To Mr Zhe Wang at Google DeepMind, an artificial intelligence (AI) company, corner kicks are like games of chess. Partly because both feature opposing sides poised to react to a single imminent move. But also, no doubt, because they may also one day be revolutionised by AI.

AI models thrive where there is abundant data. Football more than satisfies this requirement. Elite players wear vests that measure heart rate, position, speed and force exerted; team analysts watch hours of footage to tally possession percentages and numbers of passes, shots and goals. In a paper published on March 19 in Nature Communications, Mr Wang and his colleagues worked with staff at Liverpool Football Club to feed this data into a statistical model known as a graph neural network (GNN). They then were able to use this model to predict which on-field player would wind up making first contact with the ball with similar accuracy to human experts.