The Quizbowl Resource Center

Ms. Kidd e-mailed me and said she wanted to do some sort of power matching for 24, 32, 40, and 48 teams (depending on the number of registrations). There will be 9 packets. Any ideas?

Well, the 32 is obviously the easiest, and you could simply mimic NT's format for this Saturday. Is a Swiss Pairs system feasible? I've never tried it as the thought of having everyone wait to finish before announcing the next game sounds time-consuming.

I don't think Swiss pairs are anywhere near the ideal that some people hold them up to be, but Swiss pairing is far more efficient then you're giving credit for, since there is no reason to have to wait between matches to be told where to go to your next game. All you have to do is either an HSNCT style card schedule or just hand out a master schedule to teams in the beginning of the day so they can follow where they go depending on when they win or lose.

I guess I was under the impression that Swiss Pairings were dependent upon every team's performance (that is, more than just whether they won or lost) in each round. I stand corrected. Let me ask this then, is a power-matching of a 2^n set dependent on good seeding or should it not matter? That is, if your random draw had a Loyola/Auburn match-up in round 1, the loser would likely work their way back into the top tier anyway, right?

EDIT: I cun speel gud.

Woody Paige wrote:I guess I was under the impression that Swiss Pairings were dependent upon every team's performance (that is, more than just whether they won or lost) in each round. I stand corrected. Let me ask this then, is a power-matching of a 2^n set dependent on good seeding or should it not matter? That is, if your random draw had a Loyola/Auburn match-up in round 1, the loser would likely work their way back into the top tier anyway, right?

EDIT: I cun speel gud.

Well, a card system for 32 guarantees one undefeated team and five one-loss teams; the loser of that matchup could have a 4-1 record.

Anyway, yeah: in some competitions Swiss pairing is done based on other factors (I think I read a tutorial on how to do it for Magic tournaments once; I'm not sure why, as I don't play Magic) and it could be done by margin of victory or something.

I think MITBAT used to do real-time swiss pairings, maybe someone from that club could tell us how they did it?

MIT has specialized software in which the moderator enters the scores into their local computer, they're uploaded to a server, and the next round's games are computed and sent to the moderators, presumably either as soon as they're available or once all games that round are finished (which will sometimes be the same thing.) If you dudes have enough computers to try that, I'd recommend it; if not, a card system is probably the best you can do.

MaS

You can find a system for the first five rounds using 24 or 32 teams on the IHSSBCA Members Only page and in the qbwiki under tournament hosting. I can send you a system I came up with for 48 teams if you want and probably can devise one for 40 teams, though that would be a little messy. Charter has several scenarios given, but I don't always agree with all of their pairings.

Because you have only nine packets (which is probably wise using IHSA format), the playoffs will be a tight fit. You can take the top 16 teams from the morning and power pair them in the afternoon, but that is not ideal because it effectively turns the championship into a single elimination and does not pick a clear 2nd place team. However, I'm not sure how to do better.

Shcool wrote:Charter has several scenarios given, but I don't always agree with all of their pairings.

If you have an alternate version of any card system, please feel free to pass it along to me and I'll include it in those listed. The "Brackets" section of our page is intended to be a resource.

The only card system I really object to is one where the repeat matches start earlier than need be. Such as round 3 a la Weekend of Quizbowl '08. How one deals with the "one-offs" is negotiable.