DOTA2 Winner Predictor using Machine Learning

Abstract

DotA2 (Defense of the Ancient 2) is an online multiplayer game. It’s a real time strategy game played between two teams of five players. DotA2 rules are extremely mind-boggling. We have created a model to predict the winner of Dota2 by using Logistic Regression. By obtaining the information of heroes and the average MMR of all the ten players we are predicting the outcome of the match. The win percentage is dependent on hero selection and MMR of players. We also determine the optimum 10th pick by invoking a brute force execution.