The Current Iteration

Continuing with the discussion of the new tournament simulator that’s under construction, this time with the elements of the match model that relate to each individual iteration.

• TopPlayer, BottomPlayer: pointers to a player record.
• TopReadyTime, BottomReadyTime: times.
  • These elements, and the next two, I intend to use to test the efficacy of various techniques that purport to speed up the tournament, including the infamous grouping of byes.
  • These will be time zero for seeded lines.
• StartTime, FinishTime: times.
  • The time modeling of tournaments will be of no interest for many events, and I expect it will not be a common feature of my analyses. But it is important for the sort of events where rounds can overlap, and where time is a critical constraint.
  • It’s difficult to know how nuanced the time model should be, but I think it will include a parameter for game length, and another for round turnover. I might also store each player’s mean play speed, hoping to recreate in the simulator the problem that certain slow payers will tend to clog brackets, holding up progress.
  • Different games have different lengths, of course, so it probably makes sense to express times standardize the average game length at unity – that will allow the comparison between the number of rounds and the total tourney length to yield a useful efficiency measure.
  • It will also allow me to estimate the amount of dead time in a format, identifying in particular formats that are susceptible to dragging on.
• TopLuck, BottomLuck. double-precision numbers.
  • I’m not sure how I will use this, but I suspect that it will be nice to have around one day.
• MatchPlayed: boolean.
  • This will record the fact of byes. And some tournament models also need matches that will not be played under certain conditions.
• Result: string, chosen from {TopWins, BottomWins, Draw, TopBye, BottomBye}.
  • Draws will not occur in most simulations, but allowing the result will allow me to consider games like chess, where draws are common.
• SkillSpread, ResultSpread: double precision.
  • The first takes the absolute value of the difference in skill levels without regard to the luck encountered, and the second includes the luck factor for that match. This data will, presumably, be useful for calculating some coefficient of competition.