Since the 2024/2025 season, LaLiga has been using semi-automatic offside technology, a FIFA-approved system that had already made its debut at the 2022 World Cup in Qatar, in the Champions League and in Serie A. Every controversial decision reopens the same debate: is this technology perfect? At STT Systems, we work every day with motion capture and analysis systems, so rather than delving into the specific incident in question, we prefer to explain what really matters: how the system reconstructs a play and why, even though it is extraordinarily accurate, it is important to understand its limitations.
What semi-automatic offside really is
The system does not replace the referee. It generates an automatic alert and leaves the final decision to the video team, who validate or adjust the moment of contact and the offside line. That is why it remains ‘semi-automatic’ rather than ‘automatic’. Its aim is not to eliminate human judgement, but to provide a much more reliable and faster basis for measurement than manually drawing lines.
How a play is reconstructed
The most fascinating part, and the one most closely related to our work, is the capture process. Around ten dedicated cameras are installed in the stadiums, recording at a rate of around 50 frames per second. Each player is modelled using 29 points distributed across the body, including the shoulders, elbows, hips, knees and feet, the areas that determine whether a player is offside. The ball provides its own data via a sensor that detects the exact moment of the pass.
Using all this data, the system constructs a three-dimensional avatar of each player and places the line on the furthest forward part of the body that counts for the rule.
The avatar is not a photograph
Here is the nuance that is almost never explained properly. The avatar we see in the replay does not correspond to any real image from any camera. It is a reconstruction that the system calculates by combining all available angles. There is no photograph of the exact moment; there is a model. And a model, by definition, is a very good representation of reality, not reality itself.
Understanding this completely changes the conversation. When someone asks for ‘the exact image’ of the play, they are asking for something that the system neither produces nor needs to produce in order to function.
Why it can go wrong: occlusion and point of view
If you work with motion capture, you know that the biggest enemy is occlusion: when one part of the body blocks another, or when two players cross paths, the cameras momentarily lose track of a particular point. In such situations, the system has to make an estimate, and that is precisely where the risk arises of confusing, for example, a defender’s foot with that of a striker.
Added to this is the issue of perspective. To determine with absolute certainty whether there was an offside, an orthogonal perspective is required, one that is perpendicular to the offside line and has sufficient resolution. Without that reference point, no side-view image can settle the matter one hundred per cent, however convincing it may appear on television.
None of these factors makes the system unreliable. They simply describe where its real limitations lie, which have to do with the physics of the footage capture and not with some supposed quirk of the software.
Accuracy versus perfection
The difference in reliability compared to the traditional method is enormous. Whereas the margin of human error used to be measured in percentages, the semi-automatic system reduces it to a minimal fraction, in the order of one in several thousand. The ball sensor transmits hundreds of data points per second, so that the moment of the pass is recorded with a margin of thousandths of a second which, in practice, does not alter the position.
Just because the margin is minuscule does not mean it is zero. And acknowledging that gap between ‘highly accurate’ and ‘perfect’ does not undermine the technology: it is, in fact, the rigorous way of talking about it.
What this has to do with our business
STT Systems does not operate LaLiga’s system, but we speak the same language. Analysing an athlete’s gait, studying a sporting movement or reconstructing a race in 3D is based on the same principles: multiple cameras, body reference points, precise temporal synchronisation and models that convert movement into measurable data.
That experience leads us to a clear conclusion. Motion capture technology has taken the precision of football to a level that seemed impossible a few years ago, and it will continue to improve. Previously, the focus was on referees; now it is on technology. The difference is that, for the first time, the debate can be based on data.