Document Type
Thesis or dissertation
Date of this Version
2017
Abstract
The Swiss tournament structure is well-known and commonly used, particularly in chess, be- cause of its seemingly simple heuristic of preferring matchups with teams of similar win count. However, it has proven difficult to simulate. This paper presents a framework for simulation and investigation of Swiss-style tournaments and other similar tournaments. This paper con- siders the desirability of this tournament structure in different contexts, using simulation and empirical results, with particular emphasis on recovering partial and full rankings. The re- sults show that while the Swiss structure performs well under certain criteria, under many circumstances it does not outperform a simple random pairing of teams.
Keywords
pairwise comparisons, tournaments, ranking, matching
Date Posted: 14 September 2017