Expected goals, usually written as xG, estimates the chance that a shot will become a goal. It gives a way to discuss the quality of opportunities rather than counting every attempt as equally dangerous. A close-range shot into an exposed goal and a pressured effort from far away should not tell the same story about an attack.
xG is a model, not an alternative official score. The team that scores more actual goals wins. Its value is in helping us ask better questions about how those goals, misses and saves came about.
What a shot value means
A shot rated 0.20 xG has an estimated scoring probability of 20 percent in that model. Across a large collection of comparable opportunities, roughly one in five would be expected to become a goal if the model is well calibrated. It does not mean that a particular player is owed a goal after five attempts.
The model learns relationships from recorded shots and their outcomes. Distance and shooting angle are important features. Depending on the provider, body part, type of assist, defensive pressure, goalkeeper position and the positions of other players can also be included.
Those inputs explain why two attempts from the same spot might receive different ratings. One could be an uncontested first-time finish; the other a header under pressure with the goalkeeper well positioned. Location alone cannot capture every difference in the opportunity.
Adding shots to create a match total
Imagine an invented team taking three shots with values of 0.10, 0.25 and 0.40. Adding them gives 0.75 xG. The team might score none, one, two or even all three. The fractional total expresses an expectation across the opportunities, not a possible final score.
Another team might take ten shots worth 0.05 each, producing 0.50 xG. It has more attempts but a lower combined estimate. This is why shot counts alone can exaggerate the danger of repeated speculative shooting.
Be careful when converting summed xG into the probability of scoring at least once. Two hypothetical independent chances of 0.50 each add to 1.00 xG, but the probability of at least one goal would be 75 percent, not 100 percent. Summation measures expected goals, not certainty of a goal.
Why providers disagree
There is no universal xG formula shared by every company. Providers differ in their training data, event definitions, available tracking information and modelling choices. One may know the goalkeeper's position for a shot while another has less detailed information.
A small disagreement between two published match totals does not automatically show that one is broken. It may reflect different models. For comparisons across several matches, use the same provider where possible and name it. Mixing one team's total from one source with another team's total from a different source can create a misleading comparison.
Reading a performance without rewriting the result
If a team loses despite creating the stronger chances, xG can help explain why its performance may have been more competitive than the score suggests. It cannot prove the team deserved a different result. Football rewards finishing and preventing goals as well as creating opportunities.
Game state also changes the evidence. A side protecting a lead may concede possession and allow low-quality attempts. Its opponent can accumulate shots while chasing the match. Ask when opportunities occurred and what both teams were trying to do at that moment.
A penalty can dominate a modest match total. If the question is about open-play creation, look for a non-penalty breakdown rather than treating the full total as one uniform stream of attacking pressure. Non-penalty xG still needs context, but it answers a more specific question.
Important limitations
Ordinary shot-based xG misses dangerous attacks that never produce a shot. A cutback just behind an unmarked forward can be a threatening moment without adding anything to the total. Similarly, one rebound sequence may contain several attempts that could not all have occurred if the first had gone in.
Pre-shot xG is also different from post-shot measures, which can incorporate where the shot was placed after it was struck. Do not use a pre-shot chance rating as a complete verdict on a goalkeeper's save difficulty or a striker's execution.
Small samples fluctuate. Scoring well above xG for one evening may reflect excellent finishing, weak goalkeeping or simple variation. A longer pattern deserves investigation, but even then the model's limitations and the player's role still matter.
How to use xG in an analysis
Start with the actual result on Matches. If you have a sourced xG report, identify the provider and the period covered, then describe the biggest chances and the game state. TV96 Live's fixture and table displays should not be mistaken for an xG feed.
Use xG alongside goal difference, match observation and a clear account of the opportunities. It is most useful as evidence for a question about chance creation, rather than a single number expected to settle every argument about a performance.