Forecasting the auction prices of Indian Premier League cricket players with neural networks

When it comes to professional leagues in team sports, an important research question occurs in the context of the choice and selection of individuals in teams and the linkage of the auction process to the economic interest related to player’s wages/fees. The critical relationship between wage determinations in an economical manner wherein the basis of performance/output is on the contributions of an individual vis-a-vis the rest of the team, which is essential to maintain. More recently, the game of cricket has seen the start of the latest version called Twenty-Twenty (20-20) cricket, which has gained worldwide popularity. Measuring the performance of cricket players vis-à-vis their remuneration is a complex process as it involves assessing different cricketing factors (mainly batting and bowling). We collate a longitudinal and complex data set from the Indian Premier League, the world’s largest and richest franchise-based Twenty-Twenty cricket competition. We utilize the artificial neural network (ANN) models approach to aggregate the two key factors, that of batting and bowling, as it may not be able to combine the same in a linear fashion to depict the overall performance of cricket players. Therefore, our main objective was to relate with the market value of players to their overall performance. In this study, the player’s auction prices are related to the overall performance index (calculated using ANN models). This research thus contributes as it approaches the problem with ANN models-based methods to measure player’s performance on one hand and, through our investigation, further relates the same to the player’s auction prices. We argue that this study is generalizable, as it can be extended not only for cricket, but applicable to any game that has multiple unrelated performance measures.