Evolving fantasy cricket teams: Applying genetic algorithms for optimal player selection

In fantasy crickets, users are allowed to make virtual teams on the basis of real performance of a player. The classical techniques, e.g. random sampling and systematic replacements restrict optimization in search of large solution space. Genetic Algorithms (GA) exploit the evolutionary strategies such as crossover, mutation and selection to maximize team configurations. In this paper, it has been argued that GA is a better solution to team selection. GA maximizes the expected performance and ensures the compliance with the credit and role limitations. GA has produced better teams than random sampling, systematic replacements and K-means clustering by continuously adapting to the play outcomes of players. This paper shows that GA is able to perform well at the optimisation of selection of fantasy sports clubs and offers a data-driven and customizable process.