September 13, 2013 by Charlie Eisenhood in Analysis with 44 comments
We already heard the initial reactions from readers to the new Club Championships format that guarantees every team a place in the elimination bracket. Now, more and more discussion is developing around how teams are going to attack the first day of the tournament in pool play.
Here is a great discussion thread from yesterday’s op-ed about the new Club Championships format.
Henry Thorne, USA Ultimate Board Member:
[quote]The [argument] I’m most bothered by is this notion that 4 pools of 4 means team won’t worry about pool play. Just the complete opposite of hundreds of tournaments I’ve participated in that followed that exact same, and completely common, format. I can’t remember a team ever deciding that the pool play games were irrelevant cause they absolutely are not. If you don’t hold seed you play far tougher opponents far sooner. It’s just completely wrong. [/quote]
Martin Aguilera, 2013 U23 Mixed Coach, long-time coach at Paideia HS in Atlanta:
[quote]Your argument in the other article seemed based on the idea that pool play (in that case power pool) games that just determined seeding didn’t always get full effort from teams. You cited examples from last year’s Club Championships to support your point. Why are the Thursday games this year any different than the games you described previously?
You are correct that if you drop seed on Thursday you will likely play a more difficult opponent. I doubt teams will tank Thursday games for that reason. I don’t think that teams will regard the games as irrelevant. But think about it this way: no one will be playing for their season on Thursday. No matter what your record on Thursday it will take 4 straight wins to get a national championship. Do you think that shift in mentality will not impact the subbing strategy of teams?
I can understand being bothered by hyperbolic comments that imply that teams will sit their best seven players all of Thursday. But if you get past that hyperbole I think you will see a valid point about the context of those games in the overall tournament.[/quote]
Thorne:
[quote]Martin, the reason the power pool F1F2 and E1E2 games did not matter is because they only affected where you ended up in the top four teams at the tournament. When there is lots of parity amongst those teams, that position does not matter. The reason pool position matters in Thursday’s pool play games this year is that failing to hold seed means you play teams an entire bracket of four higher which is a significant difference.[/quote]
Flo Pfender, member of WFDF’s Rules Committee, formerly on the UPA’s Rules Committee:
[quote]Henry, you are neglecting two points:
1. I would venture to say that all the experience from tournaments you are citing is from “regular” tournaments, in which playing as many good games as possible is as important (or even more) to most teams as winning the tournament. Saving legs in a game that you could have won with more effort makes no sense in this set up. This is completely different at Nationals. Here, at the event everyone trains for all year, all that matters to the top teams is the final placement. Teams will do anything that increases their chances to place well. Losing all games on Thursday and then beating a pool winner Friday morning puts you in the same place as winning all games on Thursday and then beating a pool loser—only with the first options, your legs will be significantly more fresh come quarters. Some top-heavy teams may be tempted to select the first option: With fresh legs, beating a semi-tired top team may be very possible.
2. You forget the ripple effect. If only a couple of the top teams save legs on Thursday (or maybe their top two players can only show up Thursday night…) and end up not winning their pools, pool rankings from Thursday will be so skewed that the advantage of winning your pool for the bracket diminishes. Once that is the case, saving legs really becomes the strategy to go for everyone, making pool games even less meaningful.
We’ll have to wait and see if this problem materializes, if not this year, maybe next. I am sure (more precisely: I know for a fact as I was involved in some early discussions on this year’s format) that USAU is aware of this risk, and they chose to go this way. Eliminating the 1v4 prequarters would have been another option with other downsides, and in the end they decided to use this one.[/quote]
Aguilera:
[quote]Thanks for chiming in, Flo. I think the ripple effect is something that concerns me a little more. Going into a bracket of 16 teams is fine (and I preferred it at college regionals in the 90s) but you need to get the seeding correct. This is especially true if overall final placement impacts playing opportunities in the following year. If Boost Mobile plays Florida United in the round of 16, they are all of a sudden a top 8 team and in the Pro Flight (no disrespect to either of those teams).
The ripple effect seems like it may negatively impact that seeding. Which is surprising in a year when (based on more common opponents and tournament participation) seeding should have gotten easier.
Good to know (what we all should assume) that these issues were on the minds of people at USAU when they were making those decisions. I think I like the no-1v4-prequarters option better, but it is impossible to tell until after this year. We will just have to wait and see.[/quote]
Jim Parinella, long-time Boston DoG player and author of “Ultimate Techniques and Tactics”:
[quote]I think the whole schedule is more exciting for the “fans” but worse for picking out the true best team because it’s going to be more random. I think it’s more likely than not that a top 4 team loses in the Round of 16 this year because of tanking, top-heavy rosters, inaccurate seeding, and the fact that the better team sometimes loses. There will be ties in pool play that in the power pool format had some opportunities to be worked out prior to elimination games but won’t here. And will they use the USAU ranking as the primary seeding criterion?
Forget “just win when it counts” or “by definition, the team that wins is the best team”. It’s simple math to show that having to play more elimination games possibly against tougher opponents makes it less likely that the true best team will win, that the two best teams make it to the finals, etc.[/quote]
Sean Childers, Ultiworld Statistics Editor:
[quote]This is true, though I’m a bit suspicious it would happen often from a true top 4 team (not that I have any idea who top 4 teams are this year, so maybe we should say from a normal-year top 4 team). It will almost certainly happen to a top 8 team within the first two or three years of the format. But this is reality in the NFL playoffs, NCAA basketball, and Champions League soccer; is it good or bad?
Lots of people are talking ripple effects. Originally I thought eliminating the 1v4 crossover would be a superior system because it would reduce the incentive to rest in pool play (who is going to risk 4th place?). The more I think about it, the less I’m worried that anything earth-shattering or ripple-down is likely to happen (except for the Parinella point above about sheer randomness increasing).
Start with the assumption that I highly doubt Revolver or Ironside or Doublewide (or Fury/Riot/Scandal) are going to sit players on Thursday. I don’t want to overgeneralize, but I suspect they’ll probably going to play this like they play most of their tournaments; maybe a bit tight on the offensive line, a bit more rotation-y on the defensive side but tightening there as the tournament goes on (Doublewide maybe tightens up the most on that side of the disc).
So the next tier of teams (GOAT, Machine, Sockeye, Ring, PoNY?) are maybe wondering what is a rational response to that action that would maximize our chances of knocking one of those teams off. It seems like your first move is to ensure you get 2nd or, at worst, 3rd in your pool. So you’re going to play your top guys on Thursday. But you’d rather not run them into the ground.
Two ways I see a deviation: If a second tier team (say, GOAT) lost one or two early pool play match ups to a bottom-tier team (say, Florida United). They *might* rather take 4th seed with rested legs than a 50% 3rd/50% 4th place outcome with tired legs. The effect would be minimal. A bottom-tier team could say that their best chance of winning nationals would be to just toss everything into bracket play and rest the studs during the pool. I don’t think many teams would do this, because your top players are risking getting only one game out of the tournament. And even if they did try this, it would require that team to upset a 1 seed in 1v4 play which I personally think is unlikely.[/quote]
Parinella:
[quote]”it would require that team to upset a 1 seed in 1v4 play which I personally think is unlikely.”
You’re the stats guy, run some numbers. Using a Pythagorean exponent of 4 and an expected final score of 15-11, I come up with a 22% chance of the underdog winning that game. If you make the exponent 6 and the expected score 15-9, it drops to a 4.5% chance. What are the right numbers?
I guess I agree there won’t be systematic tanking going on, but the combination of other factors will make some really interesting R16 matchups.
a) Seeding. Seeding traditionally uses wins and losses while rankings use point differentials. This year’s results seem more mixed than before and some surprising teams are very high (Machine #2 and GOAT #4 in Men’s, AMP #1 in Mixed) and others low (Ironside #8). Where will they be seeded? I can easily see a team that is truly maybe #5 getting seeded around #11 (and ending up in 4th in their pool).
b) top heavy/high variance teams. GOAT is reported to be one and would be especially dangerous in any particular matchup, even if their odds of winning it all might be vanishingly low. In a power-pool format, they would have had to give it all they got just to make it to quarters and so wouldn’t have any reserves left.
c) tie-breakers. Power pools offered a chance for “correcting” tie-breakers. Finish behind someone in a tie-breaker and you can still win the Friday games to improve your standing (or they might lose). (Not everyone, though. A team at 1-2 and in fourth would not control their destiny on Friday.)[/quote]
Childers:
[quote]A) Seeding — Indeed, it will be interested to see how they do it. I don’t think GOAT at #4 is all that off but Machine at #2 probably is.
B) I agree there are a few of these teams but Doublewide won last year and maybe tightens up the most at the top (they had multiple guys move to D Line in the pro flight finale). I’m not sure where to go with this.
On the pythagorean front, I think 4 is definitely too low and my guess is that we are instead looking at something between 6-11. Plus how do you account for the fact that, unlike other games, the team’s entire roster is present and the favorite will be willing to play only it’s top players if necessary? 2011 Revolver is a type of “True 1” seed, in my book. Look back to their nationals results: 15-8 Condors, 15-7 Machine, 15-7 Southpaw. Someone elsewhere on the internet did an ultimate post and thought the exponent in pro ultimate was 6.9. Using 15-8 and an exponent of 7 I come up with a 1% chance but I’m not even sure that considers the odds that a true 1 seed has the option to start playing more studs when they go down. I was also reading a bit more this week here.[/quote]
How do you see the format affecting Thursday’s strategy? Chime in in the comments.
