Home/Soccer Analytics/Context Filtering Framework (Phase × Block × Dynamics)

Context Filtering Framework (Phase × Block × Dynamics)

Tactical AnalysisLevel 3 — Advanced

What It Is

The full context filtering framework combines three simultaneous dimensions: possession phase (build-up / progression / finishing), opponent block type (high press / mid block / deep block), and game dynamic (counter / fast attack / organized / direct). Any meaningful tactical analysis should be conditioned on a specific combination of these three dimensions. Without this filtering, analysis averages across situations that are fundamentally different problems, producing conclusions that apply nowhere specifically.

Correct Execution

For every tactical analysis question, first identify which context combination is relevant. Example: "How do we build out against a high press?" → filter to build-up phase + high press block + (counter or organized dynamic). "How do we penetrate a deep block in organized possession?" → progression phase + deep block + organized dynamic. Each combination may have small sample sizes — this is expected and correct. Small, well-defined samples are more analytically useful than large poorly-defined ones.

Progression Levels

Diagnostic Tree

Coaching Cues

  • "You're analyzing a possession as a whole without taking into account context — and that's why you can't explain what you see." — Javier Fernandez, 2019
  • "Drill it down to a more specific problem. Then you can describe it, analyze it, and prescribe for it."

Common Errors

  1. Applying context filtering retroactively as an excuse: Context filtering is a design choice made before analysis, not a post-hoc explanation for surprising results.
  2. Using too many context dimensions simultaneously: Three dimensions is usually sufficient; four or more creates combinatorial explosion with tiny per-cell samples.

Sources

  • Javier Fernandez, FC Barcelona, StatsBomb Innovation in Football Conference 2019, YouTube, 2019-10-22 — presented the multi-layer context filtering framework as the foundational methodology for the off-ball advantage model; argued this enables "fine grain analysis of what's going on and how to solve it for every possible opponent"