Set-pieces decide an enormous share of tight matches, yet for years they were judged by little more than goals scored and gut feeling. Set-piece expected goals, or set-piece xG, changes that by putting a number on the quality of every chance created from a dead ball. A corner that produces a free header from six yards might be worth 0.35 xG, while a hopeful delivery into a crowded box might be worth 0.02, and over a season those numbers reveal which teams genuinely threaten from set-pieces and which ones merely take a lot of them. For analysts, coaches, and even fantasy players, set-piece xG is now one of the most useful lenses on a part of the game that was long treated as pure chaos.

Expected goals models assign every shot a probability of becoming a goal, based on factors such as distance, angle, body part, defensive pressure, and the type of assist. Set-piece xG simply filters that model to shots that follow a corner, free-kick, throw-in, or penalty, with most analysis separating penalties out because they are worth around 0.79 xG each and would swamp everything else. The interesting part is what the model captures about delivery quality: a corner whipped to the near post with a runner attacking it produces higher-value shots than a floated ball to the penalty spot, because the resulting headers are closer and less contested. Models also credit the chance creation, not just the shot, so a team that consistently wins first contact but sees headers fly over still shows up as dangerous. That distinction matters, because finishing from set-pieces is volatile while chance creation is a repeatable skill.

Across a top-flight season, the best set-piece sides typically generate between 12 and 18 non-penalty set-piece xG, which translates to roughly a third of a goal per match from dead balls alone. The worst sides manage half that figure, and the gap is worth several league positions over 38 games. Per-match benchmarks help too: a team averaging 0.5 set-piece xG per game is genuinely threatening, while anything below 0.2 suggests the routines are not creating clean chances. It is also worth splitting corners from free-kicks, because some teams dominate one and waste the other, and splitting attacking from defensive set-piece xG, since a side can be dangerous going forward yet vulnerable at the other end. The most revealing number is often xG per set-piece, which shows efficiency rather than volume: a team taking fewer corners but creating better chances is usually the better-coached side.

Coaches use set-piece xG to answer questions that video alone cannot settle, such as whether a new corner routine is actually working or just looked good once. If xG per corner rises after a routine change, the change stays; if it falls, the routine is scrapped regardless of how clever it seemed on the training ground. The data also shapes defensive work: a side conceding high set-piece xG against is usually losing first contact or failing to mark the zones where the best headers come from, and the fix is structural rather than motivational. Recruitment plays a part too, since clubs now value aerial dominance and delivery quality as measurable assets, signing specialists whose set-piece numbers travel with them. Perhaps most importantly, set-piece xG stops coaches from overreacting to a single goal, because one lucky header does not prove a routine works and one unlucky miss does not prove it fails.

Like all models, set-piece xG has blind spots that fans should keep in mind. It measures shot quality, not the quality of the delivery that did not result in a shot, so a brilliant corner that is cleared off the line by a defender's toe may register as nothing at all. It also struggles with second-phase chances, the scrambles and cutbacks that follow a half-cleared corner, which are often recorded as open-play shots even though the set-piece created them. Context matters as well: a team chasing a game takes more attacking corners, inflating its numbers, while a team protecting a lead may deliberately waste them. Used with those caveats, set-piece xG is a powerful tool, but it works best alongside video and coaching judgement, not as a replacement for them. The teams that gain the most from it are those that treat it as a question to investigate, not an answer to accept blindly.