|
|
1 |
2013 1-0
2012 17-0 |
5/11 @ Philly |
Nothing to see here. Move along. |
|
|
3 |
2013 0-0
2012 12-1 |
6/9 vs London |
Still only have one WFTDA game scheduled in the ranking period. USARS-only rumblings continue. No action. no movement. |
|
|
2 |
2013 0-0
2012 13-4 |
5/5 @ Minnesota |
The MHC hasn't faced anyone to stretch them yet in 2013 -- MNRG's visit should give us an idea of how they adjust to Juska's exit. |
|
|
4 |
2013 1-0
2012 11-8 |
5/11 vs Rocky Mountain |
Still only the MNRG drubbing to go on; the Rocky bout on May 11 should be a good litmus test for both. |
|
|
4 |
2013 2-0
2012 9-4 |
5/11 vs Montreal |
Bay Area hold firm at 5. |
|
|
9 |
2013 3-0
2012 11-6 |
5/11 vs Gotham |
Philly climb based on Windy's fall. Their Boston game was closer than expected, but is looking like an outlier for now. |
|
|
18 |
2013 8-0
2012 7-7 |
5/4 @ Rat City |
Climb into the top ten off the back of a stunning first half and eventual win over Rose City. |
|
|
7 |
2013 0-0
2012 12-4 |
5/4 @ Angel City |
Rat's season finally gets underway this month. |
|
|
10 |
2013 2-2
2012 9-5 |
5/17 vs Victorian |
Rose climb after at least partly answering the jammer question against Bay Area... before opening it up again after Angel City. |
|
|
12 |
2013 1-0
2012 9-5 |
5/4 vs Minnesota |
No real action; get things underway properly this weekend. |
|
|
6 |
2013 4-0
2012 9-7 |
TBD |
Windy slide to 11 based on their London and Montreal results. |
|
|
24 |
2013 0-1
2012 7-3 |
6/1 @ Terminal City |
London slips a place, but climbs above Minnesota based on the closeness of their season-opener against Windy City. |
|
|
8 |
2013 0-2
2012 9-11 |
5/4 @ Rocky Mountain |
Minnesota slip by virtue of teams rising above them; they will have a chance to correct this slide against Denver and Rocky. |
|
|
17 |
2013 6-1
2012 14-6 |
5/11 @ Bay Area |
Slip one on other's action. |
|
|
13 |
2013 2-2
2012 8-5 |
5/11 vs Columbia |
Charm City performed to expectation in April. |
|
|
16 |
2013 5-1
2012 12-8 |
5/18 vs Arch Rival |
2-0 for April, but against teams they were expected to get the better of. |
|
|
15 |
2013 2-0
2012 15-5 |
5/18 vs Ohio |
Two 400+ point wins opened the Dirty South Derby Girls' account in 2013. Ohio this month should prove a sterner test. |
|
|
21 |
2013 0-4
2012 8-5 |
5/4 vs Gotham |
Again got closer to Philly than the rankings might suggest. Climb on place on Kansas City's demise. |
|
|
23 |
2013 7-1
2012 21-4 |
5/4 @ Arch Rival |
The win over Steel and loss to Charm keeps Ohio pegged between them. Four May games with teams ranked around them will be very interesting to see. |
|
|
27 |
2013 3-2
2012 5-8 |
5/17 @ Wasatch |
A big win over Madison comes with a re-evaluation of Angel's form, and sees the Sacrificers crack the power rankings. |
|
|
11 |
2013 2-1
2012 9-4 |
5/10 vs Tallahassee |
Only action in April was expected loss to Naptown. |
|
|
22 |
2013 2-3
2012 10-8 |
5/4 @ Brewcity |
Didn't get as close to Ohio as some expected, but did hammer DC. |
|
|
20 |
2013 2-2
2012 9-3 |
5/17 @ Sacred City |
Wasatch climb a place based on the fall of Kansas City. |
|
|
33 |
2013 8-0
2012 17-5 |
5/4 @ Pikes Peak |
...as do No Coast. |
|
|
13 |
2013 3-1
2012 11-5 |
5/4 @ Ohio |
Arch Rival's win over Kansas City is enough to see them sneak back into the rankings after a month away. |
Comments
Atlanta's next game
I sent a submission/update last week...
Atlanta's next game is against Tampa in a closed bout on May 17th.
Then Tampa plays Ohio May 18th (closed bout) immediately followed by Atlanta v. Ohio (public bout).
I think there will be text casts for those games.
Santa Cruz!
Other teams who may well be on a trajectory towards future rankings fame, fortune and riches include a cluster of teams around what was once the West: Sin City, Sac City and Duke City. Dust Devil was kind to all of them, and while it's hard to know exactly where they are in relation to the rest of the derby world at the moment, we think they're definitely heading in the right direction and will be keeping a very close eye on their games.
You left out Santa Cruz.
There's always one...
...that doesn't make it from the notes into the actual words. Fixed now.
POP! POP!
For those who understand, no explanation is necessary.
Can we fix the MNRG W/L, please?
There will be two more games this weekend, and another two in a month at Brewhaha, but as of the morning of May 4, Minnesota is 0-1, not 0-2. The system still seems to be counting a bout against Your Mom back in January.
Yes, but...
Not today. The fix is a non-trivial piece of coding that only Gnosis can do, and it's being rolled into a bigger project he's actively working on right now. Fruits of that labor will be live in a single-digit number of weeks.
Our sincere apologies to Minnesota! We hope you'll bear with us. New goodness coming.
I'm just curious...
...why did the system ever count non-sanctioned bouts in the first place?
two reasons. three. three reasons.
1) Four years ago, when we first built the Power Rankings tools, at least a couple teams that weren't WFTDA members were clearly capable of contending -- particularly San Diego (which spent several months ranked) and LADD.
2) We didn't have the foresight to imagine top men's and women's teams meeting up.
3) Gnosis' time was much less pressed, so the idea that he'd need to sink some of it into tweaking a W/L from time to time didn't seem like a huge thing. To be frank, it'd probably take him less than half an hour, but we don't want to take a second of his attention away from the bigger project, which will include a fix for this in its scope and will be complete in a not horribly large amount of time.
Triangulations?
I'm not sure I follow the triangulations that bring Rose up 3 places, but keeps Charm City ranked at 14, when Charm only lost by a single jam against Angel City.
Does DNN really think Charm would lose to Montreal, Minnesota, London and Windy City? Teams that Charm beat (directly or indirectly, i.e. tournament) the last time they met?
also confusing
Is windy city sliding 5 places while winning, while rose loses and jumps up...hmm
Confusion shared
Indeed. How does Rose lose to a lower ranked team and climb three spots?
War of the Roses
Indeed. How does Rose lose to a lower ranked team and climb three spots?
Well, last month Rose also lost to #5 Bay Area by a mere 29 points. Rose rose based on that result, but Angel City rose even higher by beating Rose.
Regarding Charm City, I'd guess that their big loss against #6 Philly in March has something to do with keeping them down, but I do agree that they should perhaps get at least a small bump up from the close Angel City loss.
That doesn't quite justify it
Well, last month Rose also lost to #5 Bay Area by a mere 29 points. Rose rose based on that result, but Angel City rose even higher by beating Rose.
Regarding Charm City, I'd guess that their big loss against #6 Philly in March has something to do with keeping them down, but I do agree that they should perhaps get at least a small bump up from the close Angel City loss.
So what's the argument against dropping Bay Area down to Rose's rank from that result? Then Angel wouldn't have also risen as high, and the discrepancy between Angel-Charm and Angel-Rose would've been less.
Has DNN abandoned the "Who would beat whom" metric in favor of a "What would sort of fit but have the least movement" metric"?
That's a fair point
So what's the argument against dropping Bay Area down to Rose's rank from that result? Then Angel wouldn't have also risen as high, and the discrepancy between Angel-Charm and Angel-Rose would've been less.
Has DNN abandoned the "Who would beat whom" metric in favor of a "What would sort of fit but have the least movement" metric"?
The only useful data point BAD has this year is the fact that they beat Sacred by about 100 points more than Angel City did in March. But that only tells us that BAD should be ranked several spots above Angel. It doesn't say anything about where they should be ranked relative to other teams. However, the BAD/Sacred score is very similar to their game last October, meaning neither team has shifted, at least relative to each other. But the Angel/Sacred game last September resulted in a 1pt win for Sacred, which lends evidence to the notion that Angel should move up instead of BAD moving down. It's a somewhat tenuous bit of data crunching, but there you go.
Anyway, the close Angel/Rat result means that these teams are possibly ranked correctly after you swap the Angel and Rat rankings. But that's a hindsight observation and not something the DNN guys were able to use for these rankings.
Rose, Rose, Rose
A significant part of the reason Rose had dropped as low as they had was because of the massive question over their jammer rotation.
They showed last year that without White Flight and Soulfearic playing, they were a much, much weaker team.
The addition of Licker*N*Split to their rotation, and specifically how she stood up against Bay Area, served to answer that question, and lift them back up towards Bay Area.
still
Still doesn't explain how a team can lose two bouts, and move ahead of a team that won two bouts. I get the "level"* of team lost to/beat, but a win is a win. Teams are constantly over and under achieving just about every bout. So just because rose barely lost to bad and windy barely beat some lower ranked teams, doesn't mean a lot in sports. A win is a win and a loss is a loss.
*Level is a relative term. Top 25 is to 25. After Gotham, most of the teams ranked 2-25 aren't all that far apart in terms of talent.
What you are describing...
...is a league table.
This is not a league table.
This is Power Rankings; Outrage is understandable and expected. There's a link at the top to the FAQ you might care to peruse so that it can be appropriately directed.
read
Faq have been read. Is there really much of a difference between power rankings and a league table? I guess I'm using major sports to determine rankings. Say in college basketball, if major conference team A is 2-2 and major conference team B is 4-0, team B is ranked higher in any sort of list made, even if team A lost their 2 games close and team B only squeaked out a couple victories. I guess that's my problem. I'm using my common sense and not thinking about they way DNN would rank things. I guess also when thinking on if rose city played windy city on a neutral track who would win a best of series, I think windy takes it. I just think they're the better team.
or
DO YOU JUST HATE NORTH CENTRAL TEAMS!!!!!!!!! That's probably what it is. North central racists.
The main issue is that
The main issue is that leagues pick their own games, as opposed to a season where teams play a variety of teams. Also, 2-25 are not remotely close. The difference in point spreads is hundreds of points. Peruse the DNN scores page and that becomes plainly obvious. A win is completely meaningless when a good team plays one that is not good. It's just as meaningless when two not good teams teams play each other.
DNN power rankings have an 80 percent track record of successfully predicting who will win a game, and that's a lot better accuracy than you'd get by ranking teams other than Gotham by wins only.
Up Vote!
Someday, I'm told, this will work.
yeah you're right
I guess I was a little ambitious in my comparing of teams. Yes you're right. There are clear hierarchies in roller derby. I guess my point was it's all arbitrary, teams always over and under achieve which is the reasoning win loss records are used in sports power rankings when comparing two similar teams, and I've just never seen a team win and drop in standings and a team lose and go up. It's strange to me.
it takes some getting used to.
The counterintuitive movement you're seeing is a direct symptom of derby's current "make your own schedule" approach. If derby adopted meaningful divisioning with balanced schedules within divisions/conferences (balanced schedule = each team plays every other team in their competitive grouping during a season), a simple comparison of W-L records would suffice.
That's just far from the case today, and that's why DNN Power Rankings continue to be based on a simple question: if these two teams play tomorrow, which one do we think will win?
I've been testing offering a
I've been testing offering a point spread prediction on games based on triangulation the way rankings are done. So it gives a team a handicap, an expectation in a foregone conclusion. The idea is, if your team loses by more than the predicted outcome, they've underperformed. If they lose by less, you've beat expectations.
It would put blowouts in perspective, and give you a way of looking at games that are lopsided, and still have something to root for as a fan. I think DNN has more data points to work from than we have in banked track, and it would make the sport more interesting.
I also proved how too little data can be misleading. If your fulcrum is a team that performed below expectations for any number of reasons but you had nothing to gauge expectations on that fulcrum, your prediction could be way off. My loss prediction was easy, but I was off on the spread by over 100 points (lower). I also said there were some ludicrously serious flaws with the methodology.
But who could be unhappy when the underdog does better than expected? It would be interesting to see this applied on a larger scale and see how close it comes to real results.
Which is fair enough
And a view you are totally entitled to.
Last month the rankings agreed with you; having seen both teams in action this year, they no longer do.
Would love to see that match-up, though. Would be one hell of a bout.
the missing piece
Teams aren't static entities. Gauging their future competitive prospects is, essentially, trying to predict a moving target. In some cases, such as Angel City, a deep dig at their recent results, as well as recent results of their opponents and their opponents' opponents, starts to suggest that the team in question is getting better every game.
It's a more complicated analytical approach than simple triangulation, and it produces rankings that are sometimes counterintuitive at first blush, but we've found our judgment about "team trajectory" to be an important element in successful predictions. And, it cuts both ways... sometimes the trajectory becomes very clearly downward rather than upward.
thanks guys
You know what I think my main problem is? I don't get to watch as much derby as you guys do.
This is going to be a
This is going to be a difficult season to judge anything based on past performance, just due to injuries of key players. Most of them should be fully competitive by playoffs. But there are a lot of variables there.
And of course, you might as well cast the oracle bones to predict what Oly is going to do.
Arch Rivals
They seem awfully low to me, especially with where they are ranked in wftda's system. Their only loss a one pointer on the road at no coast?
Only loss this year, perhaps
Our judgment to date has also been shaped by their performance in last year's playoffs, in which they struggled pretty severely over teams they were favored to beat. Clearly, they're demonstrating the incorrectness of that judgment (and helping us see No Coast's results in a new light in the bargain).
COURTEOUSLY EXPRESSED, IT'S
COURTEOUSLY EXPRESSED, IT'S AN OUTRAGE I TELL YOU!!!
Houston
Wasatch lost to Rose by 160, Rose lost to Angel by 46, and Houston lost to Angel by 169. Not necessarily saying that they should replace Wasatch, but I'd definitely put them in consideration before the "Duke-Santa Cruz" range of teams, especially after their performance over Duke this weekend (a 261pt victory).
We hear you
That one is *definitely* on the radar.
the Duke City data point helps a lot
Prior to this weekend, Houston's other recent results were hard to contextualize. Much more clear now (though I'm also curious to investigate whether Duke City took a short lineup to this event).
Yes we were a little short
Yes we were a little short this weekend. However Houston is looking amazing and I see them moving up as we see them play higher ranked teams.
Houston
Can I "like" Muffin's status? Jus sayin!