Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer

Ordinal regression models are frequently used in academic literature to model outcomes of soccer matches, and seem to be preferred over nominal models. One reason is that, obviously, there is a natural hierarchy of outcomes, with victory being preferred to a draw and a draw being preferred to a loss...

Full description

Saved in:
Bibliographic Details
Main Author: Hvattum L. M.
Format: Article
Language:English
Published: Sciendo 2017-07-01
Series:International Journal of Computer Science in Sport
Subjects:
Online Access:https://doi.org/10.1515/ijcss-2017-0004
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850060911037906944
author Hvattum L. M.
author_facet Hvattum L. M.
author_sort Hvattum L. M.
collection DOAJ
description Ordinal regression models are frequently used in academic literature to model outcomes of soccer matches, and seem to be preferred over nominal models. One reason is that, obviously, there is a natural hierarchy of outcomes, with victory being preferred to a draw and a draw being preferred to a loss. However, the often used ordinal models have an assumption of proportional odds: the influence of an independent variable on the log odds is the same for each outcome. This paper illustrates how ordinal regression models therefore fail to fully utilize independent variables that contain information about the likelihood of matches ending in a draw. However, in practice, this flaw does not seem to have a substantial effect on the predictive accuracy of an ordered logit regression model when compared to a multinomial logistic regression model.
format Article
id doaj-art-78b46b7977a04bfbb230a24e1eb4cc99
institution DOAJ
issn 1684-4769
language English
publishDate 2017-07-01
publisher Sciendo
record_format Article
series International Journal of Computer Science in Sport
spelling doaj-art-78b46b7977a04bfbb230a24e1eb4cc992025-08-20T02:50:26ZengSciendoInternational Journal of Computer Science in Sport1684-47692017-07-01161506410.1515/ijcss-2017-0004Ordinal versus nominal regression models and the problem of correctly predicting draws in soccerHvattum L. M.0Faculty of Logistics, Molde University College, Molde, NorwayOrdinal regression models are frequently used in academic literature to model outcomes of soccer matches, and seem to be preferred over nominal models. One reason is that, obviously, there is a natural hierarchy of outcomes, with victory being preferred to a draw and a draw being preferred to a loss. However, the often used ordinal models have an assumption of proportional odds: the influence of an independent variable on the log odds is the same for each outcome. This paper illustrates how ordinal regression models therefore fail to fully utilize independent variables that contain information about the likelihood of matches ending in a draw. However, in practice, this flaw does not seem to have a substantial effect on the predictive accuracy of an ordered logit regression model when compared to a multinomial logistic regression model.https://doi.org/10.1515/ijcss-2017-0004association footballforecastingordered regression
spellingShingle Hvattum L. M.
Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
International Journal of Computer Science in Sport
association football
forecasting
ordered regression
title Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
title_full Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
title_fullStr Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
title_full_unstemmed Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
title_short Ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
title_sort ordinal versus nominal regression models and the problem of correctly predicting draws in soccer
topic association football
forecasting
ordered regression
url https://doi.org/10.1515/ijcss-2017-0004
work_keys_str_mv AT hvattumlm ordinalversusnominalregressionmodelsandtheproblemofcorrectlypredictingdrawsinsoccer