After a big match, supporters often compare player ratings across websites and find confusing differences. A striker who scored twice might receive a 9 on one site and a 7 on another, while a defender who made no obvious mistake can be rated anywhere from average to excellent. The differences are not errors: they reflect different rating systems, different data, and different ideas about what a good performance is.

A player rating looks like a measurement, but it is really a model's opinion. Most rating systems start from event data: passes, shots, tackles, interceptions, dribbles and duels, each recorded with a location and an outcome. The system then applies weights to those events, producing a score. Change the weights and the same performance produces a different score, which is why two sites can disagree honestly.

Some systems are heavily influenced by goals and assists, so an attacker who scores twice will rate highly even if his general play was quiet. Others reward volume: a midfielder who attempts many passes and wins many duels can rate well without a goal or assist. Neither approach is wrong, but they answer different questions about what mattered in the match.

Event data captures what happened, but not always how or why. A defender who positions himself well so that an attack never develops may record no tackle and no interception, and therefore earn little credit from a data-driven system. A forward who makes a clever run that drags defenders away and creates space for a teammate may receive no assist and no shot, yet change the game.

Context is also missing. A pass played under intense pressure in the final minutes of a close game is harder than the same pass played with a three-goal lead, but most systems treat them similarly. A goalkeeper who faces few shots because his defence was excellent may rate lower than one who makes several spectacular saves, even though the first keeper's team defended better as a unit.

Player ratings are often compared across teams, but a player's rating is shaped by his team's performance. A midfielder on a losing team may complete fewer passes and win fewer duels simply because his team had less of the ball, not because he played badly. Conversely, a player on a dominant team benefits from more touches, more chances and more opportunities to accumulate positive events.

This is why ratings after a heavy defeat can look harsh on individuals who were not at fault, and why a substitute who plays twenty minutes against tired opponents can rate higher than a starter who faced fresh defenders for an hour. Minutes played, game state and team dominance all feed into the numbers, whether the reader realises it or not.

Use ratings as a starting point, not a verdict. Compare a player's rating with his own recent ratings rather than with other players in different roles, and check the underlying stats: shots, key passes, duels won and defensive actions. If a rating surprises you, the stats will usually explain it, whether it is a high rating driven by a goal or a low rating driven by lost duels and misplaced passes.

Finally, trust your eyes alongside the numbers. Ratings are useful for spotting performances you missed and for comparing players across many matches, but they cannot replace watching the game. The best use of a rating is to prompt a question — why did this player rate so highly, or so poorly? — and then to answer it from the match itself.

The healthiest way to read ratings is to use them comparatively over time. A defender whose rating rises steadily across a season is probably improving, even if his absolute number looks modest next to an attacker's. A midfielder whose rating falls after a tactical change may be playing a less glamorous role, not playing worse. Trends, roles and underlying numbers together give a fairer picture than any single rating from any single match.