Home/Soccer Analytics/Expected Ball Speed Model

Expected Ball Speed Model

Passing MetricsLevel 3 — Advanced

What It Is

A neural network that predicts, for any passing moment, the expected ball speed given the spatial context — where attackers, defenders, and the ball are positioned. The architecture is adapted from Fernandez & Bornn's soccer_map: a CNN that takes spatial input surfaces and produces a prediction surface for the entire pitch, but evaluates loss only at the actual pass destination. The key modification: replace the sigmoid activation (pass success/failure) with a linear activation to predict a continuous output (ball speed in m/s). The model captures that passes into crowded areas require higher speed, passes to open teammates can be slower, and certain pitch zones have characteristic speed profiles.

Correct Execution

Architecture: 4 spatial input channels, each a continuous surface over the pitch (1 cell per square meter):

  1. Min distance from each cell to nearest attacking player
  2. Min distance from each cell to nearest defending player
  3. Distance from each cell to the ball
  4. Difference between channels 1 and 2 (crude pitch control proxy)

Train with linear activation on the output layer; loss = absolute difference between actual and predicted ball speed. ~1200 trainable parameters (much smaller than soccer_map's ~400K). Train on ~50 games of 360 data.

Alternative approach (physics-based): iterate Spearman's pitch control across a range of ball speeds (8-20 m/s); compute pitch control along the pass trajectory at each speed; the minimum speed where pitch control stays >50% for the entire trajectory is the "minimum viable ball speed." Strengths: physics-based, accounts for player velocity. Weaknesses: slow, fails with missing off-screen players in 360 data.

Progression Levels

Diagnostic Tree

Coaching Cues

  • "The surface is the product, not the prediction."
  • "A fast pass into an open player isn't tempo. A fast pass through a crowd is."
  • "Average pass speed measures nothing useful. Context-adjusted speed measures tempo."

Common Errors

  1. Assuming zero player velocity with 360 data: 360 freeze frames don't include velocities. This introduces error, especially for passes into areas where players are running. Possible fix: estimate velocity from consecutive freeze frames using minimum spanning algorithms.
  2. Using the physics-based approach with incomplete 360 data: If the broadcast view is narrow, off-screen players are missing and pitch control surfaces become unreliable.
  3. Optimizing for point prediction accuracy: The model's value is in producing interpretable surfaces, not in predicting the exact speed of one pass.

Edges

🔑 Hidden Causal Lever

Expected Ball Speed Reveals Intent Independently of Outcome

A pass played at 20 m/s when the expected speed for that context is 12 m/s reveals urgent intent — the player was trying to execute quickly, regardless of whether the pass completed. Conversely, a pass at 8 m/s when expected was 12 m/s suggests hesitation or a deliberate tempo change. The actual-minus-expected speed delta is an intent signal that event data doesn't capture, because event data only records what happened, not how urgently the player tried to make it happen.

What most people do
Analyze pass outcomes (complete/incomplete) and destinations without considering execution speed relative to context.
What the best do
Use ball speed data (from tracking or estimated from timestamps) to compute speed delta per pass. Aggregate by player and situation to reveal who plays with urgency in key moments and who hesitates.
Why it's an edge: A player who consistently exceeds expected speed in the progression phase is a tempo-setter — they're forcing the defense to react faster. A player who consistently underperforms expected speed is a tempo-breaker. Neither shows up in standard pass metrics.
How to exploit: Build expected ball speed models from tracking data. Identify tempo-setters (consistently positive speed delta in progression/final third) and recruit them for pressing systems that demand fast ball circulation.
Devin Pleuler, Toronto FC, StatsBomb Conference, 2021-11-04. Expected ball speed as the foundation for tempo quantification.

Sources

  • Devin Pleuler, Toronto FC, StatsBomb Conference 2021, YouTube, 2021-11-04 — presented modified soccer_map architecture for expected ball speed prediction; compared physics-based (pitch control) and neural network approaches; identified surface interpretability as primary value