Reducing the non-uniformity of the group draw in sports tournaments

Abstract

The group draw of a sports tournament requires assigning teams to groups of (almost) the same size. The most important criteria for a draw procedure are balance, randomness, and transparency, which cannot be satisfied simultaneously if draw constraints exist. Organisers usually use the so-called Skip mechanism, a method based on a random sequential draw of the teams from pots, in order to ensure balance and transparency. However, the Skip mechanism is non-uniformly distributed: the valid assignments are not necessarily equally likely. We quantify this distortion if a group can contain at most two teams from a given set S, which poses a serious challenge for the Skip mechanism. Our study provides exact results for an arbitrary number of teams when there are three pots and two pots contain only one team from the set S, as well as complete enumeration for small problems with three pots and at most five teams per pot. We also analyse three real-world case studies from basketball and football. It turns out that the optimal design considers the pots in decreasing order according to the number of teams in the set S. These results can be used to identify the least distorted transparent draw procedure, and decide whether the extent of non-uniformity calls for further actions.