Abstract
Tennis is one of the few sports that allows players to challenge umpire calls. This study investigated decision-making with the use of the challenge system. Regression and machine learning models were applied to data of 1037 challenged and 154,648 unchallenged shots at the 2016 Australian Open to investigate factors that explain the use and success of challenges. On shots closer to lines, players were more likely to challenge fast shots (OR = 1.41, p < .001). For male players, associations were also found for point importance (OR = 1.12, p = .003) and a previous successful challenge (OR = 1.84, p < .001). Ball and player location were the most important factors for successful challenge. Players frequently delayed challenges and there was evidence of impact aversion among officials. Players could use challenges more effectively and tennis stakeholders could improve the integrity of the system by taking steps to reduce delayed challenges and impact bias among chair umpires.