Separating Open-Play and Set-Piece xG: the numbers

Set pieces produce a distinctive kind of chance and a distinctive kind of xG. Treating the two phases as one total hides real differences between teams, and separating them is one of the highest-value splits in the whole metric family.

The two phases are different sports

Open play is continuous and contested: the attack must move the ball past a defence that is set and shifting. Set pieces are discrete and rehearsed: the ball arrives from a dead-ball delivery into a box where both sides have had time to arrange. The events differ in their geometry, the number of players involved and the degree of rehearsal, and a single blended total averages over all of it.

The consequence is that a side can look identical to another on total expected goals while attacking in two entirely different ways. One creates through sustained possession and cut-backs, the other through corners and long throws. Blending them describes neither, and the reader has no way to tell which side of the average the team sits on.

How the split is decided

Providers classify each shot by the phase of play that produced it. The main categories are open play, corners, free-kicks, penalties, throw-ins and, in some models, second phases: the scrambled possession that follows a cleared set piece. The second-phase category is the most contested, because where a cleared corner stops being a corner is a judgement rather than a fact.

Those definitional choices propagate into the totals. A provider that counts a headed chance after a partially cleared corner as open play will report a higher open-play figure and a lower set-piece figure than one that keeps it in the corner bucket. Comparing two sides measured by two providers therefore requires checking the phase rule first.

Why it matters for judging a coach

Teams that rely on set pieces are usually evaluated harshly for a low open-play figure, and that judgement is partly unfair and partly accurate. It is accurate in the sense that a set-piece-heavy attack depends on winning set pieces, which requires sustained pressure and often correlates with a more defensive posture. It is unfair in the sense that a well-coached set-piece routine is a genuine competitive advantage that the open-play column simply does not see.

The honest reading uses both columns plus the count of set pieces won. A side with a modest open-play figure but a high set-piece figure per corner taken is not lucky; it has a rehearsed routine. A side with a high set-piece figure from a great many deliveries is living on volume, which is a fragile position.

The volatility difference

Set-piece expected goals is a much more volatile statistic than open-play expected goals, match to match and season to season. The number of corners won depends on deflections, blocks and clearance decisions that carry almost no information about the attacking side, and a routine that scores from two corners in one afternoon may go ten matches without another.

This volatility is the strongest argument for always reporting the split. A club whose total is inflated by a spike in set-piece conversions looks like a strong attacking side until the conversions stop, and the open-play figure would have described its stable level all along.

Penalties belong in neither column

A penalty is a set piece in form but not in spirit. It is not produced by rehearsal in the way a corner routine is, and its conversion rate barely varies between teams. Folding penalties into a set-piece figure inflates it and makes it incomparable across sides, which is one more reason to report penalties as their own line.

The clean presentation is therefore at least four columns: open play, set piece, penalties and the count of set-piece deliveries. Four numbers describe a team's attacking shape; one number hides it, and the hiding is what produces the confident, contradictory conclusions that analytics is often criticised for.

Practical split reporting

For a season review, the informative comparison is each side's open-play expected goals against its set-piece expected goals, with the set-piece figure expressed per delivery as well as in total. A high total with a low value per delivery is a volume story; a moderate total with a high value per delivery is a coaching story.

The split also explains why two clubs can be described in contradictory terms by different observers and both be right. One is watching the set-piece routine and the other the open-play build-up, and neither has the full picture until the phases are separated. The table is what settles the argument.

Key reference points

• Open play and set pieces differ in geometry, rehearsal and player involvement.
• Second-phase attribution is the least standardised part of the classification.
• A set-piece figure per delivery separates a rehearsed routine from pure volume.
• Set-piece expected goals is far more volatile than the open-play figure.
• Penalties should be their own line, not folded into the set-piece bucket.
• At least four columns are needed to describe a team's attacking shape.

Splitting the phases does not add information about the shots; it prevents two different kinds of attack from cancelling each other out inside one average.