the pitch, but at the edge of the frame. A passage of play, to me, is never a single event. It is a chain of technical choices: who picks the camera angle, who picks the freeze frame, who picks the start point of the clip, and how many seconds of it the referee is allowed to see. Before the whistle sounds, at least four decisions have already been taken behind a closed door. VAR, the video assistant referee, may intervene in only four categories: goals, penalties, direct red cards and mistaken identity. The intervention threshold set by IFAB, the body that governs the laws of the game, is a clear and obvious error, or a serious missed incident. There is no arithmetic definition of clear. There is no centimetre in the text of the law. There is no minimum frame rate on record. Every argument about VAR grows out of that gap. In 2026, when I was 36 and working as a VAR data analyst for a sports television channel in Chengdu, I reviewed 147 controversial refereeing situations across 28 rounds of the Chinese top flight. Twelve incorrect offside decisions traced directly to camera placement. In round 25, in the match between Guangzhou Evergrande and Shanghai SIPG, Wu Lei's goal was disallowed by 15 centimetres, yet no camera sat on the horizontal plane required to measure it. I started building my own database of viewing-angle error, one passage of play per line. The geometry of football broadcasting is harsher than viewers imagine. At 32 km/h a player covers 8.89 metres per second. On a 50 frame-per-second feed, the gap between two frames is 0.02 seconds, or 17.8 centimetres. If the source runs at only 25 frames per second, that figure rises to 35.6 centimetres. The frame the VAR team uses to fix the moment the ball leaves the foot may already have missed nearly two hand-spans of movement. In an offside call, that is the whole story. The offside line does not sit at the attacker's foot; it sits on the plane of the second-last defender. If the camera is not square to that plane, every line drawn on the screen is the output of an interpolation. An interpolation can be mathematically exact and still geometrically wrong. Semi-automated offside technology was built to fix exactly this weakness. At the 2026 World Cup the system used 12 dedicated cameras under the roof, tracking 29 data points per player 50 times per second; the ball carried an inertial sensor operating at 500 Hz. Positional error fell to millimetres. But the system answers where a player is, not when the ball left the foot. The moment of contact still depends on the sensor and on the nearest frame. It took me 37 replays of one borderline incident before I understood that the human eye is not a measuring instrument. With handball the problem is harder still. The law does not forbid the ball touching the hand; it forbids the ball touching the hand in a position deemed unnatural, or one that makes the body unnaturally bigger. That is a judgement about a silhouette, not a measurement. On the Perišić incident the VAR team chose one angle and held it for nearly two minutes. One arm, six cameras, and the referee was shown one. There is another error type rarely discussed: presentation error. When the screen shows the words offside by 3 cm, viewers assume an exact measurement. In most cases what sits behind the words is an image-based interpolation with an uncertainty band that can run to tens of centimetres. The words are not wrong; they are simply narrower than reality. At the tracking-data layer, distance covered and sprint counts are packaged as effort indicators, and running without purpose still produces handsome numbers. An analyst who trusts the appearance of data will soon be led by it; a decent one always asks where the data came from. The 2026 season played out in empty stadiums and taught me something else. Without crowd noise, referees could hear contact, studs, breathing, and players shouting from distances they normally could not hear. Tracking domestic matches played behind closed doors, I noticed the number of on-field reviews rose sharply: with no crowd applying pressure, referees were bolder about admitting they needed more information. The empty stadiums of 2026 showed me this: VAR does not rescue football, it exposes football. One variable is routinely ignored: working conditions. A standard VAR room holds four to six monitors, a replay operator, an assistant and a lead video referee, working continuously for more than 90 minutes under pressure to answer before play moves to the next incident. Semi-automated offside needs roof-mounted camera infrastructure and a sensor inside the ball. Not every league has them. The result is one law and one incident, but evidence of differing quality depending on a competition's budget, an inequality the league table never reflects. At this point the story leaves technique and enters psychology. When a stand roars for a review, most of those roaring are not asking for more data. They are asking for an angle that confirms what they already believe. We think we are hunting justice; in fact we are hunting a prettier camera angle. My trade taught me that technology does not erase subjectivity, it relocates it. Once, the decision-maker was the referee on the pitch, with his eyes and one viewpoint. Now the person choosing what the referee sees is the video operator, with a source list and a timeline. Power does not vanish; it changes seats. And when power changes seats, accountability has to move with it, otherwise every post-match explanation is no more than ritual. The most workable reform does not lie in a new sensor. It lies in a public process: publish the set of camera angles and timestamps used for every review, replay the clip on the stadium's big screen so spectators watch it too, and open the audio between referee and VAR team, as some competitions have begun to do. When the angle set is public, viewers can judge the quality of the evidence instead of judging only the outcome. Who is willing to publish the list of their own camera angles, knowing that another viewpoint could overturn what they have just declared?