A comprehensive game plan and strategic tactics are critical to winning football, but how much does a team's unpredictability in moving the soccer ball around the pitch matter?
In a new article published in PLOS One , an international team of researchers analysed event data from top-tier association football competitions to provide insights into match analysis, player tactics and game strategy.
"Soccer is low-scoring, so a couple of moments can swing a match, and simple statistics like possession or shot counts do not always capture who performed better. Our approach measures how unpredictably and widely a team moves the ball across a match," says Dr Sergiy Shelyag , Associate Professor in Applied Mathematics and Data Science at Flinders University.
"We found that 'all zones count' metric, the one that values every region of the field equally, including rarely used areas, aligns best with winning.
"Teams that spread their play unpredictably across the full field tend to succeed more than teams that confine their play unpredictability to a few hotspots."
Using the StatsBomb Open Data dataset, the researchers from Australia and Europe measured how widely and unpredictably teams moved the ball across the field using an information-theoretic metric called Spatial Event Distribution Randomness (EDRan).
In the men's dataset, higher EDRan was positively associated with winning, and the association weakened toward the end of matches, consistent with winning teams becoming more predictable late in games.
The study also built match-winner prediction models based on the EDRan differences between teams.
In men's matches, models using the equal-weight, whole-field version of the metric outperformed models based on the commonly used Shannon-entropy formulation, with the best-performing model achieving 80.61% accuracy in predicting match winners.
"A mathematically grounded, information theory-based view suggests that broad, field-wide unpredictability is a stronger path to success than being tricky in the same old places," concludes Dr Shelyag.
The research group was led by experts at Deakin University's School of IT and Centre for Sport Research, and the UK universities of Coventry and Aston, Birmingham.
The new article, 'Maximising ball movement unpredictability in Association Football: A R´enyi Entropy-based approach to optimising event distribution randomness' (2026) by Ishara Bandara (Deakin), Sergiy Shelyag, Sutharshan Rajasegarar, Dan Dwyer, Eun-jin Kim and Maia Angelova has been published in PLOS One DOI: 10.1371/journal.pone.0326800.
Acknowledgements: The authors thank StatsBomb for providing publicly available event data from association football.
