Abstract

Ultimate frisbee is an indoor and outdoor sport played, with a frisbee or disc, by many people across the world including many here, at the university of warwick. It is best described as a cross between American football and netball. You must move the disc into the oppositions end zone, without moving with the disc or making contact with any other player. Ultimate is based on an over-riding rule of fairness and good 'spirit', there is no referee and any foul resolution seeks to continue play from the position the players were in before the foul was committed and doesn't seek to penalise the fouler as it is assumed to be accidental. Although we have enjoyable games during practices and in a mini league system during the winter the real fun of Ultimate is the tournaments, which happen across a whole weekend or somethimes longer. These tournaments, inline with the 'spirit of the game' are played in such a way as everyone has the same number of games as far as is possible. There are several different formats which are used, and which try to smooth over any mis-seeding, which is common due to ad-hoc or 'pickup' teams forming regularly. These standard formats have, I assume, been arrived at through a long process of trial and error. This paper will attempt to show how an empirical modelling approach to tournament design and interaction could lead to a system with a greater potential for designing and simulating tournament formats. Such a system will also have a broader range of uses as the format of the tournament affects many things such as the scheduling of pitch time and the spacing of a teams games. I will demonstrate that this model will be useful to more than just tournament directors as by simulating the tournament, we will allow teams to predict possible outcomes based on hypothetical results, i.e. we can only finish 3rd if we loose this game, but if we win it we are guaranteed to finish at least 5th.