Bayesian forecasting of UEFA Champions League under alternative seeding regimes

The Effect of Seeding on Tournament Outcomes: Evidence From a Regression-Discontinuity Design

“…In a similar vein, Scarf et al (1980) confirm that seeding increases the correlation between team rank at tournament entry and exit. Dagaev and Rudyak (2019) and Corona et al (2019) find qualitatively similar results when evaluating the 2015/2016 seeding reform, which reserved the first pot in the UEFA Champions League for the title holder and the champions of the top-7 nations, instead of seeding exclusively based on the team coefficient. 1 Their simulations indicate that the tournaments prior to the reform were characterized by higher average coefficients for the finalists, as well as more balanced match-ups in the final.…”

“…1 Their simulations indicate that the tournaments prior to the reform were characterized by higher average coefficients for the finalists, as well as more balanced match-ups in the final. Corona et al (2019) conclude that the reform particularly favored national champions with low coefficients, such as the 2015 Dutch champion PSV Eindhoven, by placing them in pot 1 rather than pot 3, thus shielding them from facing an additional highly ranked competitor like FC Barcelona or Bayern Munich. It should be noted that the 2015/2016 reform induced perverse incentives for national competitions by awarding the eighth position in the first pot to a lower ranked league if the title holder of the Champions League and the national competition coincide (Csató, 2020).…”

Soccer Team Performance Forecasting using artificial neural network

“…Case II: at learning rate = 1 Fig. 11: MSE results for LR=1 at 1000 epochs Figure (11), illustrated that in the network trained at 1000 epochs the least MSE value is 0.04921with a 15 hidden neurons while, the maximum MSE value is 0.054578 appeared at 13 hidden neurons. 13), illustrated that in the network trained at 2000 epochs the least MSE value is 0.013125 with a 15 hidden neurons, while the maximum MSE value is 0.017422 appeared at 10 hidden neurons.…”

“…Corona et. al., in [11] proposed a Bayesian approach for simulating the UEFA Champions League results under alternative seeding regimes. They indicated that the changes in 2015 tended to increase the uncertainty over progression to the knock-out stage, but made limited difference to which clubs would contest the final.…”

Optimal Tournament Design: Lessons From the Men’s Handball Champions League

“…Most numerical studies of tournament designs apply specific models for simulating match results (Corona et al, 2019;Dagaev & Rudyak, 2019;Goossens et al, 2012;Lasek & Gagolewski, 2018;Scarf et al, 2009;Scarf & Yusof, 2011), but we do not follow this approach due to several reasons. First, general works comparing different competition formats (Appleton, 1995;Marchand, 2002;McGarry & Schutz, 1997) or ranking methods (Mendoņ ca & Raghavachari, 2000) avoid the use of specific prediction models.…”