About ACE (Beta)Follow predictions by ACE on our Twitter feed. The Athlete Contribution Estimator (ACE) is an algorithm that, for each position, determines and quantifies actions that contribute to scoring points (e.g., offloads) and conceding points (e.g., missed tackles). ACE calculates the net contribution of each player to the predicted score margin (the score of the home team minus the score of the away team). A positive number indicates that the home-team player is stronger than their opposite number, and a negative number indicates that the away-team player is stronger. The image below shows ACE’s prediction when the Hurricanes hosted the Crusaders in Round 4 of the 2018 Super Rugby season. In evaluating the number ten position, ACE predicted that Beauden Barrett gave the Hurricanes a 9.7 point advantage over the Crusaders, who fielded back-up first-five Mitch Hunt. Interestingly, if the Crusaders had been able to field first-choice first-five Richie Mo’unga (who was out injured), the advantage to the Hurricanes would have been reduced to just 4 points. Comparing Matt Proctor and Jack Goodhue, ACE estimated that the Crusaders had an edge worth 0.8 points in the (outside) centre position.
ACE calculates the predicted score margin for each game by adding together (1) net score contributions for each position, (2) home field advantage, and (3) net ‘team dynamics’. What are team dynamics? Not all attributes that contribute to scoring and conceding points are captured by available player statistics. For example, we do not have metrics that measure the scrumming ability of tight head props, or tactical player substitutions. These characteristics are captured by team dynamic variables. In the Hurricanes vs Crusaders prediction above, (1) the sum of net score contributions over all positions favoured the Hurricanes by 11.2 points, (2) home field advantage was 4.5 points, and (3) the Crusaders held a 4.5 points advantage in team dynamics. This resulted in a predicted net score of 11.2 + 4.5 – 4.5 = 11.2 points (or 11 points when rounded to the nearest whole point). This indicates that the home team, in this case the Hurricanes, was expected to win by 11 points. The actual result of the match? A 10-point victory to the Hurricanes. A Beta version of ACE will operate for selected Super Rugby games in 2018. Development and testing of ACE will continue throughout the competition. ACE is developed and maintained by Niven Winchester and Will Beard. Winchester is an economist at the Massachusetts Institute of Technology and founder of RugbyVision.com. Beard is an engineer with a background in big data and statistical analysis. To find out more about ACE or to enquire about further applications of our research, contact firstname.lastname@example.org.