A probabilistic match classification model for sports tournaments
创建于 更新于
摘要
We propose a probabilistic simulation-based match classification that measures, for each team, the gain from winning and the loss from losing (relative to a 0-0 draw) to determine whether teams are indifferent, offensively- or defensively-incentivised; this yields six match categories (stakeless, defensive/offensive asymmetric, antagonistic, defensive, offensive) and a strength metric κ to quantify incentive intensity [page::0][page::5][page::6][page::7][page::13]. The model, applied to pre-2024 group stages and the 2024/25 incomplete round-robin leagues, finds fewer purely stakeless matches but substantially more offensive matches — and an alarming relative increase in defensive matches that raise collusion risk under the new format [page::2][page::13][page::14][page::16].
