What is xG in football predictions

Quick Answer: What Is xG in Football Predictions?

Expected goals, or xG, is a probability metric that rates every shot from 0 to 1 based on how likely it is to become a goal. In World Cup 2026 betting, xG helps estimate true chance quality, fair odds, goal expectancy, and whether teams are overperforming or underperforming their results.

If France create chances worth 2.1 xG but only score once, the model says their attacking process was better than the scoreboard. That matters when you are checking prices at lunch, refreshing lineups on your phone at 4%, and trying to decide whether the market has overreacted to one unlucky 1-0 result.

For a wider betting framework, xG should sit alongside team news, tactical matchups, odds movement, and bankroll discipline. You can use it with our World Cup betting guides and compare model prices with live market numbers on World Cup odds.

Core Definition: What Does xG Actually Measure?

xG measures the quality of a scoring chance by assigning each shot a probability between 0 and 1. A team’s match xG is the sum of those shot probabilities, estimating how many goals an average finisher would score from the chances created.

A speculative shot from 30 yards might be worth 0.05 xG, meaning similar shots historically become goals about 5% of the time. A close-range chance from the centre of the box might be worth 0.40 xG. A penalty is usually valued around 0.76 to 0.79 xG, depending on the provider and historical sample used.

The key betting idea is simple: xG separates process from outcome. A team can win 2-0 from two low-quality shots and still have created little. Another team can lose 1-0 after generating 2.1 xG, hitting the post, and forcing a goalkeeper into three strong saves. The final score says one thing; the chance quality says another.

In prediction terms, xG answers: “Given the chances created, how many goals would an average team be expected to score?” If England produce 2.1 xG but score only 1 goal, that is finishing underperformance in that match, not necessarily poor attacking play. Over a tournament, those gaps can create value before the public narrative catches up.

How xG Is Calculated: Variables Inside the Model

xG is calculated from historical shot data, with shot location usually the biggest driver. The closer and more central the shot, the higher the probability; wider angles, longer distances, headers, and pressure generally reduce the xG value.

The first layer is geography. A shot from six yards in the centre of goal is far more dangerous than a low-percentage effort from outside the box. Distance to goal and angle to goal are the backbone of most xG models because they explain a large share of finishing probability.

Shot type also matters. Headers are usually harder to control than shots with the foot. Open-play chances, set-piece chances, direct free kicks, penalties, rebounds, and cutbacks are treated differently because they have different historical conversion rates. A cutback to Kylian Mbappé near the penalty spot is not the same event as a floated cross to a marked defender at the back post.

Modern models may include defensive pressure: how many defenders were nearby, whether the shooting lane was blocked, whether the attacker was off balance, and whether the chance came under immediate pressure. Some providers also use advanced build-up inputs such as cross versus cutback, counterattack speed, possession depth, and whether the chance followed a turnover in a dangerous area.

Post-shot xG is a second layer. Traditional xG asks how good the chance was before the shot outcome. Post-shot xG asks how dangerous the shot became once placement is known: top corner, low across the keeper, central, or straight at the goalkeeper.

Provider methodology varies. FootyStats describes its xG-style inputs through shot accuracy, shot frequency, attack dangerousness, possession amount, and possession depth. Sportmonks offers dedicated expected goals API endpoints for fixtures, making xG usable in automated World Cup 2026 prediction models. The numbers will not always match exactly, but the mechanism is the same: convert chance quality into goal probability.

Key xG-Derived Metrics: xGA, xGD, and xG vs Actual

The most useful xG betting metrics are xG for, xGA against, xG difference, and xG versus actual goals. Together, they describe attacking chance creation, defensive chance prevention, and whether results are likely to regress.

• xG for: Expected goals created by a team. This measures attacking chance quality rather than actual finishing.
• xGA against: Expected goals conceded based on chances allowed. This helps rate defensive process and goalkeeper workload.
• xGD: Expected goal difference, calculated as xG minus xGA. A positive xGD suggests a team is creating better chances than it allows.
• xG vs actual: Actual goals minus xG. A positive number means a team is scoring above expectation; a negative number means it is finishing below expectation.

For World Cup 2026 betting, per-90-minute normalization is essential. National teams play uneven schedules, with qualifiers, Nations League matches, continental tournaments, and friendlies all mixed together. Comparing Brazil’s total xG across 12 games with Spain’s total xG across 8 games is misleading; xG per 90 gives a cleaner baseline.

In practical terms, a side with modest results but a strong xGD per 90 can be underrated. A team winning tight matches while losing the xG battle may look strong in the pub TV glow, but its underlying numbers could be far less convincing.

xG Probability Table: Shot Types and Expected Values

Typical xG values show why not all shots are equal. A team taking five weak long shots may look active, but one clean cutback or close-range rebound can be worth more than all of them combined.

These values vary by provider and model. One data company may rate a one-on-one at 0.38 while another rates it at 0.43 because of angle, pressure, goalkeeper position, or historical training data. The broad ranges, though, are consistent enough to support betting analysis.

The cumulative effect is the part bettors often miss. A team creating five 0.20 xG chances has 1.00 total xG. That does not mean it “must” score exactly once; it means the chance package is worth one expected goal before finishing variance is applied.

Why xG Matters for World Cup 2026 Betting Markets

xG matters because betting markets often react to final scores faster than they react to underlying chance quality. For World Cup 2026, strong xGD with modest results can signal an underrated team, while hot finishing can flag a regression risk.

International tournaments are narrative machines. One 3-0 win can push a team into the spotlight; one goalless draw can make a favorite look stale. xG helps test whether that reaction is justified. If Argentina win 3-0 from 1.1 xG, the scoreline may flatter them. If Spain draw 1-1 from 2.4 xG while conceding only 0.5 xGA, their process may still be strong.

Finishing regression is one of the clearest betting uses. A team scoring far above xG may include elite finishers such as Lionel Messi, Harry Kane, or Vinícius Júnior, but extreme overperformance often cools. If a side has scored 10 goals from 5.8 xG in qualifying, the market may price recent goals too aggressively.

Goalkeeping regression works the other way. If a team has conceded 3 goals from 8.5 xGA, its goalkeeper may be in excellent form, but the defense is still allowing high-quality chances. Against stronger World Cup opposition, that can show up in over markets, both teams to score, or opposition team totals.

xG also feeds totals betting. If a model projects 1.55 xG for France and 1.05 xG for Mexico, the implied match total is 2.60 goals. If the bookmaker total line is priced as though the game expectation is closer to 2.25, there may be over value after margin is removed. Historically, World Cup tournament goal averages and aggregate xG have often clustered around roughly 2.6 to 2.7 goals per match, which supports using xG as a first-pass totals anchor.

Outright and futures markets use the same building blocks. Attack and defense xG ratings feed simulations that produce probabilities for group qualification, quarter-final appearance, finalist status, and lifting the trophy.

How xG Powers Tournament Simulation Models (Poisson and Monte Carlo)

xG powers tournament simulations by becoming the expected goals input, or lambda, for each team in a match. Poisson or Dixon-Coles models then convert those expected goals into scoreline, win, draw, and loss probabilities.

A standard World Cup 2026 model starts with historical xG. Each national team receives attacking and defensive strength ratings based on the quality of chances created and allowed. Those ratings are adjusted for opponent strength, because generating 2.0 xG against San Marino does not mean the same thing as generating 2.0 xG against France.

Recency also matters. Matches from six months ago usually receive more weight than matches from three years ago. That time decay is especially important in international football, where squads change between qualifying and tournament kick-off. A team with a new manager, injured centre-backs, or an emerging striker can look different by June 2026.

The 2026 format also creates host adjustments. USA, Mexico, and Canada may receive varying home or regional advantages depending on venue, travel, crowd composition, altitude, climate, and familiarity. A hot afternoon in Mexico City is not the same environment as an evening match in Vancouver.

Once each match has expected goals, a Poisson model estimates scoreline probabilities. If Brazil are projected at 1.80 xG and Japan at 0.85 xG, the model calculates the probability of 0-0, 1-0, 1-1, 2-0, 2-1, and so on. Dixon-Coles adds a correction for low-scoring football outcomes, especially 0-0, 1-0, 0-1, and 1-1.

Monte Carlo simulation then repeats the tournament thousands of times, often 50,000 or more. Each run simulates group matches, tiebreakers, knockout brackets, extra time or penalty assumptions, and trophy outcomes. The result is a probability table, not a single prediction.

These are illustrative model-style numbers, not live odds. Sportmonks and other data providers combine xG, squad quality, fixture data, and machine-learning layers for World Cup 2026 win probabilities, but the core mechanism remains probability conversion from expected goals.

Using xG in Match-by-Match World Cup 2026 Betting: Practical Steps

The practical use of xG is to compare your projected probabilities with bookmaker implied probabilities. If your xG-based model says over 2.5 goals is 54% likely, fair odds are 1.85; if the market offers 2.05, the bet may have value before margin and limits.

Start with pre-match xG projections. Estimate each team’s attacking expectation and defensive allowance, then adjust for opponent quality, location, rest days, injuries, and tactical incentives. A group-stage favorite already qualified may rotate; an underdog needing a win may attack earlier than usual.

For over/under betting, combine both team xG projections into a match total. A 1.45 versus 1.20 setup gives 2.65 expected goals. You can then compare that with bookmaker totals and prices. The edge is not “over because both teams attack”; the edge is whether the implied probability in the odds is lower than your probability estimate.

For match result markets, look at xG versus actual trends across qualifying, continental tournaments, and recent friendlies. If Portugal are creating 2.0 xG per 90 and allowing 0.7 xGA, their process supports short prices. If a smaller nation has won three straight while losing the xG battle, caution is needed.

For both teams to score, combine each side’s xG for with the opponent’s xGA. If one team projects at 1.60 and the other at 1.25, the probability that both score is materially higher than in a 1.10 versus 0.70 match. Poisson can convert those lambdas into BTTS probability by calculating the chance each team scores at least once.

Correct score betting is where Poisson becomes most visible. With xG-based lambdas, you can estimate 1-0, 1-1, 2-1, 2-0, and other scorelines, then compare fair odds with bookmaker prices. Correct score markets have large margins, so discipline matters.

In-play, live xG can challenge the scoreline. If Colombia trail 1-0 at half-time but lead live xG 1.4 to 0.3, the next-goal or draw markets may be more interesting than the scoreboard suggests. That is the moment of lineup-refresh anxiety, pub noise, and trying not to overbet because your phone battery is flashing red.