WFTDA has released its members' update to the ruleset for 2014. Get all the details here.
login | register
Enter your Derby News Network username.
Enter the password that accompanies your username.

you forgot it?!?

F***ing Rankings and Divisions. How Do They Work?

  • S-curve Seeding for playoffs.
  • Bout Weightings at playoffs.

Yesterday, at about the same time as we published the results of the DNN 2012 Reader Poll, WFTDA released more information on how ranking and divisioning is going to work in 2013 and onwards. As well as making both our site and theirs fall over, that information dump has left a lot of people rather confused, so we have both reached out to WFTDA sources and gone through the documents ourselves in an effort to clear things up for you.

"Divisions are set once annually and do not change, regardless of rankings. Divisions will be reset each year based on the rankings results of the previous Competitive Season."

It has been assumed by many people that this means that the teams in Division 1 at the start of the year will all be going to playoffs at the end of the year. This is NOT the case. A team’s division during the regular season is useful information for the purposes of working out who needs to play who now that in-region play requirements have been done away with; it does NOT reflect which division’s playoff tournament any given team will be going to.

Rather than needing to get in two in-region games, teams need to get a certain number of in-division games in. If you qualify for Division 1 playoffs one year, you need to get at least three games in against Division 1 opponents, and a fourth game that can be against a Division 1 or Division 2 team. Those in Division 2 need two games against Division 1 or 2 teams, and one that can be against a team from Division 1, 2 or 3. Division 3 teams only need to get two games in, and these can be against teams from any division.

"Teams are assigned to Playoffs based purely on ranking, using S‐curve seeding."

The 40 teams going to the division one playoffs in 2013 will be the 40 highest-ranked WFTDA leagues as of June 30, 2013--assuming all those teams have played the required numbers of games.

Those 40 teams will then count as Division 1 until June 30, 2014--which means teams can schedule borderline teams with confidence, knowing that for the next twelve months, even if a team is ranked #40, they will still count as a Division 1 opponent for that year, even if they have a poor run of form and are ranked outside of the top 40 when the game actually takes place.

What is this ‘S-Curve seeding’? It is the way teams are distributed. How does it work? You start with a long list of the top 40 teams, and take turns placing each team into a playoff--so the teams ranked 1, 2, 3 and 4 go into the four division one playoffs as the top seeds; those ranked 5, 6, 7 and 8 go in as the second seeds, and so on and so forth. It is S-curve because of the particular way that the ordering goes: it means that the team ranked fifth are the second seed in the tournament where the top seed is ranked 4, the team ranked sixth are second seeds to the team ranked 3, etc. The first picture inset shows how this should work.

It means that all four playoff tournaments should be equally hard and of equally high quality, higher ranked teams are rewarded with a more favourable draw, but minimal ranking changes should make minimal difference to your chances of progression--it doesn’t much matter if you’re ranked four or five, for instance, but the team ranked 1 is likely to have an easier path to championships than those ranked three or four.

Still seem a little opaque? We’ve used the rankings as provided by Flat Track Stats to show you what playoffs might look like in 2013 -- the tournaments as they stand are listed at the top, and the team’s seeding (equivalent to old regional ranking in the tournament structure) in that tournament is listed on the left.. This is just so you can get an idea of how things work.

T1: Fort Wayne, IN T2: Richmond, VA T3: Asheville, NC T4: Salem, OR
1 Gotham (1) Oly (2) Denver (3) Texas (4)
2 Minnesota (8) Philly (7) Windy City (6) Bay Area (5)
3 Rose City (9) Rat City (10) Naptown (11) London (12)
4 Atlanta (16) Montreal (15) Charm City (14) Rocky Mountain (13)
5 Victoria (17) Detroit (18) Mad Rollin’ Dolls (19) Kansas City (20)
6 Angel City (24) Steel City (23) Wasatch (22) Ohio (21)
7 Tampa (25) Boston (26) Sacred (27) No Coast (28)
8 DC (32) Houston (31) Arch Rival (30) Jet City (29)
9 Blue Ridge (33) Brewcity (34) Mid Iowa (35) Nashville (36)
10 Terminal City (40) Sac City (39) Arizona (38) Columbia QuadSquad (37)

The Division 2 playoffs will feature the next 20 teams--that’s those ranked 41-60; the top two teams from those two tournaments will head to WFTDA Championships too, where they will play off for the top three places in Division 2. As well as the honour of finishing at the top of their division, these games will reward the teams who play in them with more ranking points than their playoff or regular season games, which means strong performances here will greatly increase their chances of breaking into the next tier the following season.

Hopefully that should make it a little clearer how divisions and playoffs will work this year. Next, let's take a more detailed look at rankings, to see how teams will actually earn their seedings and playoff spots in the coming year.

Next page: The Ranking Calculator demystified.

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

We need to invent a word

This is weird, but it seems possible that a Division I team could end up playing in the Division II playoff system and at the end, still be a Division I team. And vice versa, that is a Division II team could play in a Division I playoff and at the end of the season still be in Division II.

We need to invent a word for these "divisional interlopers".

Timing

Certainly is possible, as WFTDA now see two separate calendars:

July-June: season for seeding into tournaments
Nov-Nov: season for placement into divisions

Floaters

May I propose 'floaters'?

Tweeners

Or 'tweeners'

Wrong Word

I actually think we need a new word for the Playoffs, not the teams.

They aren't "Division" Playoffs, because a team's Division doesn't really have anything to do with which playoff a team earns. Teams from Division 1 will be playing in the Division 2 Playoff, and teams from Division 2 will be in the Division 1 Playoff. Divisions are irrelevant to the playoffs really/

Maybe just call them the WFTDA Playoffs, and the WFTDA Championships? Maybe the "Division 2 Playoff" could be called something like the WFTDA Wildcard Tournament or something.

Divisions determine your schedule, but don't actually have any direct connection to the Playoffs. Just sayin'

Opportunity

Who's going to make a WFTDA Ranking app? So we all know how big point spreads need to be.

I'm working on it...

...well, a spreadsheet that I intend to make public, which may develop into more presently.

I'll also try and make sure we include relevant information for weekend previews and the like, so everyone knows what's expected of them.

I need a little bit more information to ensure it's accurate, but I'm chasing that down too!

Domain

I've got a domain with MySQL, so that could be useful. Set it up as a mobile-formatted site, where teams can enter the matchups and see what the point spread needed to climb the chart is.

This is all assuming, of course, that the WFTDA stick to the weights for bouts that they've listed, and only rank according to public scores.

Details

In digging into the the new ranking system I ran into a ton of questions, key among them, how the initial ranking will be arrived at. Sounds like you'll have at least another article to write about this subject.

Good luck chasing down those details. If not, we'll see what's what when the new system debuts in March.

WHAT??? IT'S AN... IT'S AN...

WHAT??? IT'S AN... IT'S AN... IT'S... UH... i don't know WHAT it is, actually.

Here's an observation: Using

Here's an observation: Using the same Flat Track Stats rankings provided up top for ease of reference, here's how the 2012 regional championship games were set up in real life:

Gotham (1) v. Philly (7)
Oly (2) v. Denver (3)
Windy (6) v. MNRG (8)
Texas (4) v. Atlanta (16)

Thus, positional differences of 6, 1, 2 and 12 positions, respectively.

Looking at the hypothetical tournament example listed up above, here's how the regional championships WOULD shake out using the new system, were seeding maintained:

Gotham (1) v. MNRG (8)
Oly (2) v. Philly (7)
Denver (3) v. Windy (6)
Texas (4) v. Bay Area (5)

Thus, positional differences of 7, 5, 3 and 1, respectively. In other words, if one plugs in last year's teams into this year's playoff structure, three of the four games are ((on paper, of course)) GREATER mismatches in the new structure than they were last year. The one exception is Texas v. Bay Area ((4 against 5)) as opposed to Texas v. Atlanta ((4 against 16)), but do recall that Texas v. Atlanta at South Centrals was a tremendous game, so go figure.

Therefore, whilst i find the new ranking system "interesting," at bare minimum, i really do not yet understand exactly what is perceived as being gained here in terms of competitive balance or compelling tournament matchups.

Another observation...

...is that the top 8 teams probably won't change much. The difference will be the next tier down, in the third-place games.

But that's just a gut feeling.

The difference

The difference is that these rankings are based on facts, rather than votes.

Democracy said that the earth was the center of the solar system, and (according to some polls) still says that evolution is a lie. I don't trust democracy when it comes to verifiable facts.

This system at least moves in the right direction, using bout scores to rank teams, sort them into divisions, and set scheduling expectations. I like that it penalises both sides for blow-out bout, as in the LRG-ARRG example above. I HATE HATE HATE 1-sided non-competitive bouts.

Sort of

"The difference is that these rankings are based on facts, rather than votes."

The difference being that the method used to interpolate those facts into a ranking is still highly speculative. This proposes a single model, and the test listed in the parent comment is one way to compare the two models. There are many other models one could choose: FTS uses the Elo rating system which is widely accepted as being a good model for ranking systems where there is a shortage of cross-data points.

Two run with your analogy, both the geocentric model and the intelligent design model are models that make predictions, which once we have more data (bouts) we can see that they do not fit the models and/or don't fit as well as other models. I haven't seen a good independent suggestion for an independent method of testing to see if this is a good method of ranking other than "people complain less", which seems to still be democracy based.

Sort of is better, ain't it?

Yeah, I get that. All ranking schemes are based on some form of subjectivity unless there is a system where upsets are theoretically impossible. I personally use one for UK teams based on this paper: https://umdrive.memphis.edu/ccrousse/public/MATH%207375/PERRON.pdf which seems to do a good job as well.

Thing is, the "perception factor" here only comes into play at the design phase and the confirmation phase, not at the individual ranking of teams. Sure, votes confirm or deny an overall ranking, but they do not go into whether London are #5 or #7.

I was trying to make a global ranking with the system described in that paper above, but my server begged for mercy.

RE: The difference

You, my friend, get an upvote.

The new WFTDA ranking system may not be perfect, but it's damn sure better than the borderline-abstract voting system previously used. While many votes may have been based upon view of available statistics, they were still based upon viewpoint, not solely statistical analysis. Major step in the right direction.

one big tourney

revnorb wrote:

In other words, if one plugs in last year's teams into this year's playoff structure, three of the four games are ((on paper, of course)) GREATER mismatches in the new structure than they were last year.

But is that even the point?

My assumption is that the goal here is to increase the likelihood of the 12 best teams making it to championships. The last few years, many would argue that the Western region has been consistently under-represented (if you look have faith in DNN's raking system, for example) at champs. With this new system, if six of the top 12 teams in the country all happen to be from the same geographical area, that won't keep them from being able to compete at champs. Yay for that.

The Big 5 now becomes MUCH more like one big tournament bracket. The finals of each "regional" are more like quarterfinals (sort of) in the big picture. And, as such, the top teams SHOULD have fairly easy match-ups in those bouts (just as the #1 seed in a tennis tournament would in the quarterfinals).

I suppose. So now we get

I suppose. So now we get ((say)) Rocky Mountain, Rose City and Rat City at Championships instead of ((say)) Charm City, Naptown and Kansas City. That's cool, but if, as you state, the top teams "should" have fairly easy match-ups in the regionals (("not-regionals?")) championship bouts, then the value of having the legit #10, #11 and #12 teams represented at Championships instead of having the #13, #14 and #15 teams sneaking in due to geographically-based selection is sort of minor to me, and possibly not worth eliminating the more evenly distributed geographic representation to achieve ((obviously, if i were on Rocky, Rose or Rat, i wouldn't imagine i'd feel this way)).

In other words, it doesn't make a vast amount of difference to me if it's the #10, #11 and #12 teams being eliminated on Friday, or if it's the #13, #14 and #15 teams being eliminated on Friday whilst the #10, #11 and #12 teams sit at home and gnash their teeth. Obviously, having the #10, #11 and #12 teams there is preferable, but as a trade-off for the lessened regional identity and rivalry that will come with the new format, i'm skeptical that what is being gained will offset what is being lost.

Host leagues not playing in own tournament

revnorb wrote:

In other words, it doesn't make a vast amount of difference to me if it's the #10, #11 and #12 teams being eliminated on Friday, or if it's the #13, #14 and #15 teams being eliminated on Friday whilst the #10, #11 and #12 teams sit at home and gnash their teeth. Obviously, having the #10, #11 and #12 teams there is preferable, but as a trade-off for the lessened regional identity and rivalry that will come with the new format, i'm skeptical that what is being gained will offset what is being lost.

It makes no competitive difference whatsoever. The #1 or #2 team is going to destroy their first round opponent regardless if they're seed #25, seed #26, seed #29, seed #32, or any seed in that range. The only thing this system does is force a team to travel several thousand more miles than they may have to to go 2-1 or 1-2 against a different set of opponents.

But the most glaring oversight? It's highly probable that a team will not play in the regivisional tournament their league is hosting. It's great the the Fort Wayne Derby Girls get to host a playoff site, but if they're playing in Oregon and not in Fort Wayne for the playoffs, isn't that kind of a terrible and horrible thing for their local fans?

Subject field is required.

WindyMan wrote:

But the most glaring oversight? It's highly probable that a team will not play in the regivisional tournament their league is hosting. It's great the the Fort Wayne Derby Girls get to host a playoff site, but if they're playing in Oregon and not in Fort Wayne for the playoffs, isn't that kind of a terrible and horrible thing for their local fans?

This presumes that hosts would not get first priority on which set of competitors they will host. If FMDG is playing in a divisional tournament, they are most likely to get the set of ten competitors that they are in.

Except...

Okay, so what if two or more host leagues wind up in the same bracket? Then what?

Or what if accommodating the host league forces four or five or more other leagues (including the top seeds) to travel much farther than they would have had to, if instead the host league traveled to come to them?

A rigid S-curve with no flexibility for regional accommodation is filled with little pitfalls like this. I don't understand why the WFTDA would stick to their guns on this when it has the potential to be more trouble than it's worth.

Is there any data regarding

Is there any data regarding local fan attendance at regional playoffs in the past? While I see your point, it may not be "worth it" to stick to the previous system - deemed broken by many voices, the least of which is the WFTDA membership - rather than try a new system which may be inconvenient to a very small minority? I don't doubt FWDG's local fans would love to watch them play in a divisional playoff in their hometown, but then again if only a coupe hundred fans are disappointed so that the WFTDA has a more accurate representation of the "best 12" at Champs, maybe it's worth it. I am not one to condone upsetting fans, but tournaments are different than your average home game.

Ignore what crowds look like

Ignore what crowds look like when the local team isn't playing. When the local team was playing, there were a ton of people at the tournaments this year. Some of them were jam-packed. Whether a fan shows up for one game or all of them, they're still paying for a ticket.

I guarantee ticket sales will drop dramatically if there aren't any nearby teams.

The stupid thing is that the new system can take geography into account. There's absolutely no need to have the s-curve for every seed. None. Any seed higher than 4 or 5 is very unlikely to compete for a trip to championships and the teams that have stepped up and done that in the past have all been complete surprises which means the s-curve wouldn't help with that anyway. Let fans watch their team play.

Host Leagues – an Outdated Concept

For several years now WFTDA member leagues have been delegating increasing areas of responsibility of the Big 5 Tournaments to WFTDA's administration. A year ago the membership mandated and issued the necessary resources to the WFTDA administration to take total control of the Big 5. The result being in the production of the Big 5 we experienced for 2012.

For all practical proposes WFTDA's administration does not need a “host league” to produce a Big 5 tournament, other than as a labor pool to man venue operations. WFTDA administration manages all operations of a Big 5 tournament and can produce a Big 5 tournament anywhere. This is how the majority WFTDA member leagues want the Big 5 to be run. The WFTDA administration is merely the instrument of the membership majority.

It is interesting to look at 2013 Division 1's Big 5 tournament locations.
Fort Wayne – Division 2
River City – Division 2
Blue Ridge – Division 2
Cherry City – Division 3
and Championships at Brewcity

The chances of any of these “host leagues” playing at their “hosted event” is realistically, highly unlikely. In my opinion this incongruity is something the WFTDA membership should look into and, in all probability, is already doing or has already done.

What I'd like to see …

I'd like to see WFTDA relink Division 1 Big 5 tournaments to Division 1 cities. As stated above WFTDA can produce a Big 5 event anywhere, so why not in the city of a Division 1 league? And that Division 1 leagues ought to take responsibility to make themselves available for supporting Division 1 events. Other divisions have their own division events.

At the end of the WFTDA Championship Tournament every league will know in what division they will be in for the next season. This would be the city list that WFTDA administration uses to decide where next year's divisional events are to be held. The tricky part will be designating alternate cities to take into account the variegates of the S-curve, in case of multiple hosting cities ending up in the same tournament, alternate host cities would then be availed. It is quite likely that each division will have a large core of leagues that changes very little from season to season. That should help greatly in deriving the hosting cities.

I guess we can call the hosting cities, candidate hosting cities as the final decision will not become apparent until the 1st of July. This would give everyone, fans, leagues, officials and hosts two to three months to prepare. As the situation stands now only tournament participates and fans have two to three months to prepare. Two to three months is apparently enough time for everyone to make the necessary travel reservations. Or WFTDA could stick with fixed locations and just let the chips fall where they may. I'm not a fan of that idea, but what do I know.

No host cities would be harmed in this process as WFTDA administration does all planning and organizing of Big 5 events currently. The only thing the leagues of host cities have to do is be ready to show up for work. Not any different than what most league members and volunteers do on any bout day. I have worked directly with WFTDA Big 5 administration officials and I can assure you they are dedicated and capable people that only want what's best for Women's Flat Track Derby.

WFTDA should draft a policy that leagues are responsible to be available to help produce their division's events. WFTDA can setup a bid/allocation system to distribute division events as evenly as possible among the division leagues. We'll never see perfection, it doesn't exist.

I am not an original thinker, as far as I know WFTDA has already thought of this idea, and possibility even proposed it … a year ago.

In lieu of doing Derby 101,

In lieu of doing Derby 101, the announcers should explain the new tournament structure before every bout! "How we doin' on time?"

Or...

WFTDA could do away with the s-curve (as it is completely unneeded for the lower seeds) and place teams into playoffs based on geography and you can skip that whole mess. Plus, then fans could actually travel to watch their team play. I do think having division 1 teams host is a good idea, though, so that local fans show up and buy tickets. I have a feeling this year's playoffs are going to set a record for low attendance since no host teams are likely to play in the tournament they're hosting.

Attendance, Schmattendance

thebigchuckbowski wrote:

I have a feeling this year's playoffs are going to set a record for low attendance since no host teams are likely to play in the tournament they're hosting.

I think this is a complex issue that deserves its own thread...my eyeball test at South Centrals at the Pershing Center this year indicated that three-day passes at the door cost twice as much and attendance was half as much as South Centrals at the very same venue two years previously. So, i mean, the line on the graph is already trending the wrong way, from what i've seen ((which is, admittedly, only a small fraction of the totality of things)). I guess we'll see.

To counter your argument

... maybe Division 2 or 3 leagues would LIKE to have the Division 1 tournament on their floor so they and other local leagues can see how the top leagues are playing.

People who attend Regionals are hard-core derby people. Since 2008 I've been to several regional tournaments and they're primarily filled out by skaters within driving distance of the venue, plus a small core of fans that travel with each team. The tournaments look better in smaller venues, in smaller cities*, that can fill up a greater percentage of seats (even if it's the same number of seats in a larger venue/larger city).

I couldn't bring out any study, but I think it's safe to say that if someone was going to shell out enough money to watch a tournament that wasn't already easy-access, they'd pick champs instead of a regional.

* I realize that smaller venues are not limited to small cities, but it's increasingly hard to find a professional-type venue in large or medium-sized cities, where they get bulldozed in favor of behemoths that seat 15,000+.

To follow up on your point on venues

I've noticed that since the WFTDA started hosting regional and championship tournaments, they've all been, with the arguable exceptions of Bay Area, Chicago and Philadelphia, in smaller to medium sized markets and I suspect that leagues in the large markets may be challenged not just by availability, but by the prohibitive rental cost of facilities. We may have to ultimately accept that only a finite number of leagues/metro areas, who may end up hosting multiple times, will be capable of hosting tournaments short of generous investor/sponsor intervention to assist the leagues in the larger markets.

Top 12

revnorb wrote:

In other words, it doesn't make a vast amount of difference to me if it's the #10, #11 and #12 teams being eliminated on Friday, or if it's the #13, #14 and #15 teams being eliminated on Friday whilst the #10, #11 and #12 teams sit at home and gnash their teeth. Obviously, having the #10, #11 and #12 teams there is preferable, but as a trade-off for the lessened regional identity and rivalry that will come with the new format, i'm skeptical that what is being gained will offset what is being lost.

But I think you're over-simplifying. Again, just to use 2012 (and DNN rankings) as an example, there were four teams outside of the top 12 that made it to champs and four teams in the top 12 that missed champs. That's a 33% miss rate, which I'd argue is significant. Personally, I wouldn't assume for a minute that Rat, Rocky, London, and Rose would have been eliminated on Friday. I think any of them could play a competitive game against teams like Windy/Philly/Minnesota/Charm.

Anyway...long-winded way of saying that I personally like that the new system is bending over backwards to get the best 12 teams in the biggest tournament...albeit at the expense of the "regional" tournaments possibly being less interesting (though I imagine the battles for 2nd and 3rd place will still generally be great).

The other fun thing about the "regional" tournaments not being regional is that teams will get a chance to play against teams that they probably never would have faced otherwise.

The best 12

Dave Wood wrote:

I personally like that the new system is bending over backwards to get the best 12 teams in the biggest tournament...albeit at the expense of the "regional" tournaments possibly being less interesting (though I imagine the battles for 2nd and 3rd place will still generally be great).

This is the big point for many though. There's no good reason why Champs is supposed to be the "best 12 teams". No other major sport works that way; they all have some sort of playoff structure that puts the best team and a bunch of other good teams into playoffs and then a structure that lets the best team win (assuming the best team has a 100% chance of winning every game; obviously without that it's an impossible task as there is no well ordered list to sports ranking /math).

In other words, if you couldn't get first or second place in your region, you won't win the Hydra. What this format is saying is that we have some teams that we want to award the extra recognition of being invited to a tournament that we know they won't win, and the cost of that is that we will weaken the interest of the regional tournaments.

12

N8 wrote:

There's no good reason why Champs is supposed to be the "best 12 teams".

I sort of agree with you here. I've always been basically fine with the argument that the point of champs is to declare a champion. And if that's really your only goal, the fact that the 4th best team might not even be in the tournament is irrelevant.

But in reality, the purpose of champs seems at least to be to find the top 3 teams (since they get medals). The pre-2013 structure really didn't do this very well (take 2011, for example, where most people would have put Rocky Mountain at #3, but because of the structure, they had to play Gotham in the Quarters and therefore missed the Top 3 (or 4) completely -- under the 2013 system, Gotham, Oly, and Rocky would almost certainly have had 1st round byes at champs that year). I think the fact that the four best teams are likely to get the byes is a good thing.

So I'm not sure it's reasonable to say that there's "no good reason" to put the 12 best teams through to champs. While it's not NECESSARY to have the 12 best, it'll make for a great tournament.

Agreed!

N8 wrote:

What this format is saying is that we have some teams that we want to award the extra recognition of being invited to a tournament that we know they won't win, and the cost of that is that we will weaken the interest of the regional tournaments.

Exactly! Thank you for phrasing this like a normal person. I have no more to add.

Argue, argue, outrage, argue, etc. etc. etc.

Dave Wood wrote:

Again, just to use 2012 (and DNN rankings) as an example, there were four teams outside of the top 12 that made it to champs and four teams in the top 12 that missed champs. That's a 33% miss rate, which I'd argue is significant.

Perhaps, but it's silly to suggest that #12 London belonged at Championships and #14 Charm City didn't; Charm City beat 'em fair and square! There's no real arguing that one.

Apples and oranges.

revnorb wrote:
Dave Wood wrote:

Again, just to use 2012 (and DNN rankings) as an example, there were four teams outside of the top 12 that made it to champs and four teams in the top 12 that missed champs. That's a 33% miss rate, which I'd argue is significant.

Perhaps, but it's silly to suggest that #12 London belonged at Championships and #14 Charm City didn't; Charm City beat 'em fair and square! There's no real arguing that one.

I don't believe Charm City was one of the teams he's talking about... Charm City was in the top 12.

It never ceases to amaze me

It never ceases to amaze me how people can look straight at a problem, and completely miss the obvious.

Which is, that the whole idea of playing #1 vs #8, #2 vs #7, and so on in a tournament system, is not about anything other than:
1. Making sure that the tournament takes a very specific number of games.
2. Giving the higher ranking teams more games to make the travel worth the effort.

Lower ranking teams should have to EARN their right to play a higher ranking team, by beating all their lower ranking competitors first in a battle for survival.

It should be #7 vs #8, #5 vs #6, winner of game 1, plays winner of game 2.

Historically, we know that there is a huge distance between #8 and #1-4. #5-8 are the challengers. They need to be sorted out so the strongest champion among the challengers is chosen to challenge the higher ranks. This would keep the games closer.

Winner of #5-8 plays #2. Winner of #3 vs #4 plays #1. Winners of those two games play for the championship. Losers of those two games play for #3.

Erase all concerns of the lower ranking teams playing more games than higher ranking teams, and thus being at a competitive disadvantage, or other "fairness" considerations, and concentrate on competitive merit, and you'll have fewer blowouts and still find the best team.

Battle for the Bank

Is this the new Battle for the Bank tournament structure? And WFTDA needs a structure for 10 not 8.

It almost was one year

It almost was one year (Arizona). The structure was announced late, certain teams complained because they wouldn't have a game on the first day, for which they'd already booked travel. But at least it started with the lower level teams playing each other, and the games were more satisfying.

The next year it went back to the guaranteed and pointless blowout system, because the WFTDA knows better than banked track, so everyone should imitate them.

You can make a "graduating" system out of any number of teams as long as it's an even number. The point is, we know in roller derby more than any other sport, that the skills increase at a greater than linear rate from bottom to middle, and level off at the top.

10 teams
10 vs 9; 8 vs 7; 5 vs 6. Winner of game 1 vs winner of game 2. Winner of game 3 vs winner of game 4. 3 vs 4. #2 plays winner of game 5. #1 plays winner of game 6. Loser of games 7 and 8 play for #3. Winners of game 7 and 8 play for the championship.

12 teams
12 vs 11; 10 vs 9; 8 vs 7; 6 vs 5. Winner games 1 vs 2. Games 3 vs 4. Winner of game 5 vs 6. 3 vs 4. Winner Game 7 vs #2. winner game 8 vs #1. Losers of game 9 vs loser of game 10 for 3rd place. Winners of 9 and 10 for Championship.

The problem I've heard is that this isn't fair to the lower level teams, because they have to play so many games if they're qualified to make it to the top. My answer is, you want less games to get to the top, do what it takes to get a higher ranking during the regular season.

Interesting...

Well it is an interesting alternative system and I'd like to see a non-Big 5 event experiment with it. But WFTDA originally had an alternative tournament structure and fans hated it and insisted on the traditional model. WFTDA made the change to the current model in 2008.

I'm loving the Division 2 playoff structure

As it stood in 2012, a team ranked below 15th in their region going into the year didn't have much of a chance to get into a regional tournament. There wasn't really any reason for a top 10 team to play them and give them the opportunity to move up into the top 10. So they essentially knew in January that they wouldn't have much to play for that year in terms of a post season. It wouldn't matter if they improved a lot in July and August, at least not until the next year at the soonest.

A Division 2 playoff structure changes that. A team ranked in Division 3 in January now has more of a path to play their way up. They're only 1 or 2 key wins away from finding themselves in a playoff tournament. —And the teams ranked 41-60 will likely all be very close in terms of talent level. So it's entirely possible that a team could begin the year ranked in the 70's, improve throughout the year, and find themselves in Milwaukee for the WFTDA Championships. That's an exciting thought.

Fort Wayne is not in Idaho.

;)

This is a very helpful article, however.

You make a good point

Fixed! And thank you.

delegation shenanigans

There are lots of reasons that we're enjoying having a couple of non-US-natives (Lex and Beck) step up to take on more core DNN responsibilities, not least of which is that we now get at least as many mistakes of US geography as we've always had with teams from other nations :)

Disturbing Question

DNN created their Power Rankings in response to a need, when WFTDA abandoned their national ranking system and adopted a regional based system.

With WFTDA having moved back to a national/global ranking system, how will that effect DNN's Power Rankings?

Informative

I see the power rankings as becoming an inform/educate system, and remaining in place. It is not at all unnecessary to have them. NCAA football has multiple polls, and even the NFL has power rankings run by ESPN rather than the league. If DNN keep up the rankings, they'll definitely have value to coaches and fans.

It shouldn't affect the Power

It shouldn't affect the Power Rankings at all. It will be a test of whether DNN's system is more correct in actual game play then the WFTDA's system. The WFTDA system is designed to show where teams stand based on past performance. The DNN Power Rankings are designed to predict outcomes in future engagements and often takes into account missing players who were unavailable for a given game but will be available in all future games (a Gotham game without Bonnie Thunders would be a lot different from a game with her, for example).

Also, DNN includes teams that are not WFTDA members when they have wins or close final scores against highly ranked WFTDA opponents in WFTDA regulation play.

how much time

will teams have from the last *ranking* bout played within this model, to be able to plan their travel for playoffs? Thinking particularly of international leagues and the expense of booking flights without much notice. Im not knocking this model at all just curious. Definitely prefer a ranking that's based not on votes but on actual performance :) and love the pathways from div 2 to div 1.

They will know slightly earlier...

...because they can work out their own rankings, in the most part, and won't have to wait for things such as vote submissions.

At the absolute worst case, I reckon an international team trying to break into playoffs would know they were going the exact same time as before. Previously no flights could be booked anywhere before rankings came out--under this system if you get your games in and know everyone else's schedules, you should have a very good idea of whether you're going or not when you step off the track after your last qualifying game.

Planning!

Lex Talionis wrote:

...because they can work out their own rankings, in the most part, and won't have to wait for things such as vote submissions.

At the absolute worst case, I reckon an international team trying to break into playoffs would know they were going the exact same time as before. Previously no flights could be booked anywhere before rankings came out--under this system if you get your games in and know everyone else's schedules, you should have a very good idea of whether you're going or not when you step off the track after your last qualifying game.

Yes, but before, the international team would know that if they made it, they're going to go to wherever the city their regional playoff was in. Now there are four possibilities. I'm really interested to see where London and Montreal end up this year.

I don't think it makes that much of a difference

As the fundraising can kick in earlier; the local recces are all done in advance by WFTDA and league hosts.

The difference for (say) London or Victoria in sorting out travel to any of the US locations is going to be significantly less compared to their regional tournament than domestic teams, given the difference in distance travelled is proportionally much smaller.

It will affect everyone; I just don't think it'll affect international teams more than domestic ones, but quite the opposite.

In which I provide my minority opinion and get downvoted

I still strongly dislike one aspect of this, and that's the "Game Weight" portion. The justification is that "playoff games are more important, and so we want them to count for more."

First, mathematically, those games already count more. They count more because your opponents will have a higher than average "Opponent Strength". There is no need to have an additional weighting factor just to make it worth more. I know those weighting values look small, but they're actually going to have a much larger impact than most people realize. To use the Montreal/DC example, that bout is weighted heavier than if Montreal would beat Gotham in a regular season bout, and that's not even the highest Game Weight available.

Second, there's an underlying message that it's okay to value regular season games less. Yes, I know people have lives and can't always make every game, but it really gives big fat raspberry to the regular season when the playoff games are weighted this heavily.

Lastly, this extra weighting is going to make it far harder for teams on the edge to play up their game and break into the playoffs. It's just going to make it easier for the same N teams to stay in playoffs from year to year, even if they aren't playing the best derby any more.

A weighty issue

N8 wrote:

Lastly, this extra weighting is going to make it far harder for teams on the edge to play up their game and break into the playoffs. It's just going to make it easier for the same N teams to stay in playoffs from year to year, even if they aren't playing the best derby any more.

I partially disagree with this. Teams that just missed making the playoffs will also be in the playoffs—the D2 playoffs, where games have similar weights to the consolation brackets in D1, where those teams would have wound up playing anyway. So I don't foresee an issue with teams in the 41-50 zone getting enough quality ranking points to move up into the 31-40 zone, if they have a good run in the D2 playoffs and are better than the bottom-feeders in D1.

However, I agree with your point about the same teams staying in the playoffs from year-to-year. Not that this hasn't been happening already, but a team's make-up can change considerably over six to eight months. Giving equal weight to games played 11 months ago vs. games played 1 month ago doesn't make sense if you ask me. Particularly, since the locking-in for playoff seeding includes the weight-boosted playoff results from the year before. I'd like to eventually see some kind of degradation factor to make past games (like, past six months) worth less in the total rating.

And a minor quibble

I also would have preferred that the points were gained based off of rating and not ranking as moving up ten positions is likely to take a smaller ratings gain (and thus be easier) in the middle ranks than the top.

I have a different issue with playoff weighting

My issue is that it doesn't make the result more significant -- it makes the fact you were there more significant. Everyone gets 25% more (or whatever) points for the game, and that's that.

If I wanted a system that made playoff games more important, I'd set it up so *results* of playoff games were more important, not the taking part.

One way of doing this is to tweak the formula slightly to move the weighting factor up to a power level, as it were.

Let's say you change it so that the weighting factor sits as a power of the score factor. That way you are rewarded for a game where you score more than 1/3 of the points, and penalised if you go to playoffs and lose quite heavily. This is actually quite neat -- it is akin to the bonus points you get for winning by large margins in certain other sports, such as some rugby competitions, or by losing narrowly in others. Both teams would share the bonus points for playing out a really close game--some consolation for the loser, perhaps, who gets some reward for playing in an awesome playoffs game but being dumped into the consolation bracket regardless.

You can tweak this effect by changing the multiplier inside the scoring factor to set up the level at which you want these bonus points to kick in (and the penalty for being blown out), as well as changing the power that the score ratio factor is raised to to alter the weighting of the game.

This has two obvious implications -- it will really punish teams such as Dutchland and reward their opponents (just going off last year's playoff results), but also has the potential to punish 7-10 seeds in their first game who get run over by a far superior 1/2 placed team, through no fault of their own. The latter effect might be cancelled out by the extra points to be gained in the consolation bracket, but it clearly depends on the magnitude of the blowout.

Don't Play Sanctioned Games

N8 wrote:

... there's an underlying message that it's okay to value regular season games less ... but it really gives big fat raspberry to the regular season when the playoff games are weighted this heavily.

As usual N8 gets it before I do, but I'd like to expand on N8's observation. For a team that participated in a regional and/or championship tournament those games have more points than regular season games will. Since your rank is determined by your game point average, keeping your average as high as possible assures a high ranking. By playing a large number of regular season games, you are gathering a lower game point average and actually playing yourself into a lower rank. It makes no sense to play anything other than the WFTDA minimum.

But for a team to improve it is crucial that it regularly test itself against the best competition it can find to uncover the areas it needs to improve on. I see teams traveling hundreds even thousands of miles to NOT play a sanctioned game, but to scrimmage or play a regulation game. There's even a precedent for this, last year's Minnesota/Kansas City scrimmage.

6 Month Bout Decay instead of 12?

Could a 6 month bout decay (instead of the current 12 month bout decay) help the issue of tourney weighting, as well as help the issue of "By playing a large number of regular season games, you are gathering a lower game point average and actually playing yourself into a lower rank"? I'm no math geek, so I don't really know - I'm really just asking the question.

On the surface at least, it seems like teams could then play a large number of games throughout the season (ie - year), without the fear of it necessarily meaning they would play themselves into a lower rank, while also taking out the results of the previous year's tournament weighting so that it puts all teams on equal footing from year to year without giving any unfair advantage to teams that go to tournaments (who get weighted game points) versus teams that don't go to tournaments (who only get the benefit of single-weighted games)? Maybe someone who knows more about math can answer the draw-backs to this.

I mean, I know it means people will schedule their season differently (likely scheduling riskier teams at the start and end of their seasons, while scheduling safer teams in the 6 months leading up to tournament eligibility), but at least teams wouldn't then be scared to play risky teams AT ALL (the way the system is set up now) and it promotes game-play rather than taking away the benefits of playing more games (the way the system at least seems to do now).

I'm sure I'm missing some glaring issue with doing it this way, but I know DNN reader's can at least sort it all out so we can see all sides. ;)

Or at the least, don't

Or at the least, don't partially take a previous season's results and put them towards the next season's tournament eligibility.

As I understand it, the 2013 tournament eligibility and seeding will be based on the ranking calculator's results that occurred from the cut off date in 2012 to the cut off date in 2013. So, for example, if the cut-off date for tournament eligibility ends up being June 30th of each year, the results that are going to be calculated in the rankings calculator to determine a team's tournament eligibility and seeding are a team's results from June 30th, 2012 to June 30th, 2013.

To me, this is like saying "Hey Patriots - We're going to use 1/2 of your 2011 season results and apply them towards your 2012 playoff seeding, cool?". It seems crazy to me that WFTDA would do this.

I would at least think that the fair way to do things would be to set the rankings calculator to match that of the games minimum requirement, in that rankings will be calculated from the end of Championships in one year to the end of Championships in the next year, at which point Divisions are set - not taking into consideration a team's previous season results. The calculations for tournament eligibility would then be all games that occurred from November (post-Championships)-June - taking into consideration only that season's results up until that point, and removing the serious advantage/disadvantage of the tournament weighting issue as well.

Misinterpretation

Maybe I'm misreading the document, but I think the tournament game weight is only applied to Championships. If that is true, then it's also a good idea to sand bag and qualify 7-10 seed for championships so you get an extra game in with 1.25 multiplier. Best option is to qualify 10th and win the tournament, that way you get more games in with bonus tournament multiplier, and your opponent strength multiplier is maximized. The next year you can kick back and pretty much ensure that you will be in Divisional playoff since your rating sky rocketed from championship tournament the year before.

Kind of...

You've got a great point, but there are a few corrections. Championship games are not the only weighted games. The games that get extra weight are the Division 1 tournaments, Division 2 tournaments, AND Championships. From the white pages document released by WFTDA:

-Tournament games are weighted higher than regular season games.
-Championship games are weighted highest.
-Division 1 tournament games are weighted higher than Division 2 tournament games.
-Elimination bracket games are worth more than consolation bracket games.

But, yeah, this system basically takes away the benefit of having a bye (other than you are sure not to get bounced in the first round) - but it really isn't going to help your ranking in the following year the way it would if you're a team that goes in #7-10 and wins all 4 games to get to the top. But, again, all of this would be a moot point if tournament results from one year didn't go into the tournament qualifying calculations in the next year (a major flaw IMO, as it serves to keep the people who went to tournaments going to tournaments and the people who didn't from ever getting in, merely based on the fact that a team was *at* tournaments in the previous year and not necessarily based on merits of winning games in the season the actual tournament is being held).

THANK YOU!

I was one of the many confused people assuming that Divisions determined the Playoffs. Whew! I'm so glad you cleared that up. I was actually pretty confused and (unpleasantly) surprised that teams would not be able to play their way into Division 1 or Division 2 playoffs. I'm glad I was wrong.

Anyway, great job on this to the WFTDA Games Committee! It was a ton of work and they did a great job with it.

And, thanks DNN for educating and informing the Derby public. What? An incident of totally responsible journalism? Nice!

Close only counts...

...in horseshoes and hand grenades. Oh, and roller derby.

My only significant gripe with this whole thing is that winning doesn't matter. The difference between winning 151-149 and losing 151-149 is microscopic. That's painfully unlike sports IMHO. I don't disagree with taking the score into account when generating the rankings, I just think there should be a "did you WIN?" factor in there as well! (For example, maybe you get one point for winning and the other 2 points are split up based on the scoring ratio -- that ensures that winning is at least twice as good as losing.)

I haven't done the math, but I'm pretty sure this scenario could (and will) happen:

The #10 team beats the #15 team...
...but only by one point...
...#15 jumps AHEAD of #10 team based on this result!

That just feels broken, doesn't it? (or am I missing something...?)

I think...

...it's OK, and a natural consequence of being able to set up games with big skill differences.

Winning certainly matters in the host of single-elimination tournaments that abound!

Also, running some maths with ideal ratings (ie ones set where a team's rating is unchanged if it draws with itself), if a team in 10 beats a team in 15 by one point then neither team moves in terms of ranking. This hold as long as both teams have 5 or more games in the ten months preceding the ranking period in which this game took place; fewer than 5 and teams start to move a little bit.

This may well change a little in reality (especially if Gotham's form continues and they play lots of games in 2013; they'll serve to remove points from teams below them and cause the chasing bunch to become tighter), but I think it gives a fairly good indication of things.

But losing still > winning, right?

Maybe I went too far trying to guess who would move up/down, but it's still the case that the two teams would get similar total scores for the bout, right? And the #15 team that LOST would get *more* points than the #10 team that WON.

Win/Loss Factor:
#10 team wins 151-149 = 1.51
#15 team loses 151-149 = 1.49

Opponent Strength (based on 172 teams):
#10 team - 1.83 opponent strength
#15 team - 1.88 opponent strength

Game Weight = 1 for both teams

Total Points for WINNING team = 1.51 x 1.83 x 1.0 x 100 = 276
Total Points for LOSING team = 1.49 x 1.88 x 1.0 x 100 = 280

If nothing else, it's something to keep in mind. It'll be interesting to see how it plays out.

By this math, teams are now

By this math, teams are now incentivized to maintain their ranking by winning more and by greater point margins. While the difference between the #10 and #15 teams ought to be marginal, the fact is that they both must fight for their ranking, which should hopefully be displayed in a game between the two. #15 will be fighting to even just win by one point, while #10 has to fight to win by a somewhat substantial amount. That prevents #10 from getting "lazy" or just running the clock out with a small lead. Not saying this happens, but it certainly won't now.

It's an outrage

JK.

But seriously, Winning DOES Matter or else why would people be trying to hard to win?
If it didn't matter everyone would be like, oh no big - we were just down by 1.
What if a team was amazing and lost against Gotham by 10 but also lost against the other ranked teams by 1 or 2 or 6?
When you're the one who lost, it's great to be positive and say, we only lost by 1 or 3 or 12 and know what you need to do to win the next game and that you had a good chance at winning.
When you're learning to play, winning doesn't matter - skills matter and learning strategy and teamwork matter.
But at some point, winning does matter or else why would we care who wins?

For whatever the reason - strong training, wanting it more, soul/heart, luck, good strategy, less penalties, more teamwork, etc. the team that won still played a better game than the losing team. Better enough to win and score more points. So that should still count for more points in a chart somehow, even though it messes with the other logic you guys are trying to add into the equation. The calculus isn't quite right there. So maybe a win vs. a team who is ranked higher than you is weighted higher than a win vs. a team which is ranked below you.

But you can also argue that once you get into the middle rankings of leagues, anything can happen -- so teams ranked from say, 10-15 should be relatively equal in skill but depending on the day and situation - one of them is going to prevail. Whether it's because the winning team played an amazing strong game and everything went right, or because less things went wrong than the other team - they still won.

Basically the logic for scoring should be, if #10 ranked team beats #15 team by 1 point, it is worth less than if #10 ranked team beats #9 ranked team by 1. But if #10 ranked team beats #5 ranked team by 1 that's worth even more.

OK, so now someone can go calculus this jawn up.

New WFTDA leagues....

One thing that is addressed poorly is that new "full WFTDA" leagues will be put into Division 3 regardless of skill level. I know that my league in particular, based on past wins, would most likely be placed into Division 2 via skill, but when we do graduate out of Apprentice, we will be placed into D3 for a full year. This kinda sucks. How do we get enough D2 teams to play us in the next season when we are ranked D3? And forget about D1 teams....

Sorry to be a negative Nelly, but I think there must be a better solution.

Your division doesn't matter,

Your division doesn't matter, it's where you're ranked that matters. Once you get your ranking up by playing some sanctioned games, there will be plenty of D2 teams that will want to play you if your team is truly at that level.

Divisions mean absolutely nothing except to set the minimum amount of games teams have to play.

Let me try to be helpful

You're right, it sucks being an ambitious new league. Getting the leagues you want to play to take you seriously is often is next to impossible. But this is the way it's always been for the over one hundred leagues that have come before you. Everyone understands, but you still have to prove yourself. So how do you get leagues to take you seriously? Truthfully your opportunities are rare and when they present themselves to you, they must be seized. Literally your chance to move into another division or be invited to a WFTDA tournament can hinge on a single game during the season.

As I'm sure you know, scheduling roller derby games is an exercise in the art of the possible. Everyone labors under the constraints of their schedule and budget and everyone's schedule and budget is different from everyone else's. Unless your budget is zero, which true for most leagues. If we can't afford to get to you and you can't afford to get to us, then I don't think a game is going to happen between us this season. In this helter skelter environment no one gets the schedule they want, everyone settles for the schedule they can get.

One of the surest ways of getting to play better leagues is though tournaments and multi-bout events, such as Wild West Showdown, The Big O, Dust Devil, Spudtown Knockdown, Brewhaha, Governor's Cup, Clover Cup, Franky Panky and East Coast Extravaganza or any other WFTDA sanctioned events you can find. It can be difficult to get into these events, but in many cases such events may be the only real opportunity you have all season to play the type of leagues you want to play. Apply early and inquire often.

Be ready to travel. Sadly no one is going to beat a path to the door of an unknown team. You will have to go to them. Schedule and budget constraints are not only different for every league, but they also change during the season. An unhappy fact of roller derby scheduling is every now and then a previously scheduled opponent will have to cancel their bout or not be able to attend a tournament. What was possible when the plan was made in July has become impossible to execute in March. This happens to everybody eventually. If this happens to a league you want to play or a tournament you weren't able to get into, here's your chance. Let leagues and event directors know that if they suffer a cancellation, you're there for them.

Be a good guest and good host. After parties matter, as a guest, if you have the chance, attend them. Everyone understands if it's Saturday and you've got another game Sunday or if it's Sunday and you've got to get everybody back so they can be at work Monday. Make clear to your hosts what your after party plans are well before you arrive, even just a short, non-alcohol appearance is showing respect for your hosts. As a host, be gracious and generous, no one likes to be around a bunch of stuck up bitches.

Patience. It normally takes at least two, even three seasons for your good reputation to get around. Being in a hurry doesn't make the time go by any faster.

Hope this helps.

Thanks!

Thank you for your reply. I think for us it is especially frustrating because we have been around for quite some time. The WFTDA Apprentice process has just taken us much longer than it does a lot of leagues (due to unforseen circumstances). So, we are chomping at the bit. And we really do want to blaze a trail on the scene so our ambition makes the wait that much more agonizing. It is nice to know that other leagues experience this same kind of frustration.

Also, I don't want to seem ungrateful because a lot of leagues that didn't have to give is a chance have been very gracious and free with their time and expertise. At the end of the day I do believe that most leagues want to see the sport grow and will do what they can to accomplish that goal. It just gets harder as the competition grows.

Most helpful and thoughtful DNN comment I've seen in years

... and, while I probably don't see every single comment, I surely see most of them. Trying to move up in the world? Heed this advice well!

But....

The teams are going to play their required bouts within their divisions as is logical for playoffs and then be looking elsewere, true. But, don't you think all of the teams will then be thinking of the next season and attempting to be ranked higher? Am I correct in thinking that sanctioned games don't only count or matter for the playoff seeding but towards ranking for the next season? So, ultimately, they are rewarded for playing "up" in the divisions or placements (for lack of a better term). Which means it will be harder to entice teams in a higher division to play the lowest division. Am I off base thinking this way?

Seriously,

Truly, tell me if I am wrong. It will be welcome as I am going into a panic thinking about scheduling next season!! ;-)

Yes, you are wrong

You are right about rankings affecting next season and not just playoffs. But, teams will only play "up" if they think they can win the game or at the very least keep it close. Getting blown out hurts a team's rankings even if they get blown out by Gotham. In reality, teams would want to play slightly down so that they can get a win with a decent scoring margin. What division a team is in will have no bearing whatsoever on who teams schedule (except to meet their minimum requirements) because it basically means nothing.

You're going to come in with a strength factor of 0.5 (out of 2). You may be underrated but teams at that level and even above will be unlikely to care that much about rankings since they're not anywhere near competing for playoffs. You shouldn't have trouble getting games with division 3 teams. Once you play a few of those teams and blow them out, your ranking will go up and will be more accurate, division 2 teams and maybe even low-level division 1 teams (depending on where you're ranked) would definitely schedule you.

Well now.

Thanks for the schoolin'. I will repent my former attitude and adjust properly!

What I've learned over the years...

Stab_7 wrote:

The teams are going to play their required bouts within their divisions as is logical for playoffs and then be looking elsewere, true. But, don't you think all of the teams will then be thinking of the next season and attempting to be ranked higher? Am I correct in thinking that sanctioned games don't only count or matter for the playoff seeding but towards ranking for the next season? So, ultimately, they are rewarded for playing "up" in the divisions or placements (for lack of a better term). Which means it will be harder to entice teams in a higher division to play the lowest division. Am I off base thinking this way?

One thing I've learned from entering thousands of bouts into our database each year is that those higher-ranked leagues you want to play typically book their weakest opponent as the first home bout of the season. The idea is probably that they want to play a low-risk opponent so their team can gel? On occasion they'll get upset. Usually not. Once your interleague rep picks up on this, they can strategically ASK leagues like that for an early-season bookings.

The downside is that this time of year there tends to be a fair number of horrendous blowouts. Which this year will probably have various derby pundits blaming the new rules (in past years the old rules) and using as their window of opportunity to push their proposed rules changes (or their alternate rule set). Seriously, I've been watching THAT happen for years too...

One thing with the old ranking system was that you REALLY wanted to get ranked as quickly as you possibly could. Somewhere, anywhere. Hopefully before someone realized you were any good. An unranked team that was known for being tough had a TOUGH time getting sanctioned bouts. Additionally, many newer member leagues didn't really seem to know how the ranking system worked, as they'd spend a year and change playing other unranked teams, which didn't get 'em ranked.

Everything that Southbay posted is spot-on.

new league ranking 2013

Hi, I'm looking for any help that I can get on some answers. We are currently in the WFTDA apprenticeship program with a June 2013 graduation expectation (God & WFTDA willing) and would like to be ranked as quickly as possible. I've read through all of the WFTDA info on the new rankings and I'm still unclear as to if we must play "ranked" teams in order for us to be ranked. I realize that it's more beneficial to our ranking points to play teams that are already ranked, but we do have a couple of bouts in our 2013 schedule with teams that are yet unranked. 1 of the teams is due to graduate at the same time we are and is in our schedule for Sept. Will these, as of yet, unranked teams count towards are 2 sanctioned bouts needed in division 3 to get us into the rankings? Or is it a better idea to find currently ranked teams, with a better opponent strength ranking and try to schedule with them? Also, since it won't be until June that we become full members, what will our time frame be for getting ranked in 2014?
much thanks!

Getting Ranked

You've said a lot things so I'm not sure what your priorities are. But if you what to get ranked as soon as possible, as soon as you become a full WFTDA member, play as many sanctioned games with full WFTDA members as possible, ranked or not it doesn't matter, what matters is their membership status and whether the game is sanctioned.

Of course, the higher the ranking the better, but if you're just starting out it's best to play everybody (full members) that will play you. Don't be picky.

As of now, according to WFTDA whitepaper, third division teams and new members need to get two games to be ranked. So if you play two games in June, you will be ranked in July. But I've heard rumblings that WFTDA may up the number of games for new members to be played to as high as four games. Look for announcements coming in March. If so, the above still applies, play four sanctioned games in June and you will be ranked in July.

Regional rivalries?

So I get the goal to make the whole playoff system more fair and solve the endless "if my team was in this region they would have made the playoffs"-complaints - kudos to that!
BUT... this practically eliminates the regional rivalries and that sucks :(

Now if Rat, Rose, BAD, Oly will likely play in different divisional playoffs then it just won't be as exciting to me as say Westerns 2012 and 2011. If you look at the big professional sports you'll see a lot of focus on the regional rivalries.

So there is good and bad...

Regional rivalries will persist...

...I think we might even see the rise of smaller competitive regional tournaments which will make for great spectator events, but that would never have happened before because the teams would be playing each other at regionals.

It might not -- but I think the lack of regional tournaments (which were never good local fan fodder due to the sheer amount of time required, as well as the inevitable lopsided games) gives leagues with established rivalries and strong local competition a great opportunity.

We shall see, though.

Yep

Teams can just make sure they play each other during the regular season. They could even schedule a home-and-home series every year. No reason regional teams have to only play each other in the playoffs. This would actually guarantee teams play each other every year, rather than hope they meet in the playoffs.

I hope you are right...

Lex Talionis wrote:

... regional tournaments (which were never good local fan fodder due to the sheer amount of time required, as well as the inevitable lopsided games) ...

See for me I loved the fact that I could go to ONE regional tournament and see MOST of the teams that I was seriously interested in (Rat, Oly, BAD, Rose, Jet, Denver etc.).
And I was guranteed that the regional where my local teams would play in would be at least SOMEWHAT CLOSE.

The new model means that my regional teams are DISTRIBUTED across multiple tournaments which at least 3 out of 4 will be FURTHER AWAY by definition.

Like I said: I like the fact that the new model is more "fair", I am just pointing out the drawbacks. I wonder if some sort of hybrid that stresses the regionality more would be better...

I don't think there's any

I don't think there's any need to worry about an end to regional rivalries. While certain rivalries were certainly bolstered by competition for nationals/championships there are plenty that come from several other factors. And, let's face it, people try to hype rivalries because they believe they are a draw to attendance.

And while there was a lot to like about Westerns, let's not forget that Westerns, like the rest of the regional tournaments, was fraught with blowouts. So, we don't get to see 6 of the top 10 teams in the world competing for 3 championship spots anymore but hopefully the overall quality of all of the playoffs and the championships will rise well beyond what we're used to seeing.

ELO System already tested

I'm wondering why we can't just use the ELO algorithm since it's used for many professional sporting organizations as well as for chess rankings?

blowouts

1. The lower total points in a game, the more a point matters
2. A win is a win is a win is a win - why reward a team for beating another team by 300 points rather than just by 200

I propose the following :
1. games within 5 points are considered a tie mathematically (each team awarded same points in formula)
2. games with 6+ points are treated as a small win or a loss ( ex. winning team : 2 pts, losing team 1 pt)
3. games with 30+ points are blowout (2.5 points winning team, .5 points losing team)

This way teams can exercise their benches and develop talent in blowouts- not focus on getting even more points

final thought- run the actual models!

Why can't we run the models to compare? I will set up an ELO calculator. We run last years data through my calculator and through WFTDA's model...then we compare the percentage the ranking accurately predicted the outcome.

That would give us actual numbers to compare models and their effectiveness.

Half way there

Flat Track Stats is already ELO based model, plus with predictive capabilities, and I've even seen distributions of how they do (not just how often are they wrong, but how wrong are they when they are).

In what world is a ranking

In what world is a ranking system not supposed to accurately reflect what team is ranked above another? There's a reason college football, the MLB and the chess federation use ELO...it predicts who will beat who. That's exactly what rankings are...who beats who, no matter what model you use. fTS may be ELO based, but there are many variables that affect the formula...from the bell curve used, to the k value, to game rank

I would argue research ,

I would argue research , including a few Harvard phd thesis, beat out making our own system. In fact, no other system is used for one on one competition that I know of. Sports that don't use ELO or a variant of it, are based on set schedules, and season records, not rankings

Wrong comment?

I think you replied to the wrong comment.

yes- RANKINGS ARE PREDICTIONS

I was trying to respond to this

"The point of WFTDA's system isn't to rank teams for predictive purposes"

Of course the point of WFTDA's system is to rank teams for predictive purposes. That's literally EXACTLY what rankings are. If we didn't need to predict outcomes of games that did not exist we could go off of a W-L record and playoffs. We need rankings to PREDICT who is higher ranked than another team if they HAVE NOT PLAYED EACH OTHER. Of course, the prediction is based on the outcomes of actual bouts-- but in the end-- all rankings (including DNNs, including ANYTHING that is not a W-L season) is a PREDICATION based on the data available

?

MLB uses W-L standings to determine their playoffs.
College football uses voters and a combination of computer rankings (some of which might use Elo, I'm not sure, but I seriously doubt it's all of them and the computers are only 1/3 of the ranking anyway) to determine the championship game. When they switch to a playoff, it will completely be determined by a selection committee.

I'm not sure where you're getting this idea that an Elo model is used in things that it's not? As far as I know, it's only *officially* used to rank chess which is far from a team sport.

And, WFTDA didn't invent this system, they took it from the FIFA World Rankings and adapted it.

"In what world is a ranking system not supposed to accurately reflect what team is ranked above another?"

I don't think you're understanding what I'm saying. Just because a team is ranked above another, that doesn't necessarily mean they're *supposed* to beat a team ranked lower. In W-L standings for example, if a team has had an easier schedule, they could have a better record than another team but be of lower quality than that team. Or, the team with a worse record lost a game because their star player was injured which dropped their record. Or, they just lost a game (the better team doesn't always win). Does that mean that the W-L standings aren't doing their job? Does that mean they're wrong and every pro sports organization should switch to an Elo system? No. Standings are there to measure what a team has accomplished during the season, not to try to predict future results which is what you are suggesting standings should try to do.

corrections

"Just because a team is ranked above another, that doesn't necessarily mean they're *supposed* to beat a team ranked lower." Yes it does. That is what rankings are for. That is why if a team that is ranked higher loses to a team ranked lower their RANKING CHANGES. Based on your argument there would be no need for their ranking to change if they lost.

"Does that mean that the W-L standings aren't doing their job? " Apples and oranges. W-L is STANDINGS, we are speaking of RANKINGS. They are different. Rankings are, by nature, meant to be predictive. Who do we think will beat who given that they have not played each other. Standings are standings...they are not predictive. They are based on the fact that in a given division, every team will play every other team. Derby does not have that luxury- thus the need for rankings.

"Does that mean they're wrong and every pro sports organization should switch to an Elo system? No."
No, you're right. Ideally, derby would switch to every other sport's system. But we can't- because we can't set schedules. Look for sports WITHOUT set schedules. That's where you find RANKINGS. That is where, without fail, you find ELO or a variant thereof.

A couple things

Obviously, I'm not going to convince you and that's fine but I just wanted to explain a couple things.

Rankings do change when something unexpected happens but that's because each team's season changed (not because they're suddenly a much better or much worse team). Alabama lost to Texas A&M in college football this season but they were still clearly the best team. When they lost, they didn't drop below Texas A&M. A&M didn't suddenly shoot up to #1. Alabama simply dropped below the undefeated teams, not because people thought they would lose to those undefeated teams, but because the undefeated teams were having a better season because they hadn't lost yet. If what you're saying is true, and people were only trying to predict who would win if every top 25 team faced each other then W-L record would have no bearing on college football rankings. Alabama would have stayed #1 throughout the whole season (because many many people still thought they were the best team even after their loss) but they didn't. The reason Notre Dame was #1 and Alabama was #2 in the national championship game was because Notre Dame was 12-0 and Alabama was 12-1. Alabama was favored by 10.5 points in Vegas. 60% of people picked Alabama to win in ESPN's sportsnation poll. So, if what you're saying is true and rankings are just there to predict who would win and the oddsmakers were picking Alabama and the American public was picking Alabama. Why was Notre Dame the unanimous #1 in the AP poll? Alabama didn't have a single vote. Are you saying that not a single voter agreed with the American public or Vegas oddsmakers? That seems pretty unlikely.

"That is where, without fail, you find ELO or a variant thereof." No, you don't. Where do you find this? In 1/18 of the BCS rankings (which is going away after next year) and that's it. Elo is not used to officially rank sports. Where are you getting this idea?

Two things

1. The point of WFTDA's system isn't to rank teams for predictive purposes, it's to rank teams based on what they've accomplished throughout the season. Judging the rankings based on future results is a flawed argument from the start. It would be like saying that W-L standings in the NFL shouldn't be used because someone's power rankings are more accurate at predicting outcomes.

2. You're never going to get a realistic comparison because teams will schedule based on the rankings. You can't use previous year's data because teams scheduled based on the regional vote. You can't compare Elo with upcoming data because teams will schedule based on WFTDA's model, not an Elo model.

A maximum number of games

A maximum number of games counting for your ranking could also help. If only the min to play games +2 (for instance) count, you are not discouraged to play more bouts or to play extra bouts that otherwise would cost you points.

I like this line of thought

I think the challenge is how to modify the algorithms to take it into account. Is it only the first N games, last N games or best N games? The former would be easiest to implement, but I would love to see a system that incorporated the best N games. I doubt it would be stable though.

But, why?

But, why? There's really no reason to have the playoffs count for more to begin with. There's really no reason to count 12 month old games to begin with. Just don't do those things. Problem solved.

Even if they didn't weight

Even if they didn't weight the playoffs I think the max number of games thing has potential. It would encourage teams to play more bouts since they can play teams that are ranked way high or low without fear of it hurting their rating.

But, that's part of the point

But, that's part of the point. We don't want 400, 500, 600 point blowouts so why would we encourage those games? Plus, this is already possible. There's no requirement to have every game a WFTDA team plays be sanctioned. Just don't play a sanctioned game.

Two reasons

1) We want weaker teams to get the experience of playing stronger teams. Even if that means a blowout. But it won't always mean a blowout. There are probably lots of bouts where one team would lose by 100-150 points but the current format means those teams won't agree to a bout, because there's nothing for the better team to gain.

2) Teams have done that and we've found that when the game isn't sanctioned, teams don't bring their best team, so you don't actually get the proper experience. Unsanctioned bouts = scrimmage.

First point

1) This is precisely what the system incentivises. If you want to go up in the rankings, you play a team ranked about 30 below you, and aim to blow them out. That means if you're a lower-level team looking to come up in the world, everyone is going to want to play you until they realise they're not going to kick your ass.

Once you kick a few asses and people start to realise, it's too late and you're appropriately ranked. Job done.

2) I think this will change -- it's a matter of agreement. Sometimes teams bring their best team for non-sanctioned play, sometimes they don't. If teams agree terms beforehand, I don't see this generally being a problem.

If you agree to bring your best team, then show up with one that's sub-par without a decent excuse often enough, people stop believing you. Again, job done.

On your second point

Don't need to add to what Lex said on the first.

On your second point, why do you think it would be any different in a game that would have no effect on rankings? Sure, it could be officially sanctioned so you know all the skaters are from the 20-skater roster. But, they have no incentive to play the game like they would a real game for rankings. It's basically an official roster scrimmage. Is that really better?

Fans don't want to watch games that are completely meaningless. I don't think skaters want to play games that are completely meaningless. So, why create a maximum number of meaningful games? Makes no sense.

Fair Enough

Hmmm, I guess you're right. I think I still have a little bit of the old scheduling tactics in my head I haven't shaken loose. I retract Point 2.

Initial 2013 Rankings

Quote:

The one big unknown here is how the initial set of rankings or Strength factors will be calculated. One problem with this rankings system is that all new rankings depend on the strength factor of the teams playing -- so when you start things off, you need an existing set of rankings or set Strength factors to get things moving.

The past 12 months data looks likely to feed into the initial rankings to smooth out any initial assumptions for the first release, but WFTDA has not said how the initial rankings for the calculation period will be set.

Has there been any sort of announcement to clarify this? I see on wftda.com that "The first rankings release from the WFTDA Rankings Calculator will be in March." But I can't find anything that explains how those initial rankings will be calculated.

Is one needed?

All announcements WFTDA is going to make would appear on WFTDA.com. As I understand from WFTDA's releases the initial ranking is a mixture of calculation and human judgment, so the March release will be a rather unique ranking.

With the best understanding of WFTDA's white paper that I can muster, I created my own WFTDA calculator with the following settings:

All tournaments are division 1, minimum 2 games in 12 months to be ranked, new and returning leagues start at .5 strength, used FTS game database, initial ranking was WFTDA last national ranking, 3rd quarter 2008. Ranks come out the 1st of every odd numbered month.

While this is NOT WFTDA's ranking, I gather it's not far off.

For January 1, 2013

Rank League Avg-Pts
1 Gotham 441.92
2 Denver 386.97
3 Oly 365.16
4 Bay Area 339.93
5 Windy City 324.04
6 Rat City 323.36
7 Rose 308.79
8 Philly 306.66
9 Minnesota 306.22
10 Rocky Mtn. 292.38
11 Texas 291.67
12 Detroit 290.05
13 Charm City 285.4
14 Arch Rival 278.04
15 Montreal 273.14
16 Naptown 270.56
17 Angel City 266.65
18 Atlanta 264.01
19 Kansas City 263.08
20 Boston 251.72
21 Wasatch 242.28
22 Madison 237.46
23 Steel City 233.42
24 London 232.83
25 Ohio 231.18
26 Tampa Bay 228.39
27 Houston 223.03
28 Arizona 221.82
29 Sacred 220.59
30 Carolina 216.11
31 Nashville 214.35
32 Brewcity 211.07
33 Jet City 207.82
34 Chicago Outfit 207.34
35 DC 202.24
36 No Coast 195.65
37 Cincinnati 190.77
38 Dutchland 180.07
39 Victorian 178.7
40 Omaha 177.81
41 Jacksonville 173.13
42 Mid Iowa 172.3
43 Bleeding Heartland 171.88
44 Terminal City 169.65
45 Grand Raggidy 168.3
46 Sac City 166.36
47 Tri-City 159.32
48 Columbia 157.9
49 Maine 157.44
50 Toronto 154.83
51 Burning River 154.59
52 Blue Ridge Rollers 153.56
53 Paper Valley 150.14
54 Santa Cruz 149.4
55 Emerald City 147.92
56 New Hampshire 147.22
57 Rideau Valley 145.66
58 Tallahassee 145.58
59 Tucson 141.44
60 Dallas 140.79
61 Sin City 136.57
62 Dominion 132.38
63 Silicon Valley 132.33
64 Humboldt 129.35
65 Killamazoo 128.14
66 Queen City 126.74
67 Old Capitol City 124.45
68 Pikes Peak 122.34
69 Fargo Moorhead 121.77
70 Garden State 121.53
71 Ft. Wayne 119.94
72 River City 119.21
73 Duke 117.62
74 Green Mt. 117.1
75 Sioux City 116.56
76 Bear City 113.68
77 Suburbia 113.56
78 Babe City 113.37
79 Providence 108.53
80 North Star 108.06
81 Gold Coast 105.77
82 Okla. Victory 102.85
83 Ithaca 102.09
84 Long Island 101.57
85 Demolition 98.74
86 Bellingham 97.29
87 Twin City Derby Gi... 97.01
88 Connecticut 95.71
89 NEO 94.79
90 Memphis 92.34
91 Glasgow 92
92 Cape Fear 91.16
93 Central NY 90.91
94 Green Country 89.61
95 Des Moines 89.47
96 Lehigh Valley 88.24
97 Lowcountry 85.67
98 Dixie 83.92
99 Sioux Falls 83.75
100 Roc City 83.59
101 Treasure Valley 82.27
102 Slaughter County 81.15
103 Hard Knox 80.13
104 Black-n-Bluegrass 76.43
105 Springfield 75.33
106 Charlotte 72.19
107 Assassination 71.36
108 Big Easy 70.11
109 Rage City 69.27
110 Crime City 66.06
111 Jersey Shore 64.99
112 Dockyard 60.5
113 Central Coast 59.78
114 Hudson Valley 58.13
115 New River Valley 57.86
116 ICT 57.35
117 Central City 56.84
118 Lava City 56.8
119 Chattanooga 55.82
120 FoCo 54.69
121 Helsinki 54.35
122 Derby City 53.52
123 St. Chux 52.73
124 Junction City 52.69
125 Slaughterhouse 52.47
126 Choice City 51.01
127 Auld Reekie 48.85
128 Little City 46.67
129 NW Arkansas 45.97
130 Rocktown 44.11
131 Alamo City 43.54
132 Hammer City 40.27
133 CoMo 39.72
134 Harrisburg 39.04
135 Pacific 37.92
136 Tragic City 36.24
137 Ark Valley 35.07
138 Pueblo 34.27
139 Fox City 30.67
140 Oklahoma City 30.09
141 So Ill 28.86
142 Spindletop 24.82
143 Red Stick 24.04
144 Castle Rock 23.54
145 Rockford 20.34
146 Fairbanks 17.09

Yes, one is needed...

At least I would certainly think so.

Surely they didn't use 2008 rankings (or maybe they did).

I realize it may not make THAT big of a difference (your rankings seem pretty reasonable, at least at the top, considering what you used as a starting point), but it still seems like something the WFTDA should make public just so people can understand where the numbers come from.

A want

I presume March will be THE initial ranking. I've heard that WFTDA started the calculator with 2010 data. Since games time out after a year, if you run the calculator for three years you get to something like the ranking I've posted, really no matter where the teams actually began from.