Few statistics have changed how fans talk about football as much as expected goals, usually shortened to xG. Broadcasts show it, analysts quote it and many prediction pages mention it. Yet it is often misunderstood. This article explains what xG measures, how to read it and where it stops being useful.
What Expected Goals Measures
Expected goals gives every shot a value between 0 and 1. That value is the estimated chance that a shot of that kind ends in a goal. A penalty is usually valued at around 0.76, while a long-range effort from a crowded area might be valued at only 0.02. Add up the values of all shots a team takes in a match and you get its xG for that match.
The value comes from a model that has looked at a very large number of past shots. It considers factors such as the distance from goal, the angle, the part of the body used, whether the shot came from open play or a set piece, and how the move started. Different providers use slightly different models, so the same match can have a slightly different xG on different websites.
Why It Is More Useful Than the Final Score
A single match result contains a lot of luck. A deflection, a goalkeeping error or a missed open goal can decide a game that was otherwise even. xG tries to separate the quality of the chances from the finishing and the luck around them.
Imagine a team that loses 0-1 but created chances worth 2.1 xG, while its opponent created chances worth 0.4 xG. The score says the opponent was better. The chance quality says the losing team played well and was unlucky. Over a longer run, teams usually move toward what their chance quality suggests, which is why many analysts look at xG over several matches rather than one.
How to Read the Numbers
- xG for and xG against. The first shows the quality of chances a team creates; the second shows the quality of chances it allows. The difference between them is often a better description of a team than its goal difference.
- Goals compared with xG. If a team has scored far more than its xG over many matches, it has either finished very well or been lucky. If it has scored far fewer, the opposite may apply.
- Per-match average. A single match tells very little. A rolling average across ten or more matches is more stable and easier to compare.
What xG Does Not Tell You
Expected goals is a description of chances, not a forecast on its own. It does not know about a striker who is a clearly better finisher than average, or a goalkeeper in exceptional form. It does not see a red card in the 20th minute that changed how the match was played, because a team that is ahead may simply stop attacking and allow the other side to build up a high xG late on.
It also treats all shots of the same type equally, whatever the match situation. A shot in the final minute of a match that is already decided has the same value as one in the first minute. For this reason xG works best as one tool among several, together with team news, tactics and the context of the match.
Using xG Sensibly
A sensible approach is to look at xG over a meaningful number of matches, compare it with the actual results, and ask why the two differ. A gap can point to luck that is likely to even out, or to a real strength or weakness that the model does not capture. Neither is a promise about the next match.
Football remains unpredictable, and even the best models are wrong often. Expected goals helps to describe what has happened and to ask better questions, but it cannot say what will happen.
A Short Example
Imagine a match that ends 1-1. One team had chances worth 2.4 xG, mostly from close range, while the other had chances worth 0.6 xG. The score suggests an even match, but the numbers suggest that one side created far better opportunities. That does not prove which team will do better next time, yet it is a reasonable reason to look more closely at how both teams play.
Seen over a whole season, xG often explains why a team sits where it does in the table, and why a team that looks unlucky can improve without changing anything it does.
Predictions are opinions based on analysis, and no result is guaranteed. This site is for adults aged 18+ only.