Welcome to the IQA website

Login or register to take part in discussions in the forums and comment on news items.

Login

Lost your password?

Not a member yet? Sign Up!

announcements, news, slider

New Rankings Algorithm

by Dan Panzarella | March 14th, 2012

Dear Quidkids,

The day you may or may not have been dreading is finally here. IQA Rankings are starting up again, starting… now! You can go see them at the Standings page. Some of you were not expecting rankings to happen this half-season, since it is a shortened season without a World Cup in it, but that sounded to us like a great testbed for trying out a new way to rank teams, without immediate changes to who might qualify for what. These standings will reset with the new full season in the Summer or Fall.

These next few months, though, are a great time to still enter your games since November, and see how they affect your rank. We now have an open-ranking system, meaning that we have both an in-depth explanation of how the algorithm works (see below) and even a step-by-step technical one for those math-inclined. The actual code that powers the algorithm is even open source and available to the public. The code is built on a custom quidditch-ranking library that can be used to create any custom ranking algorithm.

The idea behind the open system is two-fold:

  1. You should know how your team is placed the way it is. No smoke and mirrors, only numbers that your team creates through its playing ability. We want you to understand what factors move a team up and down in the standings
  2. We want you to help improve it! If you believe there is a better method, we want to hear it. Maybe you’ve thought of a crucial aspect of the game that isn’t accounted for in the present algorithm. Tell us! The reason why we’re doing rankings in this half-season is so that we can test out this algorithm. If (knock on broom) it totally falls apart and doesn’t work this season, there is plenty of time to revise and come up with better solutions. We want you to be a part of that improvement while we have a great opportunity in the next few months to tweak things without major consequence.

Dan Panzarella
CTO

 

 

Hello, IQA!

The Game Play department has spent months and months hashing out the final details of the rankings system for the upcoming mini-season, and we’ve finally completed it! With any luck, this will be the rankings system for the foreseeable future.

You can see this explanation here!

Thanks, and happy quidditching!

Will Hack
IQA Game Play Director

 

update 3/16: The technical explanation has been updated due to an inaccuracy in the described formula. The Gameplay department had decided earlier to remove the 1.1x boost for playing more than 5 games. The actual algorithm updated at that time to remove the boost (see the source code for yourself, iqa.js, line 384) but the explanation wasn’t updated to reflect that change, and still described a boost that wasn’t used. The explanation page has been updated to reflect the algorithm’s correct and present state.

 

22 Responses to “New Rankings Algorithm”

propic

Doug Whiston

Captain, University of Kansas; Kansas State Director

March 14, 2012

12:43 pm

World Cup games don’t count??? Okay…
I feel that all games within one year should count. It’s not like teams are different since then. That way, all teams get back on the board.

    propic

    Will Hack

    Captain, Jetpack Ninja Dinosaurs; Captain, Michigan State University

    March 14, 2012

    1:02 pm

    This is just a mini-season to test out the new rankings. For next year, and every full year, rankings will include all games from July to World Cup.

    March 14, 2012

    1:45 pm

    There are also *still* disagreements about the correct scores of some games. I’d rather not include games that are still unsure of the official score. The whole integrity goes down the toilet.

Alex Clark

Captain, Texas Tech University; Alabama State Director

March 14, 2012

3:49 pm

Haha ok to be honest that wasn’t reaalllyyy Utah at the Western Cup. I guess though that this is just a testbed that it doesn’t really matter.

    Jordon Parisher

    Captain, Texas State University

    March 14, 2012

    5:14 pm

    ^Agreed; this should be noted to make sure nobody does something like that in the future and is included in the rankings. That was more of an all-star team than Utah team.

Spencer

March 14, 2012

3:57 pm

I don’t understand why new rankings are being released now. They’re kind of meaningless at this point. Without a regular season it’s hard to say that running everything through the algorithm will show any meaningful data at this point. I mean, Tufts has 6 games listed and every other team has 3 or fewer.

    March 14, 2012

    4:04 pm

    It might not be entirely meaningful, especially to do it a few months into the “half season”, but it helps us to see how the algorithm runs with real-world data. That will hopefully expose any weaknesses or flaws in it that we mght not have anticipated. And it doesn’t really hurt to get something going. The other teams will catch-up and add their games in the next few days, I’m sure.

      Spencer

      March 14, 2012

      4:07 pm

      Yeah, that makes sense. Thanks for the explanation!

    Spencer

    March 14, 2012

    4:06 pm

    “testbed”
    My reading comprehension has failed me. Sounds like you guys realize the data doesn’t hold a lot of weight since there’s so little of it.

TX gal

March 14, 2012

4:26 pm

Do the Mardi Gras Cup games count in this?

Jordon Parisher

Captain, Texas State University

March 14, 2012

5:15 pm

I wish there was a way we could add games we played that didn’t count towards rankings. At Mardi Gras Cup we played 5 games but only two of the teams were “official”. The other teams we played were good legit teams, they just aren’t considered “official members”. Also if we could add that we played those teams, then so could other teams that played them there. Showing “oh yeah this game didn’t count towards the rankings b/c that team isn’t official, however we can compare how Texas State did against them to how UT did against them” etc etc. It’s not like we met up with a random team in the middle of nowhere to boost our rankings; these teams actually showed up to win the tournament at the Mardi Gras Cup

    Alex Clark

    Captain, Texas Tech University; Alabama State Director

    March 15, 2012

    1:41 am

    and If they became official mid-season I think it would be reasonable to count the games that happened earlier in that season.

Alex Clark

Captain, Texas Tech University; Alabama State Director

March 14, 2012

6:20 pm

Please upload! more data now=better algorithm later

    Alex Clark

    Captain, Texas Tech University; Alabama State Director

    March 14, 2012

    6:26 pm

    Dan can we also show the snitch% on the standings page? I would be interested to see that.

      March 16, 2012

      9:57 am

      No room at the moment, might have to use a wider table design. Or scrolling.

MsTake

March 16, 2012

1:11 am

So…let me get this straight. You defend, for over a year, a supposedly “state of the art” ranking system, despite the protestations that it makes absolutely no sense and isn’t in any way transparent; and then when people start leaving the IQA because no one is listening to them, you suddenly decide to do what they were asking for? And practically stealing their arguments word-for-word, no less (especially with that “smoke and mirrors” part).

Honestly it seems insulting to me that this organization’s leadership only decides to listen to its members when doing so serves their own interests. The same thing happened when some Western teams threatened to leave over the gender rule. It makes you seem manipulative and self-serving.

March 18, 2012

12:37 am

Utah is a community team now pretty much sooo yeah it was Utah. Haha

April 18, 2012

8:31 pm

Who would I talk to if some of the games from a tournament were added to our scores but not others?

Leave a Reply