Slatery field notes / SOCCER / guided learning
Expected goals vs. goals: read the chances behind the score
Three attempts. Three different chance values.
- Distance attempt.08 xG → goal
- Angled attempt.12 xG → miss
- Close-range attempt.40 xG → saved
The shot with the smallest assigned chance scored. That can happen.
The score and the chances tell different parts of the match.
A team wins 1–0. The score tells you who won, but it does not show the route to that result. The winner might have created repeated close-range chances, or scored from distance and then survived pressure. Both stories fit the same scoreboard.
Expected goals, usually written xG, helps describe the attempts. A model gives each shot an estimated scoring chance from its recorded circumstances, then the match total adds those values. It is a summary of the shots, not a replacement result after the final whistle.
You can use both views together. The goal still counts, even if its shot had a low xG value. The chance record helps you ask what happened around that goal and whether the attack repeatedly reached promising positions.
Add the shot values. Keep the actual goal separate.
Our invented shots have values of .08, .12 and .40. Add them and the team finishes with .60 xG. In the picture, the .08 shot scores while the other two do not, so the actual result is one goal.
That is not a contradiction. A low-valued attempt can score; a high-valued one can miss. The model describes the chance assigned to the attempt, while the observed result records what happened on this occasion.
Follow three attempts into one xG total.
Actual goals: 1.
Equal totals can hide different collections of shots.
A total of 1.50 xG could come from three .50 chances or fifteen .10 chances. The sum matches, but the attack created different kinds of records. One total compresses both volume and the values of individual attempts.
Look at that distribution when the source supplies it. Then match the provider, dates and shot coverage before comparing teams. Different xG models can assign different values to the same attempt, so a mixed-source comparison can add a measurement difference you did not intend.
The same total can come from very different shot volume.
3 attempts × .50 each.
15 attempts × .10 each.
The total is a summary. The attempts explain what was added.
Use a gap to choose the next question.
When goals run above or below xG, inspect the actual attempts, sample size and current attacking role. A short gap can be a useful starting point without promising that results must reverse in the next match.
Keep penalties, red cards and unusual game situations visible if they materially shaped the window. A team protecting an early lead may create a different second-half record from a team chasing the game. A larger sample adds evidence, but can include older tactics and lineups.
A clear note is specific: “The team created these attempts in this window, and scored once despite .60 xG.” That preserves the score and the chance evidence while leaving tomorrow’s volume and outcome open.