Home/Soccer Analytics/Defensive Tendency Bayesian Decomposition

Defensive Tendency Bayesian Decomposition

Tactical AnalysisLevel 3 — Advanced

What It Is

Using Bayesian hierarchical modeling to decompose a team's observed defensive behavior into three independent components: (1) baseline tactical tendency (what the team does on average against an average opponent), (2) home/away adjustment (how they modify their approach based on venue), and (3) opponent strength effect (how much the opposing team's quality forces a change). This separates a team's true tactical identity from the confounding effects of schedule and opponent quality. Without this decomposition, a team like Burnley looks like a pure low-block team — but after removing opponent strength effects, they're actually the 4th most aggressive pressing team because they push up against weaker opponents.

Correct Execution

(1) Use the GNN model to produce per-possession-state short possession probabilities for every match over a season. (2) Fit a Bayesian hierarchical model: Y = baseline_tendency + home_away_effect + opponent_strength_effect + noise. Use the in-possession team's xGD minus the out-of-possession team's xGD as the opponent strength covariate. (3) Extract posterior distributions for each component — the Bayesian approach gives uncertainty estimates, not just point estimates. (4) Interpret: baseline tendency shows the team's true defensive philosophy; home/away effect shows tactical venue adaptation; the opponent strength coefficient shows how much the meta-game (adapting to stronger/weaker opponents) drives what we see.

Key findings:

  • Opponent strength is the biggest driver: stronger teams in possession → longer possessions (they keep the ball better). This must be removed to see the true defensive tendency.
  • Burnley revelation: after removing opponent strength, they're the 4th most pressing team. Their low-block reputation comes from playing many matches against much stronger opponents.
  • Manchester City drops: their pressing intensity is partly explained by having a massive talent advantage over almost every opponent.
  • Home/away effects vary: Chelsea presses more at home; some teams (Man City) actually drop off slightly more at home.

Progression Levels

Diagnostic Tree

Coaching Cues

  • "They're not inconsistent — they're adaptive."
  • "Leeds press harder than anyone. They also concede more shots when beaten."
  • "After removing opponent strength, Burnley is the 4th most pressing team. Their schedule made them look defensive."
  • "Manchester City's pressing looks less impressive when you account for their talent advantage."

Common Errors

  1. Using raw pressing stats without opponent adjustment: A team that only plays strong opponents will look like they don't press. Adjust for opponent strength before drawing conclusions.
  2. Ignoring home/away effects: Some teams press significantly more at home. Aggregate stats hide this.
  3. Treating the decomposition as causal: The model separates correlation from confounding but doesn't prove that Burnley CHOOSES to press less against strong teams vs. being FORCED to. Both interpretations are valid.

Edges

Conventional Wisdom Is Wrong

Burnley Is Actually the 4th Most Pressing Team (Opponent-Strength Confound)

After Bayesian decomposition removes opponent quality and home/away effects, Burnley is the 4th most aggressive pressing team. Their low-block reputation comes from playing mostly against much stronger opponents. Man City's pressing intensity is partly an artifact of their talent advantage.

What most people do
Use raw pressing metrics (PPDA, pressures/90) to classify teams.
What the best do
Decompose into baseline + opponent strength + home/away using Bayesian hierarchical models.
Why it's an edge: Preparing for "low-block Burnley" when they press against similar-ranked opponents means your game plan is wrong 60% of the time.
How to exploit: Build opponent-strength-adjusted pressing profiles. Condition on YOUR team's xGD relative to theirs.
StatsBomb Conference 2021, 2021-11-04

Sources

  • StatsBomb, StatsBomb Conference 2021, YouTube, 2021-11-04 — presented Bayesian hierarchical decomposition of defensive tendencies; revealed Burnley as hidden pressing team; quantified press-shot-concession tradeoff; showed home/away tactical adaptation