Comparison and improvement of sport ranking algorithms

OData support
Imre Gábor
Department of Automation and Applied Informatics

An immense number of tournaments are organised every year. These employ different algorithms for scheduling matches and ranking players, however, most of the current tournament algorithms are suboptimal. Since many build on preliminary ranking and/or drawing, this causes the result to be inaccurate on many occasions. For example, the runner-up in more than 40% of the Football World Cups is not the second best team.

Many tournament organiser software products exist on the market and multiple online services are also available for organising given types of tournaments.

Also, multiple publications have been written on ranking players based on available results.

In this thesis, I present the benefits and disadvantages of current methods and give proof using simulation. I also describe new ideas for new methods and compare them.

First I describe the testing environment and introduce the graphical interface. Then I present widely used and newly developed tournament algorithms and give detailed comparison of them. Lastly I highlight interesting points of the implementation.


Please sign in to download the files of this thesis.