Cricket Match Winning Prediction

Abstract

The multi-billion dollar industry is cricket betting. There is also a great incentive for models which can forecast the results of games and overcome bookers' odds. The objective of this thesis was to explore the extent to which the results of cricket matches can be predicted. The English twenty over the county Cricket Cup was the aim competition. About 500 teams and player numbers emerged from the initial features alongside the engineered features. First, the versions with only team features, then all team and player features were optimized. In individual seasons, the result has been tested on the basis of each training during the past season results. The optimum model was a straightforward method of estimation paired with dynamic hierarchical characteristics and a benchmark for the gaming industry was considerably higher. It seems magic to predict the future if a prospective buyer wants to buy the goods in advance or figures out where asset prices are concerned. If we can forecast something's future accurately, we have a huge advantage. This magic and mystery have been only amplified by machine learning