Saturday, January 24, 2009
Conference Plan: Conference Bid Allocation
What are your thoughts, ideas and concerns about the allocation of nationals bids to conferences based on the results of regional events? If you disagree with the mechanism described in the plan, how would you suggest conferences be awarded bids to national events? What do you think about how individual teams qualify to participate in Div I and II Regionals?
Subscribe to:
Post Comments (Atom)
To some degree, this discussion is a bit premature until he hammer down the number of D1, D2, and D3 nationals bids that will be available. That said:
ReplyDeleteI definitely agree that the D1 nationals bids should go to the conferences corresponding to the top X places at D1 regionals. D1 nationals bids are going to be considered extremely valuable, and it's likely that two teams from different conferences vying for the last D1 nationals bid at regionals are going to regard that game with almost as much intensity as an actual game to go. This makes the goals for teams very clear and is a good thing.
But for a variety of reasons, I do not support using the same approach for D2 and D3 bids. I have a few major reasons for this.
FIRST, these double-advance regionals formats will be a nightmare to run and could produce somewhat unfair results. I don't want to get into the nitty-gritty of tournament formats here, but suffice to say that running a 20 team format that is intended to produce two games to go at two different levels is difficult at times, and really really difficult at other times. The quality of these formats, either from a scheduling or a fairness perspective, will suffer from trying to create two games to go at two levels.
SECONDLY, the majority of the bids to D2 and D3 nationals are, by necessity, going to be "auto-bids". This is because every conference in the country automatically gets a bid to at least one national championship event. The majority of conferences in each region will be "mid-major" conferences. These conferences typically (not always, but typically) won't earn any D1 nationals bids, and they aren't eligible for D3 nationals bids. Because every conference automatically gets one nationals bid, this means every one of those midmajor conferences will be assigned one D2 nationals bid automatically. Assigning one auto-bid to D2 nationals to each of these conferences will soak up the majority of the available D2 nationals bids in the country, leaving a relatively small number of "surplus" D2 bids available in each region.
Given this large number of auto-bids, assigning the other D2 bids through high finish at D1 regionals, or top finish at D2 regionals, creates a false sense of precision. When we divvy up D1 nationals bids, we can say we are taking the best N teams in the region, and giving their conferences the bids. This is simply not the case for D2, and assigning the last few bids by placement won't do a particularly good job of rewarding the strongest conferences.
Furthermore, this ruins our efforts to create a "game to go" on the D2 cutoff. Consider: what if the winner of that "game to go" is the top finisher in a certain mid-major conference? Since that conference was going to get an autobid anyway, all they have done is replaced that "autobid" with an earned bid - which opens up another bid to be earned on the field. The next highest finisher actually gets this bid, so the loser of the game to go actually gets the bid. This means we didn't actually have a game to go.
(You could avoid that problem by simply saying that midmajor conference gets two D2 bids, but then you're actually rewarding the next-best team in that midmajor, which could be a really terrible team for all we know. This completely destroys the idea of distributing the D2 bids in a way that rewards the stronger conferences.)
This problem becomes even worse with respect to D3 nationals bids. In this case, the D3 teams could be literally anywhere in the format. Not only will we not have a reliable game to go - the small school teams on the border will frequently not even play one another. We'd be assigning the bids to small schools based on whether they faced the stronger teams earlier or later in the format, or similarly arbitrary reasons. If you're familiar with the issues caused by masters teams playing at open sectionals, then this would be like that, only much worse.
THIRD AND FINALLY, I don't really believe that "games to go" to earn an additional bid for your conference to D2 or D3 nationals (in the event that the format works out and we actually get them) would be all that dramatic or exciting anyway. No team is really going to be projecting themselves finishing at that borderline spot at their conference championship where they fall short of D1 but win the D2 spot.
Furthermore, even if they did expect that on some level, the level of drama still would not be that high. I don't think it's a knock on D2 or D3 nationals to say this. This is the reality in other sports as well. Sure, teams generally accept their NIT bids in NCAA men's basketball, and teams are glad to become bowl eligible in college football, but it's not as though they usually regard these things as major victories. D2 nationals is a great idea and something that there appears to be demand for, based on player feedback, but it's never going to inspire the sort of passion that D1 nationals will - particularly when that "game to go" at regionals isn't really a game to go, but a game to (probably) get your conference another bid. That's perfectly fine, and I think it gives us further motivation to use a different approach than leaning on placement at regionals.
---
My alternative is simple - make D1 regionals a straight "N advance" format where the top N teams earn their conference D1 nationals bids. Make D2 regionals a event like any other "tier 2" regular season event.
At the conclusion of regionals, most of the D2 and D3 nationals bids are going to be assigned automatically to the conferences that don't have any other nationals bids. The remaining bids should be handed out by algorithm. Basically, we take the same approach we used to distribute D1 nationals bids to the regions, and we use it to split D2 nationals bids among the conferences in that region. The only difference is that we get to use the data from regionals as well (and I would suggest we should weight those results more heavily.)
If it was a good enough approach for divvying up bids among the regions, it's good enough to decide the last few bids for the conferences as well. Most of the time, this will result in the major conferences each picking up some D2 bids, and most of the midmajor conferences having only one bid. But this is probably appropriate from a strength perspective.
My question is regarding the declining of bids to D2/D3 Nationals. Will the bids stay in-conference unless no team wants to go? Or will they be awarded by power ranking to the next-best team that doesn't have a bid to D2/D3 Nationals? The first option can and probably will dilute the strength of D2 Nationals more than the latter option.
ReplyDeleteThere also needs to be a deadline for declining bids so that teams on the bubble can know within a reasonable amount of time whether or not they will be playing at Nationals.
Nobody's really decided that yet. I do like the idea of every conference having representation at a national event. There's also some potential gamesmanship issues if we don't keep the bids within conference. That said, if the next team also turns it down, I'd be fine with throwing it into a pool and letting a major conference team get that bid (assuming they're the next strongest team).
ReplyDeleteYou're definitely right about the need for quick decisions between conference championships and nationals, so that teams can make plans and the field can be set.
I think D3 teams should have to opt out of the D1 structure before the post-season starts. As Tarr pointed out - having them in the mix at D1/2 Regionals really mucks up your ability to rank them intelligently, and can also mess up the ability to pick teams appropriately for D1/2 Nationals.
ReplyDeleteI know it adds a level of complication to have to put in D3 Sectionals and Regionals, but I think it makes for a better system.
There are no sectionals in the conference plan. There are regionals and conference championships in the same time frame as sectionals and regionals, currently. We don't have the calendar space for a third round of mandated postseason play in the northeast and Midwest.
ReplyDeleteAs far as D3 teams mucking up D1 regionals - this is only an issue if we are using placement at regionals to determine bid allocation for D2 or D3. Since I advocate using season-long ranking algorithm results, I don't see an issue with including those teams.
Something that I've considered is giving a bid directly to the team that wins Regionals. Then you avoid the case of two teams from the same conference playing a meaningless "finals" game at Regionals.
ReplyDeleteThat's a possibility, but I'd rather not do this, as it takes away from the year-to-year continuity of the conference championships. The hope is that, in the long run, conference championships can be a showcase event, and taking the best teams away hurts that.
ReplyDeleteI think that the regionals -> conference structure presents serious and probably intractable formatting challenges--as Adam points out above, it's difficult or impossible to design a format that places multiple games-to-go for the DI, DII and DIII bids. But that's not the worst part, by a long shot.
ReplyDeleteTeams at regionals aren't playing for themselves, they're playing for their conference; their incentive is to play to maximize the number of in-conference teams finishing above some threshold, rather than to have themselves finish as highly as possible.
This is likely to create some perverse incentives. For example, suppose that two teams from the same conference meet in the finals. It's in the best interests of *both teams* for the weaker team to win, so that the stronger team can compete on behalf of the conference for a Nationals bid in the 2v3 game. This sort of collusion could happen in any round of regionals--for example, in pool play, a pool with multiple teams from the same conference might let one of the weaker teams from the conference win the pool.
Teams may not be willing to collude in this way, especially in later rounds, since they probably won't agree on which of them is stronger, but they certainly will realize that it's in their best interests to rest their starters. Indeed, it's probably in teams' best interest not to play to win *any* intra-conference game at regionals. There are likely to be a lot of such games, with the top teams in each region comprising only 1 or 2 conferences, so equalizing the number of out-of-conference games becomes a major consideration in format design.
Now, maybe it's possible to come up with some radically new tournament format that avoids intra-conference matchups altogether. This might look something like the MARS format, where matchups are dynamically assigned after each round of play, or you might take a stab at creating a new formats manual, but I can't immediately see how to make that work, or how to deal with the immense number of different formats that would be needed.
@ David S: I very much disagree with your argument that teams should have to opt out of D1 to be D3-eligible. That is like saying that if Pixar wants Wall*E to have a shot at a Best Picture nomination, it should forswear the Best Animated Film category. Forcing the best small school teams have to decide, possibly even before the season starts, whether they want to give up one dream (making D1 Nationals) or the other (winning D3 Nationals), seems cruel.
ReplyDeleteIn other threads, people have brought up the fact that the NCAA demands a strict separation of the divisions. I would argue that that is completely different. For one thing, there are different rules for the different NCAA divisions (regarding scholarships). For another, even the best D2/D3 NCAA schools would have no shot at contending in (say) March Madness. This is not the case in ultimate, where we have small schools (and not just CUT) in the mix for nationals bids pretty regularly.