If you follow football analytics at all, you have seen xG quoted after almost every big match, often as proof that a team was lucky or unlucky. Expected goals is a simple idea underneath the jargon: every shot is given a value between zero and one, based on how often similar shots have become goals in the past. Add up a team's shots and you get its xG for the match, which is roughly how many goals an average finisher would have scored from those chances.

The value of a shot depends mainly on distance and angle, but good models also weigh the body part used, the type of assist, and the defensive pressure around the shooter. A penalty is worth roughly 0.76 xG because penalties are scored about three times in four. A shot from 30 metres at a tight angle might be worth less than 0.05, because almost no comparable shot goes in. A central shot from six yards out, by contrast, can be worth around 0.40 or more.

The most common mistake is treating one match of xG as a verdict. Football is a low-scoring sport, so a team can create the better chances and still lose, and a team can win while creating very little. Over a single game, luck and finishing quality dominate. Over ten or more games, xG becomes far more reliable, which is why analysts trust season-long xG difference, the gap between xG created and xG conceded, far more than any single result.

Two companion numbers help you read xG properly. Non-penalty xG strips penalties out, giving a cleaner picture of open-play chance creation, since penalties are rare and worth a lot. Expected goals against, or xGA, is the defensive mirror: it shows the quality of chances a team allows. A team with high xG and low xGA is usually strong even if recent results have been mixed, while a team winning despite poor xG numbers is often riding a hot streak that fades.

It also helps to know what xG cannot do. Most public models treat every player as an average finisher, so they underrate elite strikers and goalkeepers. They only count shots, so a dangerous attack that ends without a shot adds nothing. Different data providers use different models, so the same match can show slightly different xG figures on different websites. Always compare teams using one consistent source rather than mixing providers.

Used well, xG answers a better question than the scoreline alone: did this team create good chances, and did it give good chances away? Check the xG over the last five to ten games, look at the difference between attack and defence, and then watch the matches to understand why the numbers look that way. That combination of numbers and eyesight is what separates genuine analysis from quoting a stat to win an argument.

A worked example helps the idea stick. Imagine a team takes ten shots in a match: several from distance worth 0.03 to 0.05 each, two from inside the box worth 0.15 each, and one penalty worth 0.76. Add them up and the team's xG is roughly 1.25, even though the actual result could be anything from a goalless draw to a three-goal win. The total describes the chances, not a promise, and that distinction is the whole point of the stat.

If you want to go further, start tracking xG difference over a rolling five-match window for your own club, and note whether results are running ahead of or behind the chances. When a team keeps winning with poor xG, enjoy the run but do not assume it will last. When a team keeps losing with strong xG, the performances may be better than the table suggests, and a change of fortune is often closer than the headlines imply.