xG (Expected Goals)

What is xG (Expected Goals)?

Expected Goals, or xG, is a statistical metric that measures the quality of goal-scoring chances. Each shot gets a value between 0 and 1 based on its probability of becoming a goal. A penalty, for example, usually has an xG of 0.76, while a 30-meter shot from a tight angle might be worth just 0.03.

The idea is simple: not all shots are equal. A team can take 15 shots and not score because they were all long-range and not really threatening. Another team can take 4 shots and score 3 because all were clear chances inside the box. xG captures that difference that shot counts don’t show.

For the bettor, this metric is gold. It lets you see beyond the result and understand which team is creating real danger. If a team accumulates high xG match after match but scores little, those goals will eventually come. And that’s where you find value.

How does it work?

The xG model analyzes thousands of historical shots and assigns a probability to each one based on factors like distance to goal, angle, whether it was headed or kicked, whether it came from a cross or individual play, and the goalkeeper’s position. Platforms like Understat, FBref, or StatsBomb calculate xG slightly differently, but they all follow the same logic.

Take a real example. In the 2024/25 Premier League season, Brentford generated an average of 1.8 xG per home match but only converted 1.2 goals on average. That means they were missing clear chances systematically. For a bettor, that’s a signal the market may be undervaluing Brentford at home, because the results don’t reflect the quality of their chances.

There’s also xGA (Expected Goals Against), which measures the quality of chances a team concedes. If a team has low xGA, they allow few real-danger opportunities, even if opponents shoot at them a lot from distance.

When to use xG?

xG is especially useful in three situations. First, when you analyze goal trends: if a team has consistently high xG but few goals, that gap will close. Second, to evaluate Over/Under markets, because a match between two teams with high xG has a higher real probability of clearing the 2.5-goal line. Third, to spot teams whose results don’t reflect their actual performance.

A common error is looking at xG from a single match. One match is statistical noise. The power of xG appears when you analyze samples of at least 8-10 matches. There the trends become reliable.

It also pays to cross xG with tactical context. A team that plays low block and counters can have low xG per match but be lethal in transitions. The number alone doesn’t tell you the whole story.

Practical example

Suppose Manchester City hosts Wolverhampton in the Premier League. You check the data and see City has averaged 2.4 xG over their last 10 home matches, but has only scored 1.6 goals on average. Wolves, meanwhile, has an xGA of 1.9 on the road, indicating they concede clear chances.

The book offers Over 2.5 goals at 1.72. You know City’s xG at home suggests they should be scoring more, and Wolves concedes quality chances. There’s value there. The Over isn’t a sure thing, but the implied probability of the odds (58%) is below what xG data suggests (closer to 65%).

If City has had recent matches where they burned clear chances, regression to the mean likely favors them soon.

Common mistakes

1. Using xG from a single match to draw conclusions. A team can generate 3.5 xG one day and 0.8 the next. You need broad samples, minimum 8 matches, for the data to be reliable.

2. Ignoring who’s shooting. Standard xG doesn’t account for shooter quality. An xG of 0.4 with Haaland executing is different from the same with a converted full-back. Some advanced models (xGOT, Post-Shot xG) do consider execution.

3. Thinking high xG always means goals. xG tells you what should happen on average, not what will happen in a specific match. It’s a probabilistic tool, not a crystal ball.

4. Not comparing with the sportsbook’s line. xG by itself doesn’t give you value. Value appears when you cross xG with the odds the market offers and find discrepancies.

Frequently Asked Questions

Where can I check xG data for free?

The best free sources are Understat (covers Europe’s top 5 leagues and the RPL), FBref (StatsBomb data for many competitions), and Fotmob. Each has its own model, so the ideal is to stick with one source and stay consistent.

Does xG work for smaller leagues?

It depends on data coverage. For leagues like MLS, the Eredivisie, or the Portuguese Liga there’s reasonable data. For South American leagues or second divisions, coverage is more limited and models less precise. There it pays to supplement with visual match analysis.

Can I base my bets only on xG?

It’s not recommended. xG is a piece of the puzzle, not the whole puzzle. You should combine it with tactical context, injuries, motivation, match conditions, and odds analysis. A bettor who only looks at xG will miss important nuances that affect the result.