No single number on a match stats sheet tells you who played better. Possession shows who had the ball, shots show who tried, and expected goals (xG) shows how good the chances were, so you need all three read in order.
This guide explains each metric with a simple example and then gives you a short checklist for reading a stats table without being misled.
Ảnh: Keystone.ajh, Wikimedia Commons (CC BY-SA 4.0)
What does possession actually tell you?
Possession is the share of time each team holds the ball, calculated from tracked passing sequences. It describes control of the ball, not control of the game.
High possession does not guarantee a win. A side that defends tightly and counter-attacks will often finish below 40% possession and still create the better chances. A team with 65% of the ball but only 2 shots on target has had the ball without making it dangerous.
My view: treat possession as context, never as a verdict. It is useful for describing style, and nearly useless for judging who deserved the result unless you read it next to shot quality.
Shots versus shots on target: which one matters more?
A shot is any attempt to score, including blocked efforts and those that miss the frame. A shot on target is one that would have gone in without a goalkeeper save, or one that did go in.
Shots on target are usually the more useful figure, because at least the keeper was tested. A team with 20 shots but only 3 on target is either finishing poorly or being defended well. Location matters too: teams that shoot more often from inside the box, especially centrally, tend to score more over a season.
That is why five close-range shots can be worth more than ten from distance. It also explains a common pattern: when a team goes ahead early and drops deep, the opponent can pile up shots, but most of them are low quality.
What is xG and how is it worked out?
Expected goals measures chance quality as the probability that a shot becomes a goal. Each shot gets a value between 0 and 1, where 0 means it could not realistically score and 1 means a chance a player would almost always convert. A shot rated 0.01 has about a 1% chance, and one rated 0.99 is close to certain.
Models are built from a very large number of past shots. Typical inputs include shot location, body part, shot angle and the type of pass that set it up. Add up the xG of every shot a team takes and you get its xG for the match.
Take a shot from Mo Salah valued at 0.08 xG: roughly an 8% chance of scoring. A basic stats sheet counts it as one shot, the same as a tap-in from 2 yards in front of an open goal. xG separates the two, which is the whole point.
| Metric | What it measures | Common misreading |
|---|---|---|
| Possession | Share of time on the ball | “More ball means better team” |
| Shots | Number of attempts | Treating a 40-yard effort like a tap-in |
| Shots on target | Attempts that test the keeper | Ignoring where the shots came from |
| xG | Quality of chances before the shot | “Higher xG means deserved to win” |
Where can you find these numbers, and why do they differ?
Match reports, broadcasters and data providers all publish these figures, and they do not always agree. Each provider uses a slightly different xG model, so a match can show 1.4 xG on one site and a different value on another. When you compare teams or games, stick to one source.
For readers who also like to browse a sports section and football player information in one place, w88.inc is an external site that lists football, basketball and tennis sections alongside player information pages. It is a betting-related website intended for adults aged 18 and over, and this article does not cover wagering of any kind.
Some benchmarks are useful, though they vary by provider. Some providers treat a shot of about 0.38 xG or higher as a “big chance”, and a penalty is usually valued at roughly 0.75 to 0.80 depending on the model. A team’s total xG in a match is normally between 0 and 2 or 3, and anything above 5 is rare and points to a huge gap in quality.
What is the difference between xG and xGOT?
xG judges the chance before the shot is taken. Expected goals on target (xGOT) judges the shot after it is struck, and only counts shots on target, taking into account where the ball ends up in the goal. In short, xG is about chance creation and xGOT is about finishing and goalkeeping.
A simple case: Erling Haaland faces an open goal, which a provider such as Sportmonks might rate at 0.75 xG. If he drags the shot wide, his xGOT is zero because the shot was never on target. The chance was excellent, but the finish was not.
The same idea works for goalkeepers. A shot with 0.3 xGOT has a 30% chance of scoring and a 70% chance of being saved. In 2019/20, Dean Henderson of Sheffield United faced shots on target worth a combined 39.4 xGOT, which is a way of asking how many goals an average keeper would have conceded from them. That is a historical example, not a current ranking.
Two related terms are worth knowing. Expected assists (xA) rate the pass instead of the shot: Bruno Fernandes led the 2022-23 Premier League with 16.7 xA but had only 8 actual assists, which usually says more about teammates’ finishing than about his passing. Expected goals against (xGA) is the quality of chances a team allows.
How do you read a stats table, step by step?
Use five questions, in this order:
- Who scored first, and when? A team protecting a lead often sits deep, which distorts possession and shot counts.
- Does possession go with shots on target, or only with sterile passing?
- Where did the shots come from: inside the box or from distance?
- How big is the xG gap between the teams?
- Is the scoreline far from the xG? A big gap suggests luck, great finishing or a goalkeeper’s bad day.
Here is an invented match, made up purely to show how the metrics can tell different stories:
| Metric | Team A | Team B | Story it tells |
|---|---|---|---|
| Possession | 65% | 35% | A controlled the ball |
| Shots | 20 | 6 | A attempted far more |
| Shots on target | 3 | 4 | B tested the keeper more |
| xG | 1.4 | 0.8 | A had somewhat better chances overall |
| Goals | 1 | 3 | B finished far above expectation |
Read alone, possession and shots say A dominated. Shots on target make it closer, xG gives A only a modest edge, and the scoreline shows B scoring 3 goals from 0.8 xG. That is a team over-performing its chances, and the result may not reflect the balance of play. If you want to look up player information to see who might be responsible for such finishing, an external reference is W88 https://w88.inc.
When should you not trust xG?
xG is most reliable over a longer run, such as a full season, rather than in one match. Goals are rare, so a single game contains a lot of luck, and a high xG does not mean a team deserved to win. Chances also go missing from the data: a dangerous cross that nobody meets is not a shot, so it gets no xG.
The practical rule is to treat one match as a small sample and look for repeated patterns across several games from the same provider. If a team keeps beating its xG match after match, that is worth a closer look at its finishing quality.
The thing to follow next is how post-shot metrics such as xGOT sit beside xG in the data that providers publish, because that pairing separates chance creation from finishing better than any single number can.
