The All-Knowing Derbytron is now on derbytron.com

This site will no longer be updated. Go to derbytron.com for all the fresh Derbytron content.

11.17.2009

A look back and a look to the future

Roller derby is obviously a young sport and with that comes the need for dramatic change from year to year. You try one thing, try something else, see what works. There's really no way around it. So, as I have you here, I might as well take this opportunity to get on my soapbox, as small as it may be, to point to a couple issues that I would like to see changed or at least looked at for the years to come.

A major issue with regionals was the setup of the second round. The WFTDA used what they called "equitable seeding" for this round of the tournament, something I had never heard of. This, in theory, is supposed to give you closely matched games as the #1 seed played #6, the #2 seed played the winner of #7 and #10, the #3 seed played the winner of #8 and #9, and #4 played #5. Most single-elimination tournaments like this would use advantage seeding, so #1 would play the winner of #8 and #9, #2 would play the winner of #7 and #10, #3 would play #6, and #4 would play #5. Only two games are different but it would make a huge change in my opinion. In the two games I have a difference with (as played: #1 vs #6, #3 vs #8/#9) the average margin of victory for the higher seeded team was 123.6 with only one lower seeded team winning (Rocky Mountain, who was clearly seeded incorrectly, over Rose City) and only one other game under a 100 point margin. Sure, a #1 vs #8 game will probably be a blowout but is there really that big of a difference between watching a 123.6 point margin game and a 200 point margin game? Is that slightly closer margin worth it rather than being able to see a #3 vs #6 matchup which would more than likely be a very good game? It's unfair to the fans and it's unfair to the teams, especially the #6 seed whose only shot at nationals is to upset a #1. Since a #6 seed is much more likely to be a nationals-caliber team than a #8 seed (that's why they're seeded higher), shouldn't they get a shot at a #3 seed rather than a #1?

Another issue with regionals that should be looked at is the number of teams invited. No #9 or #10 seed made it past the second round and the only team that was even competitive in the second round was No Coast who lost 127-110 to Tampa Bay. Aren't regionals supposed to be deciding who goes to nationals? If this is the case, why include teams that have no chance of competing for those three spots. Really, though, what bothers me about the 10-team structure aren't that those lower ranked teams make it but that three teams get a bye. It's one thing to earn a bye like in the NFL playoffs, the teams getting byes won more games than the other teams in their conference. They earned a bye. The WFTDA votes teams into a bye which doesn't sit well with me. Those teams may not have earned that bye but are given the advantage of rest over the teams they will be playing. Knock the number of teams making regionals down to eight and you fix these problems and you make regionals something a team really has to go out and earn.

Which brings me to the regular season. There isn't one. The only requirement of making regionals is a team must play two WFTDA-sanctioned games. I don't like this at all. Here is a theoretical example of why this is wrong.

Let's say the entire Texecutioners team retires. Every single one. The voters would likely not vote them number 1 but would definitely put them somewhere between #2 to #5 in the South Central. I mean, it's Texas right? They have players on their local teams with chainsaws for arms. But, what if they really suck? What if they go out and play two very low ranked teams to qualify for regionals and win those games by around 100 points? There would be no reason to rank them lower but they really haven't proven themselves worthy of playing in regionals, have they? They would finish the season ranked high enough to go but that's completely based on the last season, not the season they're currently playing in.

Why use voters (who may have biases and more than likely haven't seen most of the teams they're voting on play) to determine anything when you can have a short regular season to qualify for regionals? What I propose is this: Each region is broken down into 5 divisions (assuming 10 teams would still be going to each regional) with 4 or 5 teams making up each division (preferably 4). Each team would be required to play a three game regular season (only one more than now). Each division would play a round-robin (each team in a 5-team division would only play 3 of the 4 other teams). The top two finishers, by record, in each division would qualify for regionals.

Divisions would be set up with two priorities: proximity and an even spread of team rank. Essentially, the WFTDA would want to have teams be close to each other for obvious reasons: less travel expense, ease of scheduling games, set up regional rivalries. The WFTDA would also want these divisions to not be top-heavy or bottom-heavy as you may end up leaving out a team that deserves to go to regionals or invite a team that doesn't. The goal would be to have a top-team, two middle ranked teams, and a low ranked team.

The divisions would last for two years and then be redistributed after that. This way each team would get two home games in one season and one home game in the other. The teams scheduled to play each other could set up the game to be played at any time between January and June.

Ties aren't probable but are definitely possible. This is where other WFTDA sanctioned games come into play. Each team starts with 0 points. If you beat a team that finishes 3-0 in their division, you get +3, 2-1 = +2, 1-2 +1, and 0-3 +0. If you lose to a team that finishes 3-0 in their division, you get -0, 2-1 = -1, 1-2 = -2, and 0-3 = -3. This also helps determine seeding. Seeding in regionals would be decided by record first and then points. A team does not have to play any non-division games but it's obviously in their best interest so they can get tiebreaker points and a higher seed.

There you a go, a beautiful regular season where every game matters and teams earn their way to regionals and earn their seeds. Imagine how intense those games between two 1-1 teams at the end of the season would be to determine who qualifies for regionals. Mmm. I love it.

One last thing before I go. I have to say first that the stop play, strollerskating, whatever you want to call it used by Denver does not bother me. However, I do have to point out what I think have been incorrect calls  by the refs throughout the tournament and regular season. What I'm looking at is clockwise skating by a blocker after knocking a jammer out of bounds thus either forcing a track cut or forcing the jammer to skate clockwise to come inbounds behind the blocker.

From the WFTDA rulebook:
5.1.1 Blocking is any movement on the track designed to knock the opponent down or out of
bounds or to impede the opponent’s speed or movement through the pack.
 6.6 Skaters must not skate in the opposite direction of the pack (clockwise) when executing a block.
6.9.5 A skater may not initiate contact with an opponent who is completely outside the track
boundary.
From my reading of the rules, a block does not need contact to be considered a block. So, any movement that is meant to impede an opponent's speed is considered a block. Clearly, a blocker moving clockwise after knocking a jammer out of bounds is meant to impede their opponent's speed. If I'm reading the rules correctly, that would be considered a block, thus making it illegal to move clockwise in that instance. Notice, the rulebook does not state that it is illegal to block an out of bounds skater, but rather that it is illegal to make contact with an out of bounds skater. That is an important distinction as it allows a blocker to stop or move forward very slowly after knocking a jammer out of bounds but not clockwise as a skater cannot move clockwise while engaged in a block.

There were a few teams that used this strategy throughout the tournament and regular season, not just Denver. I believe it has been called incorrectly all season and needs to be corrected.

If you notice any holes or have any comments about this wide-ranging post, feel free to make a comment.

November 17 Analysis - 2009 Final Rankings Edition

By my count, there only 5 scheduled WFTDA sanctioned games for the rest of the year so nothing major is going to change. What better time to present the final rankings than right after nationals? The Derbytron will be back sometime in early 2010 bigger and badder than ever.

Anyway, let's get to Oly. Talk about bigger and badder, no one has yet to combine speed and teamwork as this group of former speed skaters has been able to do. They just out-athleticked (athleticed? athleted?) everybody on their way to the national title. There are no arguments against 11-0 and they're the main reason that I am going to have to dig in really hard to rework the derbytron equation. Obviously, the current rankings are a little off. Oly has beaten the current #1, #3, #6 (twice), #8, #9 etc, all in pretty convincing fashion. There are no teams left that even have only one loss (except for teams that have played two or less games).

The rest of the rankings do not appear to be too far off. Perhaps Charm City is a little high (although, this is only based on one somewhat lackluster tournament). Also an argument could be made that pretty much every Western team is ranked too low.

Every incorrect prediction in nationals was due to a Western region team being ranked too low. The reasons for the lower ranked Western region seems to be two-fold. The top teams improved dramatically throughout the year. Oly came out winning but won two of their first few games by single digits. Oly beat Denver at home in April by two points and then won in Philly by 87. Denver also improved consistently throughout the year and obviously Rocky Mountain made a giant leap after picking up the Pikes Peak skaters. The issue with this is that the games in January count just as much as the games in September so a poor start to they year could hurt a team's rating towards the end of the season.

The other factor is a lack of inter-region games. The following is a list of the teams that played at western regionals and the number of teams they played outside of their region before nationals: Oly (0), Rocky Mountain (2), Denver (2), Rat City (5), Bay Area (2), Duke City (4), Rose City (2), Angel City (0), Tucson (0), and Pikes Peak (3). Rocky Mountain was the only nationals team representing the Western region to have played another nationals team during the regular season (Boston). Obviously this is not a lot of information to go on when you compare those numbers to those of the other teams at nationals: Philly (3), Gotham (1), Boston (7), Windy City (5), Madison (2), Detroit (5), Texas (4), Kansas City (5), and Houston (1). So, with one less team, the other regions played 13 more games inter-regionally and they all played at least one. There is a clear difference there and one of the main regions the West was underrated.

These are clearly some of the unique issues roller derby presents. Every major American team sport plays a non-division schedule. The lowest number of non-division games that I know of is college football with three although most conferences play four.

Another issue that I will be looking at are what I'm calling trash games. In other words, sanctioned games that don't include a large number of the top players of one of the teams involved. Two clear examples of this were the Charm City/Philly game as well as the Madison/Detroit game. The two games skewed the rankings dramatically despite clearly being unimportant based on other results.

These are the issues that will be tackled in the offseason so stay tuned for the 2010 rankings.

2009 was the first year of the Derbytron Rankings and as such was a very successful start. Derbytron finished the year correctly predicting 84.4% of official games. Roller Derby is a sport, and a great sport at that, which means it's unpredictable, as it should be, so anything over 90% is pretty close to impossible to achieve. I'm still hoping to get there, though.

11.16.2009

2009 Final Rankings

 Team Rating2009 RecordBase RatingSOSGames Used2009 Games
1Gotham 0.7547-20.7400.625159
2Oly+50.74111-00.7430.5541111
3Texas+10.70211-20.6870.5612013
4Philly-20.6949-20.6820.6071711
5Windy City-20.69410-20.7110.6101812
6Denver-10.69011-30.6720.5501914
7Charm City-10.64712-50.6430.5832217
8Rocky Mountain+50.6267-60.5410.5701813
9Rose City+20.6235-30.5930.553138
10Kansas City-20.6148-40.6080.5641712
11Boston-10.61111-80.5750.5562419
12Detroit-30.6058-50.6030.5811813
13Madison-10.5997-40.5900.562511
14Rat City 0.5895-70.5390.6041612
15Steel City 0.5649-30.6030.4531312
16Cincinnati 0.5589-40.5670.5321813
17Bay Area 0.5553-50.5170.552138
18North Star+10.5455-30.5910.52088
19Arch Rival+10.5376-70.5130.5021613
20Harrisburg Area+10.5363-20.5670.48085
21Duke City+10.5364-80.4910.5631712
22Houston-40.5185-30.5120.504138
23Brewcity+10.5185-40.4910.542109
24Atlanta-10.5188-80.5300.4851916
25Carolina 0.5143-80.4430.5881711
26Providence 0.5035-40.4920.511129
27Sioux Falls 0.5003-20.5620.48355
28Minnesota 0.4935-30.5230.502138
29Ohio+10.4906-30.4950.472119
30Burning River-10.4893-50.4900.459118
31Tampa Bay+10.4794-50.4800.504119
32Dallas-10.4777-30.5490.4771510
33Nashville 0.4708-40.5230.4371412
34Naptown 0.4684-40.4670.43488
35Grand Raggidy 0.4503-80.4090.4921711
36Tallahassee 0.4411-10.5060.35422
37Omaha 0.4342-40.4270.47276
38Maine 0.4333-40.4500.49287
39Montreal 0.4291-60.3760.49677
40Fort Wayne 0.4274-60.4120.4601210
41Pikes Peak 0.4184-70.4410.4941611
42DC 0.3944-50.4350.458129
43Angel City 0.3915-40.4870.456109
44Memphis 0.3876-40.4520.3711110
45Tucson 0.3862-40.4150.515116
46Dutchland 0.3813-20.3740.38565
47Sacred City+10.3743-30.3430.43896
48Connecticut+10.3633-70.3570.4491110
49No Coast-20.3624-80.3640.4431612
50Suburbia 0.3591-10.3590.32422
51Northwest Arkansas 0.3502-40.3610.427106
52West Texas 0.3282-50.3870.417107
53Hard Knox 0.2961-50.3020.38766
54Green Country 0.2893-40.3570.33297
55Dixie 0.2561-50.2520.43296
56Salt City 0.2541-10.3260.35362
57Assassination City 0.2432-40.3070.36476
58Arizona 0.2371-30.2690.42154
59Bleeding Heartland 0.2171-20.3020.38433
60Dominion 0.2110-80.2270.442118
61Sin City+10.1950-30.2100.31243
62Big Easy-10.1940-60.2460.38976
63Oklahoma 0.1841-40.2280.40975
64Alamo City 0.1660-30.2470.32073
65Slaughter County 0.1630-20.2780.51522
66Gem City 0.0910-30.2450.48533
67Long Island 0.0320-30.2460.41143
         
 *Unofficial Team