Thursday, June 14, 2007
"Tournament stratification" sounds like a good title...
I'm doing requests, it seems. I mean really, when he wrote that, he may as well have put it as "Tarr should blog about tournament stratification". I guess he could have been thinking of George or maybe Dusty, but this is right in my wheelhouse. I have 3-4 other topics I've been meaning to blog, but they can wait.
We're talking about tournaments that have "elite" and "lower" devisions. Sometimes these divisions cross into one another; sometimes they do not. Let's lead right out with the main pros and cons of formats with power pools like these:
Pro: teams get more meaningful games, which is the point of playing.
Con: fundamentally not even-handed. Relies more on initial seeding.
That pro is crucial, and is enough to justify power pools a lot of the time. The bottom line on power pools is that the elite club teams are not going to pay airfare to travel to a tournament unless they know they are getting a full slate of games against other elite teams. So having, at least, a power pool (or two or four) on the top is basically a requirement of any top tournament.
The first con, i.e. lack of even-handedness, is not a major concern. You night have a harder path to win the tournament, or only be able to place 9th, but nobody's season is ending as a result. The point of preseason tournaments is to play, and teams should be judged by individual game results as oppose to the number attached to their finish.
The second con, however, is more significant than you might realize. Allow me a meandering, self-indulgent anecdote to demonstrate my point:
My new job has a sand volleyball court outside, and there is a formal lunchtime volleyball league. Rather than attempt to make balanced teams and a full schedule, the league is split into 8-person divisions, and each league game is a different 4v4 matchup within that division. There are 56 players, and consequently 7 divisions. After describing myself as "tall and reasonably athletic, but with very little V-ball experience", I got placed in div 5.
This week, one of my teammates couldn't make the game, so a player from another division subbed in. He was decent; I thought he was about average among the players on the court. I figured he was a good 6 or a bad 4, but when I looked him up on the league spreadsheet later, I was quite surprised to discover that he was div 2.
What was the point of that story that applies to Ultimate? Most people drastically overrate their ability to predict the relative strength of teams/players. In the case of the V-ball league, the idea to have divisions is a very good one. Div 1 has a guy who can two-hand dunk with ease and a former varsity volleyball player at Stanford. But with 7 divisions, the variation within a division is far greater than the variation between divisions. The league would be well served by 3 or 4 divisions. The rest is just noise.
The same is absolutely true in Ultimate. The only arguable exception is the few elite men's teams, which are pretty stable. As I said, you need to have a power pool at the top anyway, to attract the top teams. But even there, the worst "elite" teams usually have very little (if any) edge over the best "open" teams.
In conclusion, some stratification in preseason tournaments is a good thing. But there should be some room for crossover whenever possible, and care should be taken to not overstratify. It's easy to talk yourself into believing that there are several big obvious breaks in team strength, but the results will often defy such expectations.
We're talking about tournaments that have "elite" and "lower" devisions. Sometimes these divisions cross into one another; sometimes they do not. Let's lead right out with the main pros and cons of formats with power pools like these:
Pro: teams get more meaningful games, which is the point of playing.
Con: fundamentally not even-handed. Relies more on initial seeding.
That pro is crucial, and is enough to justify power pools a lot of the time. The bottom line on power pools is that the elite club teams are not going to pay airfare to travel to a tournament unless they know they are getting a full slate of games against other elite teams. So having, at least, a power pool (or two or four) on the top is basically a requirement of any top tournament.
The first con, i.e. lack of even-handedness, is not a major concern. You night have a harder path to win the tournament, or only be able to place 9th, but nobody's season is ending as a result. The point of preseason tournaments is to play, and teams should be judged by individual game results as oppose to the number attached to their finish.
The second con, however, is more significant than you might realize. Allow me a meandering, self-indulgent anecdote to demonstrate my point:
My new job has a sand volleyball court outside, and there is a formal lunchtime volleyball league. Rather than attempt to make balanced teams and a full schedule, the league is split into 8-person divisions, and each league game is a different 4v4 matchup within that division. There are 56 players, and consequently 7 divisions. After describing myself as "tall and reasonably athletic, but with very little V-ball experience", I got placed in div 5.
This week, one of my teammates couldn't make the game, so a player from another division subbed in. He was decent; I thought he was about average among the players on the court. I figured he was a good 6 or a bad 4, but when I looked him up on the league spreadsheet later, I was quite surprised to discover that he was div 2.
What was the point of that story that applies to Ultimate? Most people drastically overrate their ability to predict the relative strength of teams/players. In the case of the V-ball league, the idea to have divisions is a very good one. Div 1 has a guy who can two-hand dunk with ease and a former varsity volleyball player at Stanford. But with 7 divisions, the variation within a division is far greater than the variation between divisions. The league would be well served by 3 or 4 divisions. The rest is just noise.
The same is absolutely true in Ultimate. The only arguable exception is the few elite men's teams, which are pretty stable. As I said, you need to have a power pool at the top anyway, to attract the top teams. But even there, the worst "elite" teams usually have very little (if any) edge over the best "open" teams.
In conclusion, some stratification in preseason tournaments is a good thing. But there should be some room for crossover whenever possible, and care should be taken to not overstratify. It's easy to talk yourself into believing that there are several big obvious breaks in team strength, but the results will often defy such expectations.
# posted by Tarr : 6/14/2007 07:32:00 PM
Comments:
<< Home
The topic came up a couple weeks ago at the NJHSU state championship. The competition basically ended after the top 4 teams, 6 if you were generous. I realize that a state tournament is supposed to give the most fair path to all teams, but in reality there were 3 B teams and at least 2 others in their first year ever. I thought it would have been better overall if it was split into two divisions, giving more meaningful games to all the teams involved. Of course, that wasn't an option. The hope is that next year they will be able to have more regular-season games/tournaments.
What I was really getting at was that "meaningful" doesn't just mean "both teams have more than a 10% chance at winning". Teams below the cutoff like to get their shots at the big boys even if they know they are going to lose. But the big boys don't like to play those teams, even once, and like you say, if you give them too many games like that, they just won't bother showing up. The correct decision depends on how important it is to the teams and to the tournament organizer (which could be the UPA, as in the series) to have/avoid those games.
Prior to Boston Invite, some complained that teams 17-24 did not get the opportunity to play against 1-8 at all. It wasn't that they thought they might win, but that they could make the quarters and get to play a 1-4 team.
Another issue is whether to apply stratification to create a whole new structure like European soccer, with relegation and promotion. One problem with this is that teams come and go in ultimate. Say there was such a system in place now. Who would own the rights to the other Boston team? What if instead of merging, Sockeye split into two, not unlike what happened to Cojones (NY) in 1996, two years after the NYNY dynasty crumbled? And would we see players deserting teams which get relegated, seeking the nearest higher-level conglomerate?
Looking at my stat book, I see that we went to two stratified tournaments in 1992, Boulder and Cuervo, while playing in 4 non-stratified tournaments (two with split squads) plus 5 UPA series events (spring Reg, Easterns, Sect, fall Reg, Nats). In 2006, we went to 1 non-stratified, 4 stratified (or Elite only), and 3 UPA events. In '05, there were no non-stratified tournaments (even White Mountain Open had Elite and Open).
There is an interrelationship between tournament stratification, attendance, earlier Nationals, and roster size/conglomeration, with a bit of chicken/egg(/frying pan?)ness. In the '80s, teams used to practice during the week and go to tournaments almost every weekend, in both spring and fall. Rosters were smaller, and there were more teams on about the same level, so it made more sense to go to tournaments. Now, with a roster of 24 and fewer comparable opponents, it doesn't make sense for an elite team to go to a non-stratified tournament in the fall or late summer, since most of the players will stand around most of the time instead of getting better. Subtract out 4 weeks from the fall (Nationals was originally Thanksgiving weekend, and still in the middle of November into the '90s), and you really can't afford to "waste" a weekend going to a tournament where half or more of your games won't be closer than 15-5. And if you only go to stratified tournaments, there are fewer opportunities to play.
I like the idea of soccer-style division play, but worry that the looseness of rosters would make it unfeasible. I'm ok with the current setup where the Series is not stratified, although I would favor giving exemptions past local qualifying for teams that do well the previous year (say, Nats qualifiers get byes to Regionals, champs get bye to Nationals or super-Regionals)
Prior to Boston Invite, some complained that teams 17-24 did not get the opportunity to play against 1-8 at all. It wasn't that they thought they might win, but that they could make the quarters and get to play a 1-4 team.
Another issue is whether to apply stratification to create a whole new structure like European soccer, with relegation and promotion. One problem with this is that teams come and go in ultimate. Say there was such a system in place now. Who would own the rights to the other Boston team? What if instead of merging, Sockeye split into two, not unlike what happened to Cojones (NY) in 1996, two years after the NYNY dynasty crumbled? And would we see players deserting teams which get relegated, seeking the nearest higher-level conglomerate?
Looking at my stat book, I see that we went to two stratified tournaments in 1992, Boulder and Cuervo, while playing in 4 non-stratified tournaments (two with split squads) plus 5 UPA series events (spring Reg, Easterns, Sect, fall Reg, Nats). In 2006, we went to 1 non-stratified, 4 stratified (or Elite only), and 3 UPA events. In '05, there were no non-stratified tournaments (even White Mountain Open had Elite and Open).
There is an interrelationship between tournament stratification, attendance, earlier Nationals, and roster size/conglomeration, with a bit of chicken/egg(/frying pan?)ness. In the '80s, teams used to practice during the week and go to tournaments almost every weekend, in both spring and fall. Rosters were smaller, and there were more teams on about the same level, so it made more sense to go to tournaments. Now, with a roster of 24 and fewer comparable opponents, it doesn't make sense for an elite team to go to a non-stratified tournament in the fall or late summer, since most of the players will stand around most of the time instead of getting better. Subtract out 4 weeks from the fall (Nationals was originally Thanksgiving weekend, and still in the middle of November into the '90s), and you really can't afford to "waste" a weekend going to a tournament where half or more of your games won't be closer than 15-5. And if you only go to stratified tournaments, there are fewer opportunities to play.
I like the idea of soccer-style division play, but worry that the looseness of rosters would make it unfeasible. I'm ok with the current setup where the Series is not stratified, although I would favor giving exemptions past local qualifying for teams that do well the previous year (say, Nats qualifiers get byes to Regionals, champs get bye to Nationals or super-Regionals)
Lots of good points.
I don't think organized soccer-style division play in club is really needed. The market is already setting it it up pretty well. The only issue, really, is sectionals. I would support some sort of system where club teams get a pass to regionals if, say, 70% of the roster was on a nationals team in the same division the year before. Or some other measure of eliteness.
The only place the relegation/promotion system might be needed is college open, where:
1) Roster-hopping is not an issue
2) The number of teams in some areas is making sectionals unwieldy, and making regional qualifying too hard
3) The weather and academic schedules conspire to make it hard to expand the series in either direction.
Personally, I think we should implement subregionals for large regions, but a lot of other people think that the time pressure doesn't permit it. In that case, we should go with a relegation/qualification system, where strong finishers in the lower division can enter Div A college sectionals the next year.
Post a Comment
I don't think organized soccer-style division play in club is really needed. The market is already setting it it up pretty well. The only issue, really, is sectionals. I would support some sort of system where club teams get a pass to regionals if, say, 70% of the roster was on a nationals team in the same division the year before. Or some other measure of eliteness.
The only place the relegation/promotion system might be needed is college open, where:
1) Roster-hopping is not an issue
2) The number of teams in some areas is making sectionals unwieldy, and making regional qualifying too hard
3) The weather and academic schedules conspire to make it hard to expand the series in either direction.
Personally, I think we should implement subregionals for large regions, but a lot of other people think that the time pressure doesn't permit it. In that case, we should go with a relegation/qualification system, where strong finishers in the lower division can enter Div A college sectionals the next year.
<< Home
