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

F***ing Rankings and Divisions. How Do They Work? S-curve Seeding for 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.

What factors go into deciding teams’ rankings, and what sort of games will this encourage? Games against teams relatively close to your level are encouraged, and games at playoffs and championships carry more weight for overall ranking calculations, and huge blowouts against lower-ranked teams don’t help anyone.

How does the calculator achieve this? Well…

“Teams are awarded points for each game played based on a formula that incorporates opponent strength, points scored, and the type of game played.”

This ranking system is actually very similar to the one used by FIFA–that’s the Fédération Internationale de Football Association, or the international governing body for soccer–for international teams that you can read about here should you so wish.

There are three major differences, however. Firstly, the rating increase will be averaged out over the course of the season–so it doesn’t matter if you’ve played 10 games or 20, how well you did in those games is what matters, not how many you’ve played. Secondly, points won’t be awarded purely on the basis of who won, as with FIFA–but three points will be split between the teams based on the scoring. Finally, only the past 12 months data counts–a team’s rating will always be the average number of ranking points they have accumulated per game in the 12 months prior to the current ranking period’s cut-off date.

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.

The points earned by each team are calculated by the score ratio, the opponent’s strength, and the game weight, and increased by a factor of 100 to make things look a little nicer on the page and keep points accumulated to whole numbers. Laid out, that is:

(Win/Loss Factor) × (Opponent Strength) × (Game Weight) × 100

“Each WFTDA team will be assigned a Strength Factor based on their most recent ranking.”

To understand how that works, we need to understand the components. So let’s take three examples to see how this would work.

The Win/Loss Factor is simply the score ratio multiplied by three–rather than it simply being three points for a win as is standard in soccer, the teams share three points depending on how close the bout was.

Let’s take the current FTS rankings (as shown before) as the WFTDA rankings (just for ease of maths), and some real-life results from last year and see how those hypothetical games would change the rankings if played today.

The Opponent Strength factor is probably the most complicated element, and is based on a team’s current ranking and the number of teams currently ranked, down to a minimum of 0.5.

How would a really close game between two closely matched teams affect ratings? Let’s imagine that there was a 155-155 between Windy City and Minnesota played today, using rankings as above.

The first thing we need to do is calculate the two team’s Strength factors. Windy’s would be as follows: (172 – 6) ÷ (172/2) = 1.93; for Minnesota it would have been (172 – 8) ÷ (172/2) = 1.91. This means the ranking changes would have been:

(Score Ratio × 3) × (Opponent Strength) × (Game Weight) × 100
For Windy: (155/310 × 3) × 1.91 × 1.0 × 100 = 287 points
For Minnesota: (155/310 × 3) × 1.93 × 1.0 × 100 = 290 points

So honours even there, as appropriate — but as the lower-ranked team (just), Minnesota gets slightly more points for the game.

How about a comfortable win at playoffs? Let’s imagine Montreal beat DC 252-112 in the fifth-place game at playoffs, and see how ranking points would be distributed. The main difference here is that this is a playoffs consolation bracket game, and so has a higher-than-usual rating. The second inset image shows how ranking weights are calculated for playoff games.

Montreal’s Strength factor comes out as 1.83; DC’s at 1.63.

(Score Ratio × 3) × (Opponent Strength) × (Game Weight) × 100
For Montreal: (252/367 × 3) × (1.63) × (1.25) × 100 = 420 points
For DC: (112/367 × 3) × (1.83) × (1.25) × 100 = 209 points

So Montreal get a whole tonne of points for this game–the magnitude of their victory combined with the strength of the opposition and the significance of the game all add up to mean this would give them a pretty solid platform to work from for the following year.

Finally, let’s look at an utter blowout of a lowly ranked team. Let’s assume that London played Auld Reekie, and won 667-18. London’s strength factor is 1.86 based on their ranking; Auld Reekie’s would be 0.40 based on theirs, but the ranking calculator sets an automatic lower bound of 0.5 on this factor. This also serves to make rankings calculations for new teams easier; before a team is eligible for ranking, it will be automatically given a Strength factor of 0.5 in the games that serve to calculate their ranking.

(Score Ratio × 3) × (Opponent Strength) × (Game Weight) × 100
For London: (667/685 × 3) × 0.5 × 1.0 × 100 = 146 points
For Auld Reekie: (18/685 × 3) × 1.86 × 1.0 × 100 = 15 points

In this case, neither team really wins in terms of rankings calculations. The 146 points earned by London will pull down their average for the year if the rest of their season comprises close games against closely-matched opponents–a tie with an equally ranked team would earn them 279 points in the regular season. Auld Reekie wouldn’t gain anything in terms of ranking points either–a tie with an equally ranked opponent would earn them 75 points, so this result would serve to drag down their average too.

“Member Confirmation”

Finally, WFTDA has specified that during the initial implementation period of this calculator, each set of rankings (which will be every two months) will have to be ratified by a simple majority vote.

If the rankings are not approved by a simple majority, individual leagues who refused to ratify that period’s vote will be asked why; if 51% or more of of the voting member leagues cite the same problems, the WFTDA rankings committee will propose a new ranking that addresses that problem. That new ranking will then be submitted to the membership for ratification, and the process begins again.

If you want to look into this in even more detail, WFTDA has released two documents outlining Divisioning and Ranking (both links go directly to the PDFs for download).

Lex Talionis

Lex is the Content Director for Derby News Network.