| Summary: | A sports rating system is a system that analyses the results of sports
competitions to provide ratings for each team or player. Usually, in a football match,
the audience will predict which team will win based on the goals scored by the team
at half-time or penalty. This prediction is important because when evaluating match
results, it is important to first compare the potential strength of the teams involved
in the match. Due to this, the main goal of this research is to compare the
performance of the team rating system using Elo Rating and Pi Rating when
forecasting match outcomes in association football. The well-known Elo Rating
system is used to calculate team ratings, whereas a Pi Rating is used to predict the
football match results based on a team’s performance to win the match when
playing home or when playing away. Two different techniques are used to generate
forecasts. Both types of models can be used to generate pre-game forecasts. The Elo
ratings worked better when predicting matches from a large data set. The Pi Rating
system applies to any other sport where the score is considered as a good indicator
for prediction purposes, as well as determining the relative performances between
adversaries. Data used in this study focuses on the dataset football match by
Switzerland Super League. The dataset from the Football-Data.co.uk website is a
dataset composed of around 1421 data of matches of the Switzerland Super League.
The research figures out the classification model based on the Decision Forest
classifier is an effective classifier with 68% f-measure for Pi Rating and 73% for f�measure Elo Rating. Therefore, Elo Rating is the best team rating system to predict
football competitions.
|
|---|