Cricket is a dynamic sport, making the selection of key performance indicators (KPIs) challenging. Objective: The study aims to identify KPIs in Twenty-20 (T20) cricket affecting match outcomes. Methods: Cricket performance data was analysed from three seasons of male T20 matches, identifying 136 performance indicators (PIs). The random forest algorithm and lasso logistic regression were used to develop a model to predict match outcomes. Results: The hybrid model achieved 85.9% accuracy with leave-one-out cross-validation statistical analyses. Sixteen KPIs were identified and ranked by importance including wickets lost in the last six overs, two or more wickets in the second innings, run rate in the last six overs, wickets by seam and spin bowling, batting strike rate, singles percentage in the second innings, sixes in the first innings, overs bowled by seam, runs in last six overs, sixes in middle overs, total catches in second innings, dot ball percentage, opening partnership runs, dot balls in the opening six, and singles in the last six. Conclusions: Cricket match performance in the final overs, especially bowling strike rate and scoring runs, were crucial for successful match outcomes. These KPIs offer insights into team strategy, player selection, and match performance evaluation in T20 cricket.

Based on a random cluster representation, the Swendsen–Wang algorithm for the Ising and Potts distributions is extended to a class of continuous Markov random fields. The algorithm can be described briefly as follows. A given configuration is decomposed into clusters. Probabilities for flipping the values of the random variables in each cluster are calculated. According to these probabilities, values of all the random variables in each cluster will be either updated or kept unchanged and this is done independently across the clusters. A new configuration is then obtained. We will show through a simulation study that, like the Swendsen–Wang algorithm in the case of Ising and Potts distributions, the cluster algorithm here also outperforms the Gibbs sampler in beating the critical slowing down for some strongly correlated Markov random fields.