Tuesday, May 8

Goddamned formats

After reading all of the talk on rsd (again), tarrblog, and more...

The rematches still don't bother me.

What does bother me (as I briefly posted a while back) is the sheer number of games played in a given day. Toss aside the "two games and go home" argument. As Colin writes, you can always schedule more consolation games. The larger issue, exacerbated in a 14 team format, is playing more than 3 games per day.

By making Regionals an endurance test, are we actually selecting the best teams to compete at Nationals? (Necessary disclaimer: The Formats Committee put in tons of hard work and I love y'all for it)

I've looked through the format manual (which is surprisingly interesting, I might add... wait... I mean... I'm not an ultimate nerd, I swear!) and even just looking at the 16 team 1-4 advance formats, some ugly situations can occur. 8 games in 2 days (Most Pool play formats, and on occasion in brackets)? That's so much ultimate! In the 16.2 Pool format, one team in the game-to-go can be on their 2nd game of the day and 6th game of the weekend, while their opponent can be on their 4th game of the day and 8th of the weekend? Just because the original seeding was messed up? Note that this would also be that team's fifth consecutive elimination game as opposed to the other team's first.

Similarly, other formats can be similarly back-breaking in terms of endurance, but are often more considerate in terms of the number of games played (specifically on the same day) by two teams in the game to go.

After looking through the formats and comparing them with the Nationals format (for Club) I feel that there are a couple things that come to mind:

  • Rarely, if ever, have teams play 4 games on Sunday. If teams must play 4 games, make it on Saturday if at all possible.
  • If a team must play 4 games on Sunday, they should be playing against a team that had also played 4 games on Sunday.
  • Never have teams with a difference of 2 games played in one day play each other. (We can also call this "For the love of God, stop using the 16.2 Pool Play format.")
  • Even the regional winner should not have just one game on Sunday. In reality, there are only two teams (in each division) who will play at a tournament where there is one game on a day. Why would we want to mimic this at Regionals? Shouldn't we work to select teams that can play 3 games on day one and 3 on day two?
Complicating thoughts:
  • That damn three way tie for first. If we could determine a way to adequately (see this episode of Newsradio [also, part one and part two] for a hilarious extended rumination on adequacy) resolve this, we could get make life much easier.
  • What about the three-way tie for last? Why do we eliminate the bottom of those three teams, but not the bottom team in a three way tie for first? Sure, in the tie for first scenario, a team may have only lost to the best team at the tournament, but in the tie for last scenario, two teams get to move on but another does not. If we feel comfortable with this elimination, why do we not feel comfortable with the 3-way tie for first elimination? The point is that after 3 rounds of play in a bracket (16.2 and 16.3. 16.1 has no bracket and 16.4 is a whole different animal), you arrive at 8 teams. After 3 rounds of pool play (16.2 and 16.3), you're down to 12 because we want to avoid the 3-way tie. Why not use the more efficient format? Why pack in 4 games?
  • More games for teams that lose early. If they want more games, provide consolation games. If they want to delude themselves into thinking they can advance, why should we lead them on? I mean, if you lose two games on the first day, there is a very good chance that you should not advance.
  • The argument that came up in the aforementioned rsd thread about the relative strength of teams at Nationals vs Regionals seems like a load of crap to me. We should not base our formats on the expected relative strength of teams. Just because we expect there to be a bigger discrepancy between 1 and 16 at Regionals than at Nationals does not mean that we should play more games at Regionals.
  • If we want to make the formats independent of seeding, why do we bother seeding the teams? Why not use a random draw? Ideally, and I think everyone agrees, there would be a "regular season" of some kind that would establish seeding (this would simplify the formats as well), but there is no regular season in ultimate. So why bother seeding? Why are we trying to assign seeds that the formats are designed to make irrelevant? Either the seeds are relevant and the formats are not properly designed, or the seeds are irrelevant and it is pointless to assign them. Seems unnecessary somehow. Also, we could get rid of inevitable discussions like "Team X beat Team Y in January in an unreported scrimmage" or "Team A beat Team B but team B was missing 72 vital players" or "You can't count that tournament because we played under a foot of snow."
Anyway, I think that's it for now.
Workout Total:
10 min core
15 mintues wobble board
10 minutes shoulder rehab
15 minutes stretching


Vancer said...

I agree a great deal about the idea of not seeding teams. However, I would make one adjustment. I think that the World Cup does it right. They come up with a group of 1 seeds for each pool, and then fill in underneath. There are a lot of similarities with the lack of a regular season, other random tournaments, and an "interesting" ranking system.

This would prevent the few strongest teams (which are normally known) from ending up in the same bracket, while still giving a random draw concept. One seeded team per pool would work well, so 4 seeded, 12 random for a 16 team draw.

Ryan Vance
Medicine Men #11 (continued Medicine Men reader base)

dusty.rhodes said...

I was actually just reading about the way that the WC does this. I had erroneously assumed that it was a completely random draw.

Do you have any info on how they determine who is, esssentially, the last "seeded" team? Do you know if there have been any large blowups concerning who was left out of the top seeds? I love following soccer, but this I don't follow it as closely as I might like. I'll likely try to ferret out this info on my own, but I don't know how successful that will be.

Perhaps that could be the prize for winning your section? I know that we range from 3-5 sections per region, and taht in some sections you've got more than one top team, but that could provide a little more emphasis on winning at sectionals instead of just advancing. Or maybe not-- the seeding arguments would be simplified to "Who are the top 4?" regardless.

Vancer said...

As far as I know, there has not been a major blowup. The biggest controvery in the last 3 or 4 was actually the last one, when the US was not seeded despite having a better ranking than a team that was. The seedings are done MAINLY by the FIFA ranking system. I believe that this is done with an algorithim, like the RRI.

With a 16 team, 3 section region, I like the sectional champs seeded, plus one at the RC's discretion (must be a 2nd place finisher). The 4th would obviously cause a commotion, but then the commotion is over one seed, not 15. With the 2nd place requirement, the eligible teams last year would have been Medicine Men, Pittsburgh, and X from the NC section. Since they beat us at Chicago 3 weeks before, Pittsburgh would have clearly been the 4th.

Again, not perfect, but I do like the idea.

dusty.rhodes said...

Hi-Rates were the second NC team. Burgh ended up seeded 4th last year . That being said, last year featured the clusterfuck 14 team format that included wonderful experiences such as:

-- 5 games on the first day for 12 teams
-- Potential for 5 games on Sunday for 2 teams
-- 6v7 seeds at 6:45am on Sun
-- Potential for A7 to play ultimate for 11 out of 12 hours on Sunday
-- 6 seeds starting Sunday with a game a mere 12 hours and 10 minutes after their last round on Saturday finished
-- 9 games (4 must wins on Sunday) for the bottom team in the game to go vs 7 games (2 games and 0 must-wins on Sunday) for the team that lost in Finals
-- Med Men and Warriors get eliminated on a Tie-Break
-- William and Mary bagging day two despite having 4 games scheduled

That's all that comes to mind immediately, but that was a horrible format exacerbated by teams not playing consistently from Sectionals to Regionals. Not that I've hunkered down to look for a solution, but still...

Disclaimer: I feel sour about that tournament, but we knew the format and lost games that would have put us in a better position to qualify. I do not think, however, that we fully *understood* the format and the ramifications of not finishing first in our pool.

Mackey said...

I agree that ultimate generally necessitates too many games. To say nothing of series tournaments, my sophomore year I had the distinct pleasure of playing in two over-leaded tourneys--College Vegas, which in '06 had a 4/4/2 schedule (they changed to to a more palatable 3/3/3 this year, i think), and Centex '06, which had a ridiculous 5 games to 15! on the first day of the tournament. I think we managed to pull off a 2-4 record that day simply because our numbers lessened more slowly than our opponents'.

I think the longevity vs. 1-game strength question is a valid one, though. To draw upon NE regionals, for example, I think my team (Dartmouth) was better equipped than any team at the tournament in terms of depth, save perhaps Brown--however, because of the discrepancy between the winner's and backdoor bracket's number of games in 16.2, like you said, a slightly-depleted but still strong Dartmouth had to deal with a rested Williams on Sunday. While I won't say it was a deciding factor, they definitely made more plays on the 50/50 discs than we did.

I agree with you that the long-term solution is, I think, a regular season, since that will allow formats to avoid seed speculation.

It's a tough matter to deal with in any way that'll be satisfactory for all, I think. Definitely credit the formats committee for doing the best they could (which is better than most).

I spent a good hour or two reading the formats manual when I first started coordinating club sectionals here. It gives you a lot of appreciation for what makes a good format, and gives you a great grounding for criticizing the lazy/inexplicable formats you sometimes see at non-series tourneys. They really account for everything.

dusty.rhodes said...

I should note that I don't give a damn about non-series tourney formats. Since we have no regular season, those tournaments are all about putting everybody in the game and letting them learn.

You're lucky you never went to a Mike G tourney (or have you?) I'm pretty sure we played 5 games to 15 on two consecutive days at one of those fuckers. With 10 dudes for the first game. 8 for the rest of the weekend. 7 for the last game in which we lost to Dartmouth B. 11-10 or something. Yech.

Bill Mill said...

Let's start a Metro-East regionals league. Sunday mornings, home-and-away series for the top (4? 6? 8?) teams from last year's regionals. Seeding regionals will be a piece of cake.

How sweet would it be to play one game in a day? Just once, I'd like to try it.

gapoole said...

I played high school Ultimate for four years in NJ, so I got a good taste for the general pros and cons of both tournaments and individual games. I find that tournaments are a far superior experience; a single game is often unsatisfying, and it's harder for your team to get in the groove. Some tournaments were rough, especially with a small squad, but you learn a lot more from 6 games in two days than you do in 6 games over the course of six weeks. I think single games should be limited to the tune-up variety.

Tarr said...

Very interesting discussion here, for sure. When making the new manual a few years back, the formats committee basically looked at existing formats/tournaments and decided we would look to stay within the established constraints on number of games (4/4 Sat/Sun max, with occasional exceptions for 5/4 with shortened pool play games) but not worry too much about it other than that.

I know how exhausting four straight elimination games to 15 can be. The issue is, do we trade off the robustness of the format for fewer games? There are plenty of ways you can abbreviate things, but you end up relying on initial seeding more, and/or having less insurance that you can cover for three-way ties in pool play.

I am increasingly sour on the (increasingly popular) 16 team, 2 advance pool play format. The reason being, as Dusty pointed out, the potential 2-game discrepancy in the final game-to-go. From a technical perspective, it's hard to argue the extra game significantly improves the format as compared to the 16 team, 2 advance bracket format.

I would love to be able to take sectionals and farm it out over several weekends of formalized league play, as Bill suggests. 3 games per day, maximum, three or four weekends to work with... ah, it would be great. Of course, writing formats for that would be a whole 'nother manual.

dusty.rhodes said...

Thanks for the comments.

Sectionals AND Regionals could be spread over two weekends each, really. 3 games per day for two weekends. Regionals could actually just follow a modified Nationals Format (given that the Nationals format is actually an acceptable format in the first place) and sectionals could actually find a way to deal with a tournament nearing 30 teams.

I would argue that the three-way tie is less of an evil than 4 games in a day. After looking at the manual for a while, it was clear that the reason for 4-5 games in a day was nearly always caused by that 3-way tie for first. Could there be a way to conditionally alter the tournament based on 3-way tie or lack thereof? Or would that possibly affect the "gamesmanship argument" by having teams conspire to force a 3-way tie in order to force more games on teh whole field? It seems more unlikely at first glance, but I don't have the time to give it a second glance at the moment.

Essentially, we don't always have a three-way tie for first, so why should we play a format that corrects for a three way tie for first?