Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

An Adapted Off-Ball Scoring Opportunity Framework for Evaluating Crossing Situations in Football

Holmstedt, Tor LU (2026) STAN40 20261
Department of Statistics
Abstract
Traditional football analytics largely evaluates players through on-ball events such as shots, passes and dribbles, leaving the much longer periods of off-ball play poorly measured. This thesis develops a spatial framework for evaluating individual player contributions in crossing situations, where a cross is defined as an aerial or driven delivery from a wide area into the central penalty zone. The framework adapts the Off-Ball Scoring Opportunity (OBSO) model of Spearman (2018), replacing its general-purpose transition component with an XGBoost cross difficulty model that estimates the probability of a delivery reaching a teammate, and its scoring component with a logistic model trained exclusively on post-cross shots. The
training... (More)
Traditional football analytics largely evaluates players through on-ball events such as shots, passes and dribbles, leaving the much longer periods of off-ball play poorly measured. This thesis develops a spatial framework for evaluating individual player contributions in crossing situations, where a cross is defined as an aerial or driven delivery from a wide area into the central penalty zone. The framework adapts the Off-Ball Scoring Opportunity (OBSO) model of Spearman (2018), replacing its general-purpose transition component with an XGBoost cross difficulty model that estimates the probability of a delivery reaching a teammate, and its scoring component with a logistic model trained exclusively on post-cross shots. The
training models are fitted on approximately 36,000 crosses from the Big 5 European leagues in 2015/16, and the framework is applied to ten A-League matches from 2024/25 with synchronised tracking and event data. Three metrics are derived from the resulting OBSO field: Decision Quality, which evaluates the decision made by the crossing player; Player OBSO, which quantifies the off-ball positioning value generated by attackers in the penalty area; and Receiver Overperformance, which captures individual qualities involved in successfully reaching crosses. The metrics are qualitatively validated and applied to the A-League sample to illustrate how the framework can identify player contributions that event-based statistics do not record. (Less)
Please use this url to cite or link to this publication:
author
Holmstedt, Tor LU
supervisor
organization
course
STAN40 20261
year
type
H1 - Master's Degree (One Year)
subject
keywords
Football analytics, Sports analytics, Tracking data, Off-Ball Scoring Opportunity (OBSO), Pitch control, Crossing
language
English
id
9231638
date added to LUP
2026-06-16 10:13:23
date last changed
2026-06-16 10:13:23
@misc{9231638,
  abstract     = {{Traditional football analytics largely evaluates players through on-ball events such as shots, passes and dribbles, leaving the much longer periods of off-ball play poorly measured. This thesis develops a spatial framework for evaluating individual player contributions in crossing situations, where a cross is defined as an aerial or driven delivery from a wide area into the central penalty zone. The framework adapts the Off-Ball Scoring Opportunity (OBSO) model of Spearman (2018), replacing its general-purpose transition component with an XGBoost cross difficulty model that estimates the probability of a delivery reaching a teammate, and its scoring component with a logistic model trained exclusively on post-cross shots. The
training models are fitted on approximately 36,000 crosses from the Big 5 European leagues in 2015/16, and the framework is applied to ten A-League matches from 2024/25 with synchronised tracking and event data. Three metrics are derived from the resulting OBSO field: Decision Quality, which evaluates the decision made by the crossing player; Player OBSO, which quantifies the off-ball positioning value generated by attackers in the penalty area; and Receiver Overperformance, which captures individual qualities involved in successfully reaching crosses. The metrics are qualitatively validated and applied to the A-League sample to illustrate how the framework can identify player contributions that event-based statistics do not record.}},
  author       = {{Holmstedt, Tor}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{An Adapted Off-Ball Scoring Opportunity Framework for Evaluating Crossing Situations in Football}},
  year         = {{2026}},
}