Winter League 2010 3-Division Analysis
Mar 8th, 2010 | By AndyBandit | Category: Features, winter league
Three Divisions. Did it work?
You clamored for it. And you know who you are. Last year, when people took a break from the usual Brian Calle chatter, it was all about moving Winter League to a 3-division format. I mean, maybe most of you never heard a peep about this, or even cared. But believe me, I heard about it a lot. Mostly from the mid-level teams who said they were tired of getting blown out by the top teams.
But I was really curious - is this just a handful of vocal players (you know, the squeaky wheel gets the grease thing) or are most players thinking this same thing? So I put out a poll (shocking), and I was really surprised to find out that the large majority of you were willing to move to three divisions.
My first thought was, don’t you want to play better teams, to help you improve? Beating someone you know you can beat does tend to simply validate bad habits (see hare/tortoise.) Plus, I thought everyone dreamed of getting to that next level? It’s a little unusual hearing people dream of moving to the next level down. Still, there’s something to be said for the thrill of an evenly-matched battle (notwithstanding 1960’s Vietnam.)
So the 3-division format was born. And after the first season under the new system, I bet you’re all dying to find out how it worked. (OR… you’re just reading this cause you’re bored at work and can only play so much on-line Scrabble.) One thing’s for damn sure. You all believe it worked. In the poll I put out last week, over 75% of you said the 3-division format worked in some capacity, and that’s higher than the percentage of people who voted for it in the first place back in December.
So let’s find out. How did W.L. 2010 compare to previous seasons?
I compared the results of all the scores from this year to the past 3 years, because that’s when we moved from 17-point games to 15-point games. And here’s something you might be shocked by. The average point differential in all the games played this year (excluding forfeits of course) was greater than in the past three.
I also looked at the percentage of “blow out” games. What counts as a blow-out is probably up for interpretation, so I used two measures that could arguably count. I think anyone would say that a 15-5 game was probably a blow-out. But some might even go further and say if you nearly doubled the other team’s score, that’s a blow-out too. So I used both scenarios. Again, the numbers show that this year the percentage of blow-out games was as high or higher than any of the past three. See table below. (Why do I even write that? Like you couldn’t see the table below.)
| Year | Division | Total games played | 7 point or higher wins | 10 point or higher wins | % 7 point wins | % 10 point wins | Total Point Diff. | Average Point Diff. |
| 2010 | A | 18 | 4 | 0 | 22% | 0% | 86 | 4.78 |
| B | 30 | 15 | 6 | 50% | 20% | 196 | 6.53 | |
| C | 23 | 14 | 6 | 61% | 26% | 150 | 6.52 | |
| Total | 71 | 33 | 12 | 46% | 17% | 432 | 6.08 | |
| 2009 | A | 34 | 16 | 3 | 47% | 9% | 204 | 6.00 |
| B | 40 | 16 | 4 | 40% | 10% | 218 | 5.45 | |
| Total | 74 | 32 | 7 | 43% | 9% | 422 | 5.70 | |
| 2008 | A | 31 | 5 | 2 | 16% | 6% | 140 | 4.52 |
| B | 21 | 10 | 7 | 48% | 33% | 142 | 6.76 | |
| Total | 52 | 15 | 9 | 29% | 17% | 282 | 5.42 | |
| 2007 | A | 36 | 16 | 2 | 44% | 6% | 187 | 5.19 |
| B | 38 | 19 | 7 | 50% | 18% | 223 | 5.87 | |
| Crossover | 4 | 2 | 1 | 50% | 25% | 19 | 4.75 | |
| Total | 78 | 37 | 10 | 47% | 13% | 429 | 5.50 |
What??? Wasn’t the 3-division format supposed to ensure closer games? And yet, the opposite is true. So how did this happen? Well, the problem clearly wasn’t in A Division. Avg. pt. diff. was down from last year, and the number of blow-out wins was as low as it’s been the past few years. That’s great, but from people I talked to, few players in A requested the change. So really it’s important to note what happened in B division (where the complainers reside). As you can see, there definitely were more blow-out games this year. In fact, half of the games played in B were decided by 7 points or more. And the avg. pt. diff. was over a full point higher than last year.
Here’s a possible theory. It could be that since there’s more teams in B than there are in A or C, there’s a greater difference between the highest seed and lowest seed. That could lead to more blow-outs. But… C Division, with fewer teams, had more blow-out games and the same avg. pt. diff. Another way to look at this theory is to look at the middle-seeded teams only. The 4, 5 & 6 seeds - Killer Chillers, Homer’s Disc and Caltech. They should have a lower number of blow-outs and closer games than all the higher and lower seeds. But their totals are pretty close to consistent with the overall average. 53% 7-pointers, 18% 10-pointers, and a 6.29 avg. pt. diff.
Another theory is the possible misplacing of some teams. If Ultimate Scorgasms for example is really a C-Division team, and The Zoo is really a B-Division team, and they were switched, that could make a difference. Consider that Ultimate Scorgasms had an 8.19 avg. pt. diff. and The Zoo had an 8.63 avg. pt. diff. Of course, unless we saw those two teams play eachother, there’s no way of truly knowing that a switch between the two would make any difference. For all we know, Ultimate Scorgasms would beat The Zoo head-to-head. (Yes, I had scheduled that game for Winnetka. Stupid rain, now we’ll never know.)
Also, Tarmac for example may have been more correctly seeded in A. Last year in A, half their games were 7 points or greater with an avg. pt. diff. of 6.17. This year in B, all but one of their games were 7 points or greater with an avg. pt. diff. of 8.00. But last year they weren’t the lowest seed in A, which makes a small difference. They would more benefit from being able to play only the bottom teams in A and the top teams in B, but unfortunately there was no great way to create that schedule.
*Note for stat geeks, when I mention an individual team’s avg. pt. diff., I’m adding the pt. diff. from each game regardless of whether the team in question won or lost, because that’s not relevant to the argument. Whether you won by 10 or lost by 10, it’s still a blow-out. So for example, if you play two games and you win one by 8 and lose the other by 6, your avg. pt. diff. is 7, not 2.
Now, one thing is fo shizzle, whether it’s a 3-division or a 2-division format, they’re both obvious improvements over the terrible terrible format we used to have when I ran Winter League. In 2004 and 2005 when I took over, and foolishly moved the league from A & B back to 2 equal divisions, it was just disastrous. The avg. pt. diff. for all teams was 8.33 and 8.98!!! In both years, nearly half of all games were 10-pt. or greater blow-outs! Over 2/3 of all games were 7-pt. or greater! Thank God we abandoned that nonsense. Too many rum and cokes for me back then.
So what’s my conclusion? Did the 3-division format work? For A Division, it looks like it did, although the trade-off was not getting to play as many teams, and it’s debatable whether that’s a good or bad thing. But if the purpose of the 3-division format was to create fewer blow out games for all teams, and make for closer scores, then I have to conclude that it did not achieve that objective.
However, having said that, I acknowledge perception can sometimes be more important than reality. Take, for example, witch doctors. Do not underestimate the power of the placebo! And based on that poll from last week, most of you really liked the 3-division format and think it worked. Perhaps it could be that all it needs is a little bit of tweaking, and then it’d be perfect. You know, like that guy that keeps treating you like his girlfriend, but he won’t call you his girlfriend, and you figure if you just keep having sex with him, eventually he’ll come around. But you know what? He’s never gonna come around. He just likes having sex with you. He’s only telling you what you wanna hear. And you’re just feeding into it. …….. Whoa, I got off track there. Sorry about that. But you know what I’m saying.
Anyway, that’s my two cents.
Loading ...
I’ll repeat what I said somewhere last week (Facebook?): The format is good because there are so many teams that in a two-division format, we wouldn’t play everyone and the tourney seeding would be too influenced by scheduling-chance. I think the reason the blow-out numbers were off were that there were a handful of teams put in the wrong places, which is completely understandable.
Look at the biggest margin of victory offenders:
Tarmac (+8) was a bottom-A team last year and pushed for the 3-division format. They got it and they got demoted and mowed down the B division. Maybe the squeeky wheel shouldn’t have gotten the oil?
Abominable Snowmen (+7.1) was a new team that got put into C and crushed most teams. Bad guess, but not really a placement error.
Ultimate Scoregasms (-7) torched the B-league last year and played (admirably) in A in the tourney. So naturally you should have assumed they were at least a B-league team this year. They weren’t. Should have been C, who know?
Slow Children (+6.3) was a mid-level B team last year that went nuts and won the B-tourney. We were clearly not an A-team and just put together a great run including a somewhat flukey 15-0 game that skewed the numbers way off. No placement-error there.
The Zoo (+6.2) was a middle-of-the-road B team that vanished in the tourney last year. They were a reasonable choice for demotion (though those of us who played them last year could have told you this was not a great call). Should have been B.
Those 5 teams combined for a 7.2 point differential in their games. The other 16 teams averaged a differential of 2.8, with the highest among them at 5.2.
So it’s not the league format; it’s the team placement. If Tarmac was in A, Zoo was in B, and Scoregasm was in C, the numbers would have come out far more evenly, but they (and the other two big-margin teams) were all where it seemed they should have been. Considering that one new stud player, one woman on vacation for a month, one injured star can sway scoring margins drastically and the organizers cannot really take that into account, I say the league was run very well.
hi!
2 division complainer here with a few of things to add:
-people play this sport for many different reasons. it shouldn’t be assumed that stepping up to the “next level” is a shared goal. sure a lot of people want to improve, but not everyone wants to eventually become a club player (most of A) or play at that level. ultimate is awesome but it’s not a priority in everyone’s lives and we don’t all have pipe dreams about it. i would say that most kids in this league simply want to have fun with their friends and get a good workout. get off your high horse you A level snob!
-i petitioned for 3 divisions assuming that the bottom 4 teams from A last year would be playing each other in B. given the standings from last year it would have been 6 teams in A, 7 in B, and 7 in C. BUT, two bottom A teams from last year (Pist and Jarnspikar) that beat Tarmac in previous years, disbanded, which threw things off. if they had stayed together and played in B (safe assumption) things would have been different.
-Ultimate Scorgasm, like Scott mentioned, was also an aberration. Tarmac was looking forward to play the top B team from last year but they had trouble even getting a team together this year. we were also looking forward to play the B tourney winners but i guess they had more important things to do than to show up to a make-up game (ahem).
-why were crossovers not possible this year? bottom A playing top B and bottom B playing top C would had made for less blow outs.
Actually, Erick’s question (”why were crossovers not possible this year? bottom A playing top B and bottom B playing top C would had made for less blow outs.”) makes me think of a possible way to incorproate this into league play:
What if, for example, the last 2 games of regular league play are scheduled for crossovers? So instead of playing the same team twice in A division, you rank the teams, and have A6 play B1 or something like that for the last couple weeks of regular play. It might be a logistical nightmare, but it might help to seed for Finals, and maybe even those could be considered pre-pre-1/4 Finals game or something…
I don’t get the benefit of “crossover” games. If I am in C, and I play C division teams all year, I should play for the C division tourney championship. Why would I play a B team?
Same argument in baseball. Why would the Dodgers play the Angels in a regular season game when they aren’t even playing in the same league?
It makes for interesting matchups sometimes, but games outside your league should matter in the standings and if it doesn’t matter in the standings, why play it? (Unless it’s all just for fun and we should all relax)
Oops, I meant “…games outside your league should NOT matter in the standings and if it doesn’t matter in the standings, why play it?”
unlike scott>>>> I LOVE the idea of cross over games! I would totally want to know if out “B” division team could beat the “A” league-ers. It would suck to lose to a lower division team, and I get that. (I will totally swallow my words if we get our asses handed to us). But don’t we want to know if the Divisions accurately depict how the league is and should be set up?
I played on the C league team snowmen this year. I would have liked the crossover game. In fact, I was looking forward to playing a B league team after we won the C division tournament. After our first game of the season in which we lost to the zoo, no team scored more than 8 on us. Wasnt even close. A game or two against a B league team would have been fun, even if it didnt count towards our record. For the record, I voted for the 3 div. format, but that was when i assumed our team would be in the B league.
First off, great work Andy! The depth of thought you put into this article is very very impressive and serves as a really good reference point to guide the discussion. Personally, I’ve always liked a weighted schedule without leagues. I think the games for the A bracket were fantastic this year, and on that level it worked out really well. But one of the things I miss is not meeting/playing against people I don’t already know. Perhaps that’s the job of summer league and St. Pats Hat.
The main point I’m hoping to convey is that if the goal is close matchups, brackets are beneficial for the very top and very bottom teams in WL and that more brackets creates greater inequity.
A misplaced team is a greater issue in a bracketed system, even more so the more brackets you have. It’s more likely to leave teams winless/ with very few wins (there’s a bottom in each bracket that someone must occupy). That carries over to the tournament as well. Take Tarmac/ Top shelf/ Milquetoast/Slow Children - Where would you place them come next year? The Zoo vs. Ultimate Scoregasm. These are all games that should happen. If we could perfectly seed all teams, brackets are great. But we can’t.
As it stands, I think the great advantage in the current system is for the bottom of the C league and the top of the A league. Top of A league (Empire, Formidable, Kong, Marshall) got to double up games against competitive teams at the expense of Milquetoast/top shelf getting to play more equitable games. Same is true in reverse for the bottom of the C pool.
But if those teams gain, who loses in having equitable games? Bottom of A (as explained above) Top of B (They don’t play their next closest matchup, bottom of A). Bottom of B (same as bottom of A explanation) and the Top of C.
The model I always liked was for a seven game season is no brackets, but weight the teams and strive for a schedule that gives
3 games that are a tossup,
1 games that you have a shot at, but you are underdogs
1 games you are favored
1 game against an elite team
1 game against a beginner team
(If you are one of the very top teams, you get 2-3 games that you will win without much effort, and if you are one of the weakest teams you get 2-3 games you probably can’t win no matter how well you play. It’s the price for more equity everywhere else.)
Now if you are at the top or bottom of that scale, it won’t work out quite like that for you. But for everyone else, it’s pretty good. The only problem is you have to weight the wins/losses for the tourney on strength of schedule
Lets look at a specific example as to how it illustrates equity in games. Here’s Tarmac’s schedule from 2009
Kong, Milquetoast, Pirates, Marshall, Jarnspikar, Homers disc, Pist, Top Shelf, Empire.
That’s 4-ish likely loses, 4 competitive games, 1 possible but tough win.
In 2010, the new brackets made Tarmac, Milquetoast, and Top shelf have less equitable games. If Top shelf and Milquetoast dropped down as well, I think it’s pretty clear they would run over most B teams.
Lets be clear about what these brackets are doing in practice.
Favors the top A and bottom C teams
Gives more shots at championships (something that was really unattainable for most players in WL unless you played at a pretty high level of frisbee … or have/had a legacy spot on Marshall or Damn)
Hurts equity of games for teams on the top of B and C, and the bottom of A and B
Incorrect ranking of a team has a dramatic effect. (Marshall in 2003 was put in B league, I don’t think anyone got within 7 points in any game in our bracket)
As to the desire not to get blown out by the top teams, that’s a personal preference. Perhaps these days are gone, but I fondly remember getting spanked by Quake (and I’m hoping they fondly remember our “big cocks” cheer), being amazed that someone could throw a disc THAT far and high on a pull, and playing against people that were wildly out of my league. I don’t expect everyone to like it, but I sure would have enjoyed a game against Slow Children this year or even one against the Zoo and meeting a few new people.
I agree with Deg, that the schedule should be dynamic and allow those teams on the cusp to play each other in the final week. The top 2 teams in each division play the bottom 2 teams in the next higher division. If the team from the lower division prevails, then they have the opportunity to chose to move into that division for the tournament if they dare. The rest of the teams would play amongst themselves within the division according to seed.
Look at the B standings. Tarmac in first. Slow Children 1/2 game back. Awesome Express 1 game back.
So suppose Tarmac and Slow, being the top 2 teams, both play an A-team in the final week and both lose. By Awesome Express plays a B team and wins. Suddenly, Awesome Express gets the B-Tourney 1 seed based on the fact that Tarmac and Slow lost to teams that were not in their league.
And if they don’t count in the standings and are just scrimmages, then why did we play league fees to play them? If we built a rain-delay week into the schedule for last-week make-ups, but there weren’t rain-outs to make up, then I think doing inter-league games would be fine. But I am in the B-league this year; my opponents should be also.
I feel that the 3 divisions worked.
There is a great discrepancy in the level of play between the top of A and the bottom of C, and while I agree with Seth, that it would be fun to play against new teams and meet new people, I do also agree with Erick, that not everyone is out there to become a club level player. Some of us play at the level that we do, may never get that much better than that, and are looking to have a fun, while still being competitive on the field. Lets save meeting new people for Summer League, St Pat’s Hat and pickup. We all assemble teams that we feel are going to play to the best of our abilities and for some teams, that’s just not at the same level as others.
In any other city, where people didn’t have to drive as far, and could change their schedules easily, making a flexible schedule works. That would allow for initial seeding through several weeks, and possibly reseeding for the last few (and that may work in the beach league, where everyone shows up at the same place and the same time any way). But it just doesn’t seem feasible in LA for this league.
I think the solution should be to keeping 3 divisions, but possibly asking for suggestion of which division teams should be in. Should Tarmac and Slow Children move to the A division next year? And do they even want to? Should Top Shelf and Milquetoast move down? And would they play as a team if they were asked to play in the B division. There’s not perfect answer.
Keep 3 divisions, maybe have help choosing which division teams play in and in the perfect world, where Ultimate is recognized as a sport that isn’t just for dogs, we’ll get a giant field in a dome and there will be free beer for everyone (except Andy, who will get free Rum and Cokes).
The key to acing any test is to not put down any incorrect answers. Similarly in ultimate, win all your games and you will end up #1 (unless we go to a BCS ranking).
I played on Tarmac in A division and in B division. And even though we had some close games last year in A it wasn’t very fun to know we were going to get spanked on the drive up. Like Seth said, last year I knew there were 3 games that might be fun/competitive and all the others we would lose (some more than others). But as a B+ team this year it was kind of boring too. I thought the B division would be tougher than it was and was surprised by how easily we breezed through some of our games. Honestly we are not that good and think that either Milquetoast or Top Shelf (both winless in A) would beat us handily if they brought the people listed on their roster, but who knows? As B league shows, you can’t always go by rosters or past years’ performances (see Ultimate Scorgasms). Compared to some of the other teams in A division, Marshall doesn’t have the most intimidating roster but they have played together for so long it makes them tough to beat. On top of having the skills they each know what the other is going to do and that is what separates them. No B division team I saw this year has that groove yet and maybe after a few years that will change – Slow Children seems to be getting there. Ultimate is a game of anticipation and timing, and not always raw skill and athleticism. Fact is, no matter how you split it, there are going to be teams on the cusp. Tarmac just happened to be the beneficiaries of a winter league designed specifically for our team to be successful because some of our players complained the loudest. I’m just as torn as anyone about what to do with a B+ team like Tarmac.
The regular season standings don’t matter that much anyway as long as the games are competitive. It is nice to be the top seed at the end of the regular season, but if you don’t win the tournament then who cares? I’ve seen Empire limp through a regular season before bringing their studs to dominate the tournament. I’m not bagging on them, but that is where they have set their priority. For others it is excelling in the B division regular season. Playing the crossover games based on regular season play is just a way to see how your team stacks up against A division teams. Maybe the sucky A division team wants to see how they stack up against a B league champ. I think it would be interesting for a team that wins a crossover to have the option to move up for the tournament (if they are bored of slaughtering everyone in their sucky division) and have more competitive games and potentially win the Grand Prize instead of the second-rate B championship. Teams would have to decide if getting 5th or 6th out of 6 in A, would be better than the B division title. Or they could just stomp the A team and remain in the B division for the tournament and see if they can shut everyone out. At least a team on the cusp controls their own destiny. If one or both top teams lose their crossover, then at worst they get bumped down to 3rd in their division’s regular season and in the tournament the two best B teams might meet in the semis instead of the finals. Big whoop. Happens all the time. You still have to win them all to win the tournament.
Another scenario could be to keep it as it is but set the tournament up so that everyone has a chance to win the whole thing - seeding based on divisional placement. Then the crossovers will happen naturally and the losers should fall out into brackets where they belong and at the end of the season there is reasonable certainty where your team belongs in the grand scheme.
I believe that our league commissioners do they best they can to provide us with the best product possible, but unfortunately you can’t create a specific formula to perfectly bound every season. And you definitely can’t make everyone happy every year. I liked the three divisions. It wasn’t perfect but it was pretty darn good (minus the make up games on a dark field that we paid for in addition to league fees that go to rainouts).
just a thought
I’m sure it will never happen
but a fun way to end all this shoulda/coulda/woulda’s with top c teams playing bottom b teams and top b’s playing bottom a’s
in honor of my favorite sporting event of the year
end of season tourny march madness style
single elimination (or double and make it more exciting)
Will Afternoon Delight Beat Marshall? Will Kansas lose to Cornell… no probably not
But could the Snowmen have a fighting chance against Slow Children
Sure why not, The University of Vermont beat Syracuse a few years back!
Make it even a two day affair,
sure the odds are against a c team winning it all, but the idea that awesome express could be this year’s cinderella and win it all just tickles me
pass out brackets, get vegas involved
A massive tournament with an equally massive consolation bracket would be pretty fun. Last teams standing from B and C could be crowed B and C champs even.
But with all this back-and-forth and different scenarios being thrown out/whined about, I do want to reitterate that the folks running the show have done a great job, especially considering they’re volunteering.
If you want to meet new people, or play against a good portion of Slow Kids, come out on Wednesday nights to La Canada high school. It’s not league but it’s still awesome. We’ll be out there tonight at 7:30pm.
On the subject. I like the 3 league system. I’m pretty sure everything that can be said has. See Scott Bergen’s post on March 8th, 2010 7:52 pm. I think that says it pretty well. The only thing missing is that it seemed this season, most of the teams we played only had 2 girls come out or only 9 players total. The lack of subs makes winning very difficult.
And Tarmac. We wanted to play that game too. Rally the troops. See you on Sunday.
We’ll try to come up with some cheer about Soul Plane or Snakes on a Plane.. or something.
couldn’t agree more
I appreciate anything and everything done by the winter league organizers, especially sarah for helping get the pick up team together that WON FOUR GAMES and finished third in C
back to the massive tournament
what if a C team outlasts all other B teams, are they b league champ?
And if B won it all, does that put A league to complete shame?
of the many fun and ridiculous scenarios this type of end of season tourny could produce
maybe it would help with division placement the following season
think of the endless heckling to that last place A team that manages to lose to the first place C team… it’s a stretch, but if they winter league tournies were run this way, EVENTUALLY it WOULD happen
That A team would have to dispand faster than the Marlins after a world series championship
Yeah definitely got to thank the league organizers! I know you put in a large amount of thankless work (except for, you know, the thanks we just gave you). I kind of like an idea that was put out there above, which was having a 2 day tournament, that was all inclusive. Make it pool play on saturday, and then go to brackets on sunday based on pool play. This way the teams fall where they may, and on saturday you get to play other teams. Itd be fun if a pool had an A team, 2 B’s, and a C. It would give everyone a chance to play another divisions team, and if a top B/C team was in a pool with a bottom A/B team, it would give them a chance to see what they really got, and a change to play in a higher bracket on sunday.
That would be fun JP, but it makes the regular season irrelevant. I think if we did something like this, it would have to be a huge tournament with every team, using the final divisional standings as seedings. A 1-7, then B 1-9, then C 1-7. Perhaps it would take two days to get through a 23-team single elimination tourney anyway, so it would be a nice weekend-festival type thing,
Here are my thoughts…I was disappointed in the end. Oh yeah, I should say that I am on First Phrisbeetarians. The brackets need to be more even. By even I mean, same amount in each bracket. In C we played each team twice, so by the time we got to the tourney we were playing teams for a third time. In B - some teams didn’t even face each other. I just wanted to play more teams - gets tiring playing the same teams over and over again. We ended up 3rd in C bracket behind Snowmen and Zoo, but it could have been due to 2 forfeits at the tournament.
Granted this was a building year for us, because last year we beat Barrel and other B brackets teams, but this year we just weren’t as strong. I have no issue playing in C - I had a great season and love my team. There are some amazing players and some of the nicest people in all of Ultimate on the team…I would play with them anywhere and everywhere. However, at tourney we had 2 teams forfeit to us. First one was Flick You, next was Chopped Liver. Granted some could have been the change of day - maybe they would have had more players the week before, I know we would have had at least 4 more people had the tourney been the week earlier. But I have to think that last game against Chopped Liver might have been played if we were new to each other and the game had more excitement to it. I think both teams, while we partied well together on the fields and had a blast, felt like we had been there and done that and the drive was lost to play each other again. Those are just my thoughts…
Thanks for listening…