Expected Goals (xG): The Number That Changed Football Betting

Football used to be judged on the scoreboard and not much else. A team that scraped home 1-0 was "clinical"; a team that lost 3-1 while dominating was "wasteful". Then a statistic called Expected Goals, or xG, arrived and started telling a very different story — one that a lot of professional punters now take more seriously than the final result itself.

The video above, from analytics channel McKay Johns, walks through the actual mathematics behind xG: how a single shot gets turned into a probability, and why that probability is more useful than most people assume.

What the number actually measures

At its simplest, xG assigns every shot a value between 0 and 1 — the estimated chance that particular shot results in a goal, based on thousands of comparable shots in the past. A close-range tap-in might sit above 0.7; a speculative strike from outside the box might be worth 0.03. A standard penalty is generally rated around 0.76, reflecting how often penalties are historically converted.

As the video explains, these probabilities come out of a statistical technique called logistic regression, which is well suited to yes/no outcomes like "goal" or "no goal". The model is trained on historical data and weighs up a handful of key inputs: distance from goal, the angle of the shot, which body part was used, the type of build-up (through ball, cross, set piece, rebound), and how much defensive pressure the shooter was under. Add up a team's shots across a match, each weighted by its own probability, and you get a match xG total — a far more stable measure of the chances a team actually created than the scoreline alone.

Why it matters more than the scoreboard

This is where xG earns its keep for anyone who follows the sport analytically. A team can win 1-0 while being heavily outshot on quality chances, or lose 2-1 after generating three or four clear-cut opportunities. Over a single match, finishing and goalkeeping — plus a good dose of randomness — can easily overwhelm the underlying pattern of play. xG strips a lot of that noise out, comparing what a team actually did (goals) against what an average side would be expected to do with the same chances.

That gap between actual goals and expected goals is exactly why markets and modellers pay attention to it. A team consistently scoring well above its xG is often riding hot finishing that regresses eventually; a team scoring below it may be unlucky, or just wasteful, and due a correction either way. Betting markets, especially in-play ones, move on exactly this kind of information.

The takeaway for punters

None of this makes xG a crystal ball. It's a probability estimate built from historical patterns, not a guarantee of what happens next — and, as the video notes, different providers calculate it differently, so numbers aren't always directly comparable. For anyone using football markets, the value of xG is in reading form more accurately than the ladder or the last scoreline can, not in promising a result. Treat it as one more piece of context, weigh it against price, and never mistake a good model for a sure thing — the ball still has to go in.