Scenario Note: The following analysis is based on a hypothetical case study involving fictional clubs and players. Names, match events, and statistics are constructed for illustrative purposes to explore the concept of substitution impact metrics in football analytics. No real-world data or outcomes are claimed.
Expected Goals from Substitutions and Impact Metrics: Bench Analysis
Let’s set the scene. It’s the 65th minute of a tight Premier League match. The home side, let’s call them Riverside FC, is trailing 1-0 against a compact 4-2-3-1 defense. The manager looks to the bench. He has a choice: bring on a direct winger to stretch the play, or introduce a creative midfielder to break the lines. Which decision maximizes the chance of scoring? This is where Expected Goals from Substitutions (xG from Subs) becomes a critical analytical lens.
Traditionally, substitution analysis was limited to "goals scored by substitutes" or "assists from the bench." These are outcome-based and prone to noise. A substitute might score a deflected winner in stoppage time, but that doesn’t tell you if the tactical change was smart. xG from Substitutions measures the quality of chances created or taken after a player enters the game, isolating the impact of the change from the run of play. It answers: Did the substitute improve the team’s shot generation relative to the player they replaced?
Let’s walk through a case. Riverside FC’s starting left winger, operating in a 4-3-3 shape, has an average xG per 90 of 0.25 and averages 1.2 shots per 90. By the 60th minute, he’s generated only 0.08 xG total. The manager brings on a fresh winger, known for high pressing intensity and direct runs. Over the next 30 minutes, the substitute generates 0.45 xG from three shots, including a header from a cross. The xG delta—the difference between the substitute’s per-minute output and the starter’s—is stark. This isn’t a fluke; it reflects tactical intent. The substitute exploited tired full-backs in a 3-5-2 system that left space in wide areas.
To operationalize this, analysts use several impact metrics:
| Metric | Description | What It Captures |
|---|---|---|
| xG per 90 (Sub) | Expected goals per 90 minutes when entering as a substitute | Shot quality in a fresh-legs context |
| xG Delta vs. Starter | Difference in xG per 90 between sub and the player replaced | Tactical upgrade or system mismatch |
| Shot Volume (Sub) | Shots per 90 after substitution | Aggression and involvement |
| PPDA Change | Change in passes per defensive action after sub | Pressing intensity shift |
The table above shows that substitution impact isn’t one-dimensional. A player might have a high xG per 90 but a low shot volume, indicating they take high-quality chances but aren’t getting involved enough. Conversely, a high shot volume with low xG per shot suggests wasteful finishing or low-percentage attempts.
Now, consider the tactical context. A team switching from a 4-3-3 to a 4-2-3-1 with a substitute can change the attacking focal point. In our case, Riverside FC’s substitution shifted their shape. The new winger’s ability to cut inside forced the opposition’s 3-5-2 to compress, opening space for the central midfielder. The xG from the substitute wasn’t just his own; it also increased the xG of teammates due to defensive disorganization. This is the "gravity" effect—a metric that’s harder to quantify but visible in post-shot expected goals (PSxG) and assist networks.
But there are caveats. xG from Substitutions is sensitive to sample size. A substitute who plays only 10 minutes per game might have inflated xG per 90 due to variance. For example, a striker who scores a 0.5 xG chance in a 15-minute cameo appears elite, but over a season, the average might regress. The metric works best when aggregated over multiple appearances, contextualized by the opponent’s defensive strength and the game state.
Let’s break down a hypothetical season for Riverside FC’s bench:
| Substitute | Total Sub Minutes | xG from Subs | xG per 90 (Sub) | Shots per 90 (Sub) | PPDA Change (Team) |
|---|---|---|---|---|---|
| Winger A | 450 | 2.1 | 0.42 | 3.8 | -1.2 (more press) |
| Midfielder B | 320 | 0.9 | 0.25 | 2.1 | +0.5 (less press) |
| Striker C | 200 | 1.5 | 0.68 | 4.5 | +0.8 (less press) |
Striker C’s numbers jump out: 0.68 xG per 90 from the bench is elite. But look at the PPDA change—the team pressed less when he came on. This suggests a tactical shift, not necessarily a superior player. The manager might have been chasing the game and sacrificed structure for directness. This is where impact metrics need to be read as a constellation, not in isolation.
In practice, clubs use these metrics to decide transfer targets, rotation policies, and in-game tactics. A player with a high xG per 90 (Sub) but low starting output might be a "super-sub" specialist—valuable for breaking down tired defenses. Conversely, a player whose xG drops significantly when coming off the bench might lack the mental readiness for high-pressure cameos.
For a deeper dive into how wide players contribute to chance creation, see our analysis on wide players crossing accuracy and expected assists. And for a look at how full-backs influence attacking output, check out full-backs overlapping runs and expected assists.
The bottom line: xG from Substitutions and impact metrics turn the bench from a black box into a quantifiable asset. They measure not just what happens, but why it happens—and whether the manager’s gamble paid off.