IBM Watson Drives Overwatch League’s Power Rankings

Is there a way to objectively determine the best player or team in the Overwatch League at any given time? This is a question fans and analysts have been trying to answer for as long as competitive Overwatch has existed. The game produces an enormous range of stats and numbers for each of its 32 heroes, so analyzing both individual and team performance can be a complex matter.

Previous statistical insights introduced by the league — including Player Impact Rating, teamfight win rates, and more — laid the groundwork for understanding the nuances of the game at a professional level, and IBM is providing fans with even more precise data in the 2021 season by introducing Power Rankings with IBM Watson.

Click here to learn how IBM is enhancing the understanding of the Overwatch League — and changing the way fans watch and discuss the game.

Power Rankings with IBM Watson — developed by a team of data scientists at IBM alongside Overwatch League experts — uses AI algorithms to analyze every player and team’s performance in every Overwatch League match. Trained on data from previous seasons, Watson’s algorithm correlates those statistics and assigns a weight to each, determining how each factor tips the balance toward a win or a loss. The combination of the data capture, correlations, and weighting creates the Power Rankings for each player and team.

Every week, IBM will provide a comprehensive list of the top performers in the league on the Power Rankings with IBM Watson page on the company website. Here were the top-ranked players from the 2020 season as determined by Watson’s algorithm:

Updated Power Rankings will be posted every Monday during the 2021 season, with analysis following on Wednesdays.