Quantifies each player's contribution to opening valuable space for teammates through off-ball movement. For every tracking frame, pitch control is computed with and without each off-ball player; the EPV-weighted difference in controlled space measures how much valuable territory each player creates through their positioning and movement. Heuristic counterfactual following Fernández & Bornn (2018); no trained weights for the per-player attribution.

Part of the (Right! Luxury!) Lakehouse soccer analytics platform.

Method Description

Algorithm

For each tracking frame:

Compute the baseline pitch-control surface with all players present (from wf-pitch-control).
For each off-ball player p:
- Recompute pitch control with player p removed.
- Compute the EPV-weighted difference between the baseline and the counterfactual surface.
- The difference is player p's space-creation value at this frame.
Aggregate per-player across all frames in a match.

The EPV weighting (from the OBSO trained grids) ensures that space opened in high-value areas (e.g., inside the penalty box) counts more than space opened on one's own goal line.

The implementation uses the "factor-out loop-invariant computation" optimisation: the transition × EPV grid product is a match-level constant, so the counterfactual surfaces differ only in their pitch-control component. This collapses what would be N sequential pitch-control evaluations into a single vectorised NumPy broadcast, cutting GPU runtime by over an order of magnitude.

Reference

Fernández, J. & Bornn, L. (2018). Wide Open Spaces: A Statistical Technique for Measuring Space Creation in Professional Soccer. MIT Sloan Sports Analytics Conference. https://www.sloansportsconference.com/research-papers/wide-open-spaces-a-statistical-technique-for-measuring-space-creation-in-professional-soccer

Inputs

No training data — this is a heuristic counterfactual. Upstream EPV transition and reachability grids are trained, but they are consumed as a static input here.

Runtime inputs:

Input	Source
Tracking frames	`luxury-lakehouse/pitch-control-tracking`
EPV transition × reachability grid	`luxury-lakehouse/obso-trained-grids`
Pitch control surfaces	From `wf-pitch-control`

Execution

Batch computation: HF Jobs GPU (l40sx1), script scripts/compute_space_creation_hf.py. Typical duration: 45 minutes for full corpus. Output dataset: luxury-lakehouse/space-creation-values.

PR 7 (ADR-011 close-out): the downstream gold mart fct_space_creation now carries Kimball-conformed FKs (match_key, player_key, data_source) alongside the legacy native columns during the 2026-07-22 dual-column window. The per-row team column remains a 'home'/'away' role string (the source data does not carry a real team identifier at frame grain); team-level resolution is deferred until a use case demands it.

See workflow-cards/wf-space-creation.yaml for the full operational contract.

Intended Use

Off-ball contribution ranking: Identify players who open the most EPV-weighted space for teammates
Tactical profiling: Surface run patterns and decoy movements that create value without touching the ball
Research: Reproducible implementation of Fernández & Bornn (2018) on open tracking data

EU AI Act — Intended Use and Non-Use

This method is published for research and reproducibility purposes on public, open-licensed tracking data. It is not intended for, not validated for, and not supplied to any use that would fall within Annex III §4 (Employment, workers management and access to self-employment) of Regulation (EU) 2024/1689 — including recruitment or selection of natural persons, decisions affecting work-related contractual relationships, promotion, termination, task allocation based on individual traits, or the monitoring and evaluation of performance and behaviour of workers for employment decisions.

Any deployer who wishes to use this method for such a purpose is responsible for performing their own conformity assessment under Article 43, for drawing up the technical documentation required by Article 11 and Annex IV, for implementing the human oversight measures required by Article 14, for declaring accuracy metrics under Article 15, and for ensuring the data governance obligations of Article 10 are met. Note specifically that the tracking corpus contains no protected attributes and therefore cannot support the group-fairness audits required by Article 10(2)(g) without ingesting additional personal data.

See the AI_GOVERNANCE.md gap analysis in the source repository for the project's full risk classification, re-classification triggers, and governance posture.

Limitations

One-player counterfactual only. Removing player p does not model the defensive reorganisation that their absence would trigger. The metric therefore slightly under-credits players who attract multiple defenders.
No ball-trajectory adjustment. The ball position is held fixed across the counterfactual; in reality, removing an attacker would change the decision space of the ball carrier.
Tracking-data dependence. Event-only matches do not produce space-creation values. Only matches with tracking coverage are processed.
Upstream grid dependence. Inherits biases in the pre-trained EPV transition and reachability grids.
Inherits pitch-control limits. All limitations of pitch control (2-D, no ball physics, no defender intent model) propagate.

Files

No persisted model weights. The method is implemented in source:

scripts/compute_space_creation_hf.py

Outputs are published as the HF Hub dataset luxury-lakehouse/space-creation-values.

Citation

bibtex

@inproceedings{fernandez2018wideopenspaces,
  title={Wide Open Spaces: A Statistical Technique for Measuring Space Creation in Professional Soccer},
  author={Fern{\'a}ndez, Javier and Bornn, Luke},
  booktitle={MIT Sloan Sports Analytics Conference},
  year={2018}
}