Skip to main content
ExplainerOffside TechExplainer· 6 min read· in Sports

The Mechanics of Semi-Automated Offside Technology: How Skeletal Tracking Works in the Premier League

Modern soccer leagues rely on networks of high-speed cameras and artificial intelligence to track 29 skeletal points per player in real time. By generating millions of data points per match, semi-automated offside technology replaces manual crosshairs with 3D mesh models to determine infractions in seconds.

By Meera Iyer

Match Officials 40%Match-Going Supporters 30%Sports Scientists 30%
Match Officials
Prioritizes the accuracy of decisions and the reduction of high-pressure manual triangulation.
Match-Going Supporters
Values the preservation of the game's emotional flow and the reduction of agonizing stadium delays.
Sports Scientists
Views the technology primarily as a revolutionary tool for capturing high-frequency biomechanical performance data.

Perspectives this story doesn't cover

  • Broadcast Directors
  • Defensive Tacticians

At a glance

  1. SAOT replaces manual crosshairs with high-frequency optical tracking and artificial intelligence.
  2. The system tracks 29 specific skeletal points on every player, excluding the arms and hands.
  3. High-speed cameras capture data up to 100 times per second to pinpoint the exact kick-point.
  4. Human officials must still validate the automated alerts and interpret active interference.
  5. The technology generates 3D animated renderings to visually explain confirmed offside decisions to fans.

Why it matters now

The implementation of skeletal tracking eliminates the agonizing multi-minute delays that have plagued soccer since the introduction of video review. By automating the geometry of the offside rule, the technology restores the spontaneous joy of goal celebrations while generating a revolutionary new dataset for biomechanical analysis.

For years, the most frustrating aspect of the Video Assistant Referee (VAR) system in the Premier League was the manual drawing of offside lines. Fans in the stadium and viewers at home were forced to watch as officials painstakingly manipulated digital crosshairs on grainy broadcast footage to determine if a striker's toe was marginally ahead of a defender's shoulder. This process routinely caused delays of three to four minutes, draining the emotional momentum from the game. The introduction of Semi-Automated Offside Technology (SAOT) fundamentally changes this dynamic, replacing human estimation with high-frequency optical tracking and artificial intelligence.[1][3]

At the core of the SAOT system is a dedicated network of specialized tracking cameras installed beneath the roof of the stadium. Unlike standard broadcast cameras, which typically capture footage at 50 frames per second and are operated by human camera operators following the ball, the SAOT cameras are entirely automated. They are strategically positioned to provide overlapping, unobstructed views of the entire pitch, ensuring that every player is constantly monitored from multiple angles simultaneously, regardless of where the play is developing.[1][3]

These high-speed cameras feed raw optical data into a centralized artificial intelligence system that utilizes advanced computer vision algorithms. The software is trained to instantly recognize the human form and map exactly 29 distinct skeletal data points on every single player on the pitch. These points correspond to the specific anatomical extremities that are relevant for determining an offside position under the Laws of the Game, including the shoulders, knees, toes, and the top of the head. The arms and hands, which cannot legally be used to score a goal, are excluded from the offside calculation.[1][2]

The tracking system operates at an incredibly high frequency, capturing the exact three-dimensional coordinates of all 29 skeletal points up to 100 times per second. This generates a staggering volume of biomechanical data. At 50 frames per second, the system records 1,450 discrete skeletal data points per player every single second. Over the course of a standard ninety-minute match, this translates to nearly eight million skeletal coordinates for a single player, and hundreds of millions of data points across all twenty-two players on the pitch.[3]

The optical tracking system generates hundreds of millions of biomechanical data points over a 90-minute match.

This high frame rate is critical for solving one of the most difficult challenges in offside officiating: determining the exact "kick-point." In a manual VAR review, the official must select the single broadcast frame that best represents the moment the ball was played by the attacker's teammate. However, at standard broadcast frame rates, the ball can travel a significant distance between two consecutive frames, creating a temporal blind spot. By doubling or even quadrupling the frame rate, the SAOT system dramatically reduces this margin of error, pinpointing the exact millisecond of contact.[1][3]

However, at standard broadcast frame rates, the ball can travel a significant distance between two consecutive frames, creating a temporal blind spot.

In some iterations of the technology, the optical tracking is supplemented by a connected ball sensor. A small inertial measurement unit (IMU) is suspended in the center of the match ball, transmitting acceleration data at a rate of 500 times per second. When a player kicks the ball, the sensor detects the exact moment of impact and synchronizes this timestamp with the optical tracking data. This fusion of sensor data and skeletal tracking provides an irrefutable, mathematically precise determination of the kick-point.[1][3]

Once the kick-point is established, the artificial intelligence instantly evaluates the three-dimensional coordinates of the attacking player relative to the second-last defender. Because the system already knows the exact location of every relevant skeletal point, it does not need to draw two-dimensional lines on a screen. Instead, it mathematically calculates whether any valid part of the attacker's body is closer to the goal line than the defender's deepest valid body part. If an offside position is detected, the system immediately generates an automated alert in the VAR control room.[1][2]

It is important to note that the technology is "semi-automated," not fully automated. The system does not blow the whistle or stop the game. When the VAR receives an offside alert, the human official must still validate the system's findings. They must confirm that the correct kick-point was identified and, crucially, they must determine whether the player in the offside position was actively interfering with play or interfering with an opponent. The subjective interpretation of the offside law remains entirely in the hands of the human referee.[1][2]

While the system automatically detects offside positions, human officials must still interpret whether the player was actively interfering with play.

If the VAR confirms the offside offense, they relay the information to the on-field referee via the earpiece. The entire process—from the moment the ball is kicked to the final confirmation—typically takes only a few seconds, reducing the average delay by roughly thirty seconds compared to the manual crosshair method. This speed allows the game to flow more naturally, preserving the immediate joy of goal celebrations while maintaining the integrity of the offside rule.[1][3]

To improve transparency and fan understanding, the SAOT system also generates a 3D animated rendering of the offside incident. Once the decision is confirmed, the system uses the skeletal data to create a precise digital recreation of the players' positions at the exact moment of the kick. This animation clearly highlights the offside line and the specific body part that triggered the infraction. The rendering is then broadcast to television viewers and displayed on the stadium's giant screens, providing a definitive visual explanation of the ruling.[1][3]

The implementation of SAOT represents a massive logistical and financial investment for the Premier League. Upgrading the stadium infrastructure to support the specialized camera networks and the massive data processing requirements requires significant coordination between the league, the clubs, and the technology providers. However, the unanimous agreement among the clubs to adopt the system underscores a collective desire to eliminate the highly publicized controversies and frustrating delays that plagued the early years of VAR implementation.[3]

SAOT is designed to reduce the average delay for close offside calls by more than 30 seconds.

Looking ahead, the unprecedented volume of biomechanical data generated by the SAOT system has implications far beyond refereeing. While the primary purpose is officiating, the skeletal tracking data provides teams with a revolutionary new tool for performance analysis. Coaches and sports scientists can analyze a player's exact posture, balance, and mechanical efficiency during a sprint or a change of direction, variables that were entirely invisible to standard event data. The technology designed to rule out goals may ultimately become the key to understanding how to score them.[3]

Terms to know

Semi-Automated Offside Technology (SAOT)
A support tool for referees that uses optical tracking and artificial intelligence to quickly and accurately determine offside positions.
Skeletal Tracking
A computer vision technique that maps specific anatomical points on a player's body, such as the shoulders, knees, and toes, in three-dimensional space.
Kick-Point
The exact millisecond that the ball is played by an attacker's teammate, which serves as the temporal anchor for determining an offside position.
Inertial Measurement Unit (IMU)
A sensor suspended inside the match ball that detects acceleration and impact, transmitting data hundreds of times per second to synchronize with camera feeds.
Video Assistant Referee (VAR)
A match official who reviews decisions made by the on-field referee using video footage and communication headsets.

Questions readers ask

Does SAOT completely remove human referees from offside decisions?

No. The system automatically flags when a player is in an offside position, but a human Video Assistant Referee (VAR) must still determine if that player was actively interfering with play.

How does the system know exactly when the ball was kicked?

SAOT relies on high-speed cameras capturing up to 100 frames per second, and in some iterations, a connected sensor inside the match ball that transmits impact data 500 times per second.

Are a player's arms considered when drawing the offside line?

No. The artificial intelligence is programmed to track 29 specific skeletal points, deliberately excluding the arms and hands, as they cannot legally be used to score a goal under the Laws of the Game.

Will fans in the stadium see the SAOT offside lines?

Yes. Once a decision is confirmed, the system generates a 3D animated rendering of the offside incident, which is broadcast on television and displayed on the stadium's giant screens.

Sources

Source coverage

3 outlets

3 viewpoints surfaced

Match Officials 40%Match-Going Supporters 30%Sports Scientists 30%
  1. [1]WikipediaMatch Officials

    Video assistant referee

    Read on Wikipedia
  2. [2]WikipediaMatch Officials

    Offside (association football)

    Read on Wikipedia
  3. [3]Factlen Editorial TeamSports Scientists

    Synthesis by Factlen editorial team

    Read on Factlen Editorial Team

Comments

Stay informed

Every angle. Every day.

Get Sports stories with full source coverage and perspective breakdowns delivered to your inbox.