Modelling Football as a Markov Process : Estimating transition probabilities through regression analysis and investigating it’s application to live betting markets

Detta är en Kandidat-uppsats från KTH/Matematisk statistik; KTH/Matematisk statistik

Författare: Gabriel Damour; Philip Lang; [2015]

Nyckelord: ;

Sammanfattning:

This degree thesis aims for a modeling of football set pieces (i.e Throw Ins, Free Kicks, Goal Kicks and Corners) through the use of Markov theory. By using regression analysis on a various range of covariates we will try to estimate the transition probabilities of such a process from a state to another and investigate what factors might have an impact on these probabilities. Although not reaching a sufficiently high level of variance explanation, the model constructed shows strong significance and let us believe that an articulation of it could lead to a strong model for these set pieces. Furthermore we will proceed with an analysis addressing the application of such modeling within the pricing processes of betting companies, based on a case study of Metric Gaming. Undertaking an operational management perspective, we will assess which level of implementation of such modeling is the most efficient, and what consequences it will have in two sub-perspectives; the risk management and branding of the company

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)