Multi Criteria Decision Making in Fantasy Sports

Football is a very famous sport worldwide and continues gaining traction to this day. As the interest in the game rises, so do the methods of interacting with the game. One such avenue is the FPL (Football Premier League). Recent years have seen a huge spike in interest for virtual, fantasy playing applications like the FPL. A person participating is usually heavily biased towards certain players or teams. This inherent bias causes them to make certain irrational decisions which may not provide the best results. There exists a huge gap in the market for an algorithm-based approach to team selection and the aim of this research is to make an optimal application capable of filling that void. The premise of this research is to create a team recommender system that optimizes one's team selection strategy for the FPL. This research works by introducing objectivity and eliminating biases in the team selection process. It incorporates the usage of TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) and criteria-based metrics along with an XG-Boost regression predictor, with the aim of filling the gaps in the current systems and to maximize the returns while minimizing the cost associated with it. The approach implemented involved the usage of XGBoost, yielding a standard deviation of 0.041188.