Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Strategiskt spelande för en positiv avkastning

Wallman, Jesper LU and Demmler, Sebastian LU (2012) STAH11 20112
Department of Statistics
Abstract (Swedish)
This study compiles a total of five statistical models of how to predict the probable outcome in football games in the German football league, the Bundesliga. The models are based on quantitative historical data from 1683 games and estimate probabilities of the outcomes home win (1), draw (X) and away win (2). The purpose of the study is to test whether a more accurate prediction can be formulated with increased historical data added to the models. From this we examine if these statistical models can outperform the probabilities generated by the gambling company Digibet. It is further tested whether the models through strategic gambling may generate a positive return. The models used are based on an ordered-probit regression with 13... (More)
This study compiles a total of five statistical models of how to predict the probable outcome in football games in the German football league, the Bundesliga. The models are based on quantitative historical data from 1683 games and estimate probabilities of the outcomes home win (1), draw (X) and away win (2). The purpose of the study is to test whether a more accurate prediction can be formulated with increased historical data added to the models. From this we examine if these statistical models can outperform the probabilities generated by the gambling company Digibet. It is further tested whether the models through strategic gambling may generate a positive return. The models used are based on an ordered-probit regression with 13 variables. In conclusion, the study shows that more historical statistical data results in a more appropriate model for prediction. However, there is no evidence to demonstrate that the statistical models can predict probabilities better than Digibet. A final conclusion is that applying strategic gambling in accordance with the model can generate a positive return. (Less)
Please use this url to cite or link to this publication:
author
Wallman, Jesper LU and Demmler, Sebastian LU
supervisor
organization
course
STAH11 20112
year
type
M2 - Bachelor Degree
subject
keywords
Odds, Bundesliga, Ordered-probit, Predict and Strategic gambling, Probability
language
Swedish
id
2338507
date added to LUP
2012-02-09 18:00:02
date last changed
2012-02-09 18:00:02
@misc{2338507,
  abstract     = {{This study compiles a total of five statistical models of how to predict the probable outcome in football games in the German football league, the Bundesliga. The models are based on quantitative historical data from 1683 games and estimate probabilities of the outcomes home win (1), draw (X) and away win (2). The purpose of the study is to test whether a more accurate prediction can be formulated with increased historical data added to the models. From this we examine if these statistical models can outperform the probabilities generated by the gambling company Digibet. It is further tested whether the models through strategic gambling may generate a positive return. The models used are based on an ordered-probit regression with 13 variables. In conclusion, the study shows that more historical statistical data results in a more appropriate model for prediction. However, there is no evidence to demonstrate that the statistical models can predict probabilities better than Digibet. A final conclusion is that applying strategic gambling in accordance with the model can generate a positive return.}},
  author       = {{Wallman, Jesper and Demmler, Sebastian}},
  language     = {{swe}},
  note         = {{Student Paper}},
  title        = {{Strategiskt spelande för en positiv avkastning}},
  year         = {{2012}},
}