Please use this identifier to cite or link to this item: http://hdl.handle.net/1959.3/51468

Title

Predicting the number of runs scored per over in one-day international cricket matches

Author(s)

Bailey, Michael;  Clarke, Stephen

Abstract

One-Day International cricket (ODI) is a popular sport worldwide. The advent of the Internet has increased opportunities for punters to wager on differing outcomes associated with each match, with one outcome of interest being the number of runs scored per over. Using information gathered from 627 past ODI matches (55000 overs), specific 'within match' and 'between match' variables were identified and explored to determine the statistical significance of their relationship with the number of runs scored. Separate models were constructed for both first and second innings with ten first-order variables and four second-order variables found to be highly significantly related to runs per over in both innings with a p-value less than 0.0001. These variables include the number of overs bowled, the stage of the game, the number of wickets fallen, recent wickets fallen, the magnitude of the existing partnership, current run rates, host country, team batting strength and opposition bowling strength. A multivariate linear regression model was used to combine and weight the relative importance of all variables with goodness of fit determined using the absolute average error between predicted and actual results, the maximum log-likelihood and an R-square statistic. This process serves to ascertain and rank the relative importance of various predictors and shed some light on the predictability of runs per over in ODI cricket.

Publication type

Conference paper

Research centre

Swinburne University of Technology. Faculty of Life and Social Sciences

Source

Proceedings of the 9th Australasian Conference on Mathematics and Computers in Sport, Tweed Heads, Australia, 30 August - 3 September 2008, p. 23-31

Publication year

2008

FOR Code(s)

010401 Applied Statistics;  010406 Stochastic Analysis and Modelling;  140305 Time-Series Analysis

Keyword(s)

Linear regression;  ODI Cricket;  Predictors of runs scored

Publisher

MathSport (ANZIAM)

ISBN

9780957862340

Publisher URL

http://www.anziam.org.au/MathSport+Conference+2008

Peer reviewed