Estimating the peak age of chess players through statistical and machine learning techniques

Because chess is played in formal tournaments and competitive environments, it requires physical and mental endurance. This endurance declines as the years progress and can decrease the player's performance. As the player's age increases, elements such as strategic thinking, game analysis, and psychological endurance come to the fore. In chess, age is the most important variable, although it is not the sole determinant of a player's abilities and achievements. In this study, the age at which Grandmaster level chess players reach the highest ELO levels and the 2,700 ELO threshold was predicted. For this purpose, 12 forecasting models were created using 11 machine learning methods with various variables. The model results were interpreted and the age at which some promising young players reached the 2,700 ELO level was determined. This study finds that the average peak ELO age for Grandmasters is approximately 30.65, with variations based on factors such as early attainment of the GM title and gender differences. To enhance the reliability of prediction results, the percentile bootstrap method was employed across all machine learning models. This approach allowed for the calculation of confidence intervals, providing a more reliable interpretation of the predicted values. These results provide insights into the career trajectories of chess players at the highest levels. This study provides a good alternative for the calculation of classification scores in sports that are uncertain and difficult to predetermine.