The scoreline lies in one match in four: 460 games analysed

We analysed 460 matches across twelve competitions for which complete expected-goals data exists. 349 of them had a winner. And in 97 of those — more than one in four — the side that created less came out on top.

That is not a curiosity or a scandal. It is the normal rate, and it rarely gets said out loud when a single match is being argued over.

What was measured

Expected goals (xG) price the quality of every chance created: where the shot came from, at what angle, out of what situation. The match total shows how many goals a side's play was worth. We compared that with the final scoreline.

A match entered the count if both sides had a computed xG and the game did not end level. There are 111 draws in the sample — almost a quarter of all 460 games — and they sit outside the comparison: there is no winner whose result could be disputed.

More than one match in four is won by the side that created less

97 of 349 is 27.8%. For context, in the same sample home wins came in 43.0% of matches, draws in 24.1%, away wins in 32.8%.

Another number from the same family: in 34.7% of decisive matches the winner took fewer shots than the loser. More than a third.

Average home xG is 1.48, away 1.21. Average home possession is 51.6%. Home advantage exists in the data, but it is modest: roughly a quarter of a goal and a point and a half of the ball.

Where the scoreline and the play part company most

One caveat, without which the table should not be read. Europe's top leagues started in mid-August, so the Bundesliga has seven decisive matches here, Ligue 1 eleven, the Premier League sixteen. At those counts a five or ten point difference means nothing: one result either way moves the share by several points. Those rows are in the table, but we draw no conclusions from them.

The one comparison the sample can carry

Two competitions have accumulated enough matches to be set against each other honestly: the World Cup with 81 decisive games, MLS with 55.

Between them the gap is twofold: 14.8% against 30.9%.

The explanation suggests itself, and the data does not contradict it. At a World Cup the gap in class between sides is wide: the favourite creates more and almost always sees it through. MLS is built to level squads deliberately — salary cap, draft, allocation rules. When sides are near-equal, the outcome is settled by a single episode rather than by superiority over ninety minutes.

So an "unfair" result is not a property of luck. It is a property of a competition without a clear favourite.

Half of all matches are settled after the eightieth minute

A separate figure explains a good deal of why scorelines diverge from the play so often. Of the 395 matches where we know the goal times, 205 had their final goal scored from the eightieth minute onwards. That is 51.9% — more than half.

In 108 matches a goal arrived in the ninetieth minute or in stoppage time.

A side can hold the ball, shoot more often and create better chances for eighty minutes, then concede once, and all of that statistical work stays in the report while the points travel with the opponent.

What a reader should do with this

The plain conclusion: one match is not enough to judge a team. The scoreline is the result, but it is not a description of the game, and in 27.8% of cases it tells a story that did not happen.

None of this makes xG "more correct" than the scoreline. Points are awarded for goals, and in that sense the scoreline is always right. But if the question is not "who took three points" and instead "who played better and what should be expected next", the scoreboard alone falls short roughly a quarter of the time.

These figures will be updated as new matches enter the base. It currently holds 516 analysed games, 460 of them with a complete set of metrics — and every number above is built on those.