Simulação do campeonato brasileiro de 2013 e estimativa de lucros em um sistema de apostas hipotético

Abstract

The study develops a probabilistic model for the 2013 Brazilian Championship, using exponential smoothing to update the probabilities of each team winning, drawing, or losing. The smoothing parameter h was estimated by comparing the model’s probabilities with those implied by Bet365’s opening odds, resulting in h = 0.9. With the model trained up to the 28th round, one million Monte Carlo simulations were performed to estimate title probabilities, relegation risks, and the teams’ expected final points. Cruzeiro showed approximately a 90% chance of winning the championship, and the average difference between actual and estimated points was 2.7 points. In the second stage, the model was used to simulate the profit of a betting house through inflated probabilities. The results showed that as the inflation rate increases, the ratio between actual and expected profit decreases.