Effect of Match Length on Ranking Accuracy of round Robin Tournaments: A Computer Simulation Study |
| |
| Authors: | Harry A King |
| |
| Institution: | Physical Education Department , San Diego State University , San Diego , CA , 92182 , USA |
| |
| Abstract: | Computer simulation techniques were used to study the accuracy with which true rankings of table tennis players could be estimated from round robin tournaments based on matches of different length. Tournaments based on best-of-three-games, 21-points-to-win, and 11-points-to-win matches were studied. The simulation design also allowed study of the effect of both the number of players and the homogeneity of their ability on the efficacy of the three types of tournaments. Accuracy of the rankings produced by different types of tournaments was ascertained by computing Spearman's Rho rank correlations between the known, true ranking (known from the statistical model on which the simulated model was based) and empirical rankings in repeated tournaments. Results demonstrated that for groups of players of mixed ability, tournaments based on matches of shortened length were highly effective in estimating the true ranking, Spearman's Rho coefficients being generally well above .90. Tournaments based on longer matches exhibited only a small increase in accuracy of estimation of the true rankings. The difference between mean Spearman's Rho coefficients for tournaments based on best-of 3-games and 21-point matches approximated .02; between coefficients for tournaments based on 21-point matches and 11-point matches this difference approximated .03. |
| |
| Keywords: | |
|
|
