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
