Bayesian Change-point Analysis of the Effects of 3-points-for-a-win Rule in Soccer

Abstract

One of the aims of the three-points-for-a-win (3pfaw) rule in world-soccer was to make soccer more exciting. Aylott and Aylott (2007) explore the effect of this rule by defining an ‘excitement index’. In this work, we argue that the proportion of decided matches captures the effects of the rule more correctly and use Bayesian change-points modeling (Raftery and Akman, 1986) to re-analyze the data from seven leagues reported in Aylott and Aylott (2007). Preliminary results support a change-point model in which the proportion of decided matches increased after the introduction of the 3pfaw-rule in most of the leagues; but the time-to-change varies across the leagues studied.

References:

Aylott, M., and Aylott, N. (2007), A Meeting of Social Science and Football: Measuring the Effects of three points for a win. Sport in Society 10, 205-222

Raftery, A. E., and Akman, V. E. (1986), Bayesian Analysis of a Poisson Process with a Change-point. Biometrika 73, 85-89.