Dataset Open Access
Harris V Georgiou (MSc,PhD)
Soccer bets and teams DB from coupons
Greece, Aug.2003 - Mar.2007
Copyright (c) 2016 by Harris V. Georgiou
Release: Aug 3, 2016
- Version: 1.1a
- Format: .xls
This file contains important information about the current version of the dataset package. Downloading and using this material hints that you accept the EULA/Terms-of-Use (please read carefully).
We welcome your comments and suggestions.
WHAT'S IN THIS PACKAGE?
- Available file formats
- Files and Datasets
- License Agreement
Athletic betting and successful risk management is one of the most challenging tasks in Machine Learning and A.I. in general. With only limited information and very sparsely distributed data, an algorithm has to model the teams' competency and predict outcomes of games, even when two teams face each other for the first time.
This package contains two main datasets, generated from soccer betting coupons in Greece throughout a 3.5 year period and including every game that was included for any championship and any team. The first dataset corresponds to 43.758 matches with complete odds, final result and goals. The second is team-based statistics from these matches for a total of 1.126 teams from various championships (European and other).
These datasets can be used as benchmark training sets for odds-oriented or team-oriented predictive models and algorithms, relevant not only to athletic betting but also other contexts, lile in in financial risk management. In the case of team-oriented modeling, special ELO-like ranking algorithms can be developed and tested against the official team/country rankings, as in UEFA and FIFA, but tailored to specific teams or championships, in order to evaluate and fine-tune the models.
Note: Team names are in Greek, but they can be used as-is for proper referencing to unique IDs, i.e., practically they can be ignored.
AVAILABLE FILE FORMATS
The datasets are available in the following formats (included):
*.xls : MS-Excel/LibreOffice worksheet data file (exportable)