tadhg.com
tadhg.com
 

Logistical/Organizational Difficulties

23:50 Wed 13 Dec 2006. Updated: 01:12 14 Dec 2006
[, ]

Anything involving group organization tends to be tricky. Any kind of logistical enterprise needs improvement and will breed some kind of discontent with at least some people. Sometimes this discontent brings better ideas, sometimes not. It continually surprises me, though, just how tough it can be to get right even small-group logistics. The example I have in mind is how to organize the pods for sfmagic.org drafting each week. How complicated can it be to get some drafts together on a Wednesday night?

Organized MTG competitions tend to use either Swiss or single-elimination formats, sometimes a combination of the two (Swiss for enough rounds to theoretically have a single player at the top of the standings, then single-elimination quarterfinals).

The ideal number of players for a draft pod is eight. High-level official tournaments are always Swiss, and sometimes they will do pods of 7 if they have to, but they tend to have a lot of players to put into pods, making it quite likely that they can do groups of 7 and 8 only.

Store (and MTGO) drafts are always eight-person, and often single-elimination.

Our group doesn’t have the luxury of doing that, because our time slot is quite constrained, and we are unwilling to turn people awawy. So we have to deal with less-than-ideal numbers. Just how to do this has been the subject of a lot of debate over the years.

The minimum number of players for a draft pod is 6, the maximum 11 (and that only as a last resort, if we have 11 players total). Our attendance fluctuates from highs of 36 to lows of 15. Rules of thumb are to avoid 11-pods and maximize the number of 8-pods.

8-pods are easy: the Swiss pairing system, matches of best-of-three games, time limits of 50 minutes—all the same as the high-level tournaments.

That just leaves 6, 7, 9, 10, and 11.

Realistically, each best-of-three round will take one hour. Drafting itself will take between forty minutes and an hour. The prize-drafting that we do at the end takes at least fifteen minutes. So three best-of-three rounds plus drafting plus prize-drafting equals about four hours. We start at 19:30, and prefer to finish by 23:30.

Because prizes are affected by where players place, tiebreakers are quite important.

Because Swiss best-of-three is accepted as the standard, and is used in official tournaments, there is a strong push to use that format.

Because people have come to play and practice, rather than simply to win prizes, long downtime for any given player should be avoided.

So, with all that in mind… what to do with the non-8-pods?

6 is relatively easy: just do Swiss, and let the tiebreakers take care of it. It’s possible that no clear winner will emerge before tiebreakers, which is unfortunate, but nobody has to sit out, and it’s still best-of-three.

7 is tough because doing it “the official way”, using best-of-three, means a) a three players will have to sit out for one round each, and b) tiebreakers get messy with byes, providing a significant advantage for the players who get the byes. The other option is to do round-robin, best-of-one, where each round only takes 18 minutes but there’s only one game per round, 7 rounds. The downtime for each player is shortened to the length of a round, and is equally distributed among the players. But best-of-one takes away a certain amount of skill, as part of typical match play is adjusting your deck to your opponent’s after the first game.

9 is tough for the same reasons as 7, and in addition rarely finishes since it requires 9 rounds, leading to incomplete round-robin, where playing a subset of the other drafters can significantly alter the difficulty of a player’s games, while not using the “self-selecting” Swiss approach. Using Swiss for 9 players is also tricky, for the same reasons as 7, but additionally the limitation of only playing 3 rounds means that you may end up with two players with 3-0 records at the end, so the draft winner must be determined by tiebreakers.

10 has that same problem, but it seems as if 3-round Swiss rather than 9-round round-robin is a better approach, albeit one that we’ve only started using recently.

11 is a terrible number. We do round-robin (best-of-one, aspirationally 11 rounds) at the moment, but Swiss looks better and better.

Tiebreakers appear to be a problem here. A standard answer would be to use what the official bodies use, since they’ve presumably thought about the issues. Looking more closely at how tiebreakers work for a Swiss 7-pod will illuminate why this kind of thing is so tricky to get right.

3 players will get byes. The first bye goes to an effectively random player. The others go to players who have losing records. Our tiebreakers are the same as the official ones: the first tiebreaker is “opponent match percentage”, meaning that the better your three opponents do, the better your first tiebreaker. It’s essentially “strength of schedule”. So, after 3 rounds, assuming no draws, there will be 1 3-0 player, 3 2-1 players, and 3 1-2 players. One or two of the 2-1 players may have gotten a bye. The standard rules state that a bye simply doesn’t count at all towards opponent match percentage. A sample calculation:
Player 1 played a 2-1 player, the 3-0 player, and a 1-2 player. So their opponents went 6-3: 66% OMP.
Player 2 played the bye, a 2-1 player, and the 3-0 player. Their opponents went 5-1: 83% OMP.
It is almost impossible for any of the other 2-1 players to exceed the OMP of the player with the first-round bye if that player wins their second round. I can’t think of any scenarios where this happens.
But if the bye is treated like a 1-2 opponent for the purpose of determining OMP, the player with the bye, through no fault of their own, ends up at a disadvantage.

This is an ongoing debate, and I’m still trying to figure out the best answer. But it’s amazing to me how many issues crawl out of the woodwork with these things.

Leave a Reply