Clan Rankings Explained

[chemefreak]

Thanks to ViperOverLord for asking the question. Thanks to Leehar for doing the legwork and for a great article. Leehar's original CC Dispatch Article can be found in the dropdown below. The various ranking systems mentioned in Leehar's article and those that were not are listed below with links to their threads. I will update as new ranking systems become available. [Note: At this time there is no "official" ranking system for clans. This is simply informational.]

My Clan's Better Than Yours - Rankings Explained by Leehar (Nemesis)

Spoiler

There's been a lot of debate lately with all these different rankings systems coming out, so after Qwert released his provisionally named ELO ranking, I decided to go out and find out what each of the creators had to say about their own and some of the other ranking systems.

We'll start by seeing what they had to say about the latest ranking system:

• Active Clan ELO ranking 2011:
  - The Active Clan Elo (or ACE) is presumably loosely based on the ELO Ranking system from Chess fame. From my understanding, it's applying the CC method of scoring whereby clans are given average scores from the useful Clan Rank add-on, and thereafter scores are gained by winning and losing. Since it's a new ranking there are obviously quite a few niggles with it, so it may just be best to adopt a 'wait and see' approach with regards to it's suitability.

  I think that i find very accurate clan ranking ,based on score and calculate in results- so far,they show quite acurate,because in 95% of all finished clan chalenges, highest ranked clan defeate lowest ranked clan when you look on average clan point score. Ofcourse for full effect need to get more and more results,because these is only a begining of 2011 year.

  qwert's is the biggest joke by whoever has the best winning %, for example we barely beat your clan 21-20, and lets say the clan ranked 21st beats the clan ranked 40th like 50 to 10, does that mean the 21st clan is better thna nemesis and thota? NO, but, qwert would have you beleive yes, seriously, it is a disgrace and joke to anyone who knows anything.

  ELO - The first ladder I used was founded in ELO and the disadvantage is lack of data. This works well for very active clans and with gaming privs, the active clans essentially have their own ladder. This ELO fits better with the times than my original ELO.

  I think Qwert's ELO rankings are interesting. I see some value in having them just to look at but I feel they are impossible to actually look at and give any credit to them as an actual ranking of the clans. If credit was given to them and that is what clans used to prove themselves then there would be too many top clans just trying to take on smaller clans in order to run their win% up and move themselves up the rankings. Obviously this would not be a showing of how good a clan really is but instead just who can manipulate the points the best.

• Perception of Top 10 Clans:
  - There has been some vigorous debate in the Perceptions thread lately about the suitability and application of the ranking, perhaps even some of the mudslinging between this and the Elo was spurred by my search for comments on the topic. It certainly has been informative with regards to historical data, but going forward we will need to see how it can adapt and ensure it has continuing purpose not supplemented by the Power Rankings

  chuuuucks and mine are the most accurate as it involves the clans and clan leaders to vote and poll and opinion's based on clans experience with war agains teach other and overall competition and records

  I think Blitz's Perception of Top 10 clans is really a great tool for past years and showing the history of who has been good. I think today it needs to evolve a little. First, I think their is a misunderstanding through the community and the people who vote of what exactly the votes are supposed to represent. Are they a snapshot at the end of each year of how the true rankings of the clans is perceived or do they represent a list of how actual clans performed in that given year only not taking into account past years? If they are just a snapshot of the rankings then in that case I feel they are very similar to the Power Rankings and there may no longer be use to run both side by side. I feel Blitz did a great job of going back and putting together a list for past years because it is impossible to go back and ask the people of those times to vote, but I feel moving forward with new rankings coming out, there needs to be a more defined process of exactly what his ranking represents and a fair system of how it is calculated.

  Well name say all-first page dont have any explanation on what parameters blitz decide who its second,third,fourt....(we know that no1 place are reserved for THOTA).These is totaly subjective formula for clan ranking.

  This has been a limited pool of contributors, but at least has some of the history of clans.

• Power Rankings:
  -An important compilation that seems to accurately gather clans based on Perceptions, perhaps it still has a ways to go before clans have accurate enough data to make informed decisions on clan strengths, but it's gone a long way in determining positioning in the clan structure.

  As far as Power Rankings go, I feel they do a lot to satisfy the biggest question on everyone's mind. "Who does everyone think is the best?" By no means do I think they are the most accurate way to rank the clans, but I think they take the important step of at least letting you know how all of the clans are viewed by their peers. And I do think they do get fairly accurate of rating most clans in the area they probably are really in. I think the clans that are ranked the most inaccurately are the medium range clans who are pretty good but don't have a lot of presence in the forums and don't have any particularly popular players. I think this is largely just due to most people don't know who they are or take the time to look.

  chuuuucks and mine are the most accurate as it involves the clans and clan leaders to vote and poll and opinion's based on clans experience with war agains teach other and overall competition and records

  Very good explane definicion of hes ranking,and its based on how other clan think who its in what place. Its not related on score and results,and its can be subjective ranking.

  Also based on perception but is more inclusive and has a better structure. I would prefer data based ladders, however, this one is very close to the leap ladder I'm currently working on. I would love to automate this and I have very distant plans to do so.

• Clan Wars Rankings Ladder - 2 Year Weighted RPI:
  -A mathematical ranking that seemed to get a lot of heat when it was established, tho largely sidelined lately and the data is now obviously pretty stagnant. A lot of effort went into the formulae and I'm sure it still holds appeal for those looking for a numerical benchmark.

  Factually based on results and is the best ladder out there but is not being updated as I'm currently trying to get a leap ladder in place.

  Well i look on table and formula explanation,and i must say that i find myself very confusing to understand what its what. To long formula,and to much comlicated for me-one thing what comlicated even more its A and B clans, Also its based on 2 years, and i think that maybe its best to try to create board only from 2011 games,and will be interesting to see,if mine ranking are similar to these ranking.

  Jp's is flawed by idiotic math and he weighs 8 and 16 game challenges in small legues for clan seasons too heavy, it is too small a sample of games, the challenges should not count at all unless they are 40 games or more

  I think the RPI is good work and I give a lot of credit to jpcloet for all of the work he puts in. I think it is largely criticized because I system like that is just tough to run with such little data among most of the clans and such varying results that are interlocking between the clans. You can have a clan that has beat two great clans near the top of the rankings while another who has beat 2 mediocre clans in the middle but when running an RPI it can be very difficult to distinguish between these two clans when they don't have any common opponents. It is a tough problem to find a solution for and I am not sure I could do any better than jpcloet is already doing which is probably why I have never tried.

• There are also rumors flying about the new leap ladder currently being discussed in the CLA, so some of our respondents had a bit to say on that as well:

  The model is built, however, it is currently in the CLA for discussion on how to make it work as it does require some foundations as to who can challenge and when and what format etc. I can see it being in place for the fall of 2011.

  I am more excited to see his next leap ladder he is going to come out with. Although I think there will be times of certain leaps that it will appear to be inaccurate, I think it will give each clan the most fair chance at moving themselves and will make it more exciting in the clan world to constantly try to improve your standing.

So there you have it, the word from the horses mouths themselves. Whats best is obviously largely up to debate, but hopefully we all have a deeper insight on the ranking systems available for our perusal.
I'd also like everyone to give a round of applause to these fine, upstanding citizens who took time away from their busy schedules to give us their views. We all know Qwert as an established map maker, and perhaps play often on his numerous maps, Blitzaholic is obviously always involved in his various Conquering endeavors and the many stat threads in General Discussions, Chuuuuck while working on releasing the eagerly awaited Power Rankings also had to organise and ensure both editions of the Conqueror's Cup are progressing successfully, and then of course there's jpcloet our Head Clan Director who tirelessly works behind the scenes on all clan issues.

Various Ranking Systems:

jpc's Clan Wars Rankings Ladder - 2 Year Weighted RPI

Chuuuuck's Power Rankings

Blitz's Perception of Top 10 Clans

Qwert's Active Clan ELO

FD's Algorithm of 400

The CC Dispatch contains a really cool clan section that surely takes a lot of time and effort on behalf of its contributors to create. Here is the link for the Dispatch:

CC Dispatch

Instructions on how to subscribe

[danryan]

I think of all the possibilities shown so far, my 2 preferred methods are FarangDemon's algorithm and Chuck's power rankings. The first is a solid number crunching ranking method, the second a pure opinion poll. If you combine the two, you get something like a BCS ranking for college football - not perfect, biased toward the bigger names, but certainly better that either one on its own.

[Leehar]

It's a pity FD's came out so soon after my article (maybe as a result of it?;)), I wouldn't have minded trying to get feedback on that as well.

[laughingcavalier]

Our client the Conquer Club Dispatch has issued instructions with respect to the gross infringement of copyright you have perpetrated in this thread. You are hereby required to send a cheque for £1,000,000 to the Dispatch Writers Retirement Charity (Ripoff Villas, Cayman Islands). This corresponds to Leehar's weekly Clan Reporter fee (£500K), loss of reputation (£100K), punitive damages (£1) & our own consideration (£399,999). Alternatively you can provide readers with a link to the Dispatch where they can regularly read articles as thought provoking and informative as this one - http://www.conquerclub.com/forum/viewforum.php?f=430 - and instructions on how to subscribe.
PS Leehar, any more nicey-nicey from you & we'll cut your fee again & give Marf another pay rise.

[chemefreak]

Our client the Conquer Club Dispatch has issued instructions with respect to the gross infringement of copyright you have perpetrated in this thread. You are hereby required to send a cheque for £1,000,000 to the Dispatch Writers Retirement Charity (Ripoff Villas, Cayman Islands). This corresponds to Leehar's weekly Clan Reporter fee (£500K), loss of reputation (£100K), punitive damages (£1) & our own consideration (£399,999). Alternatively you can provide readers with a link to the Dispatch where they can regularly read articles as thought provoking and informative as this one - http://www.conquerclub.com/forum/viewforum.php?f=430 - and instructions on how to subscribe.
PS Leehar, any more nicey-nicey from you & we'll cut your fee again & give Marf another pay rise.

Dear CC Dispatch:

Please be advised that the undersigned represents the CC Clan Section. Due to the fact that the above quoted material was compiled from multiple sources by the underlying creator it qualifies for the "fair use" exception to Copyright Law. We see nothing in the formatting that would qualify the writing as a "unique" piece of work under Copyright Law and therefore the work itself is not entitled to protection under the Law. Additionally, if any of you have ever seen those stupid Coke vs. Coke Zero commercials, this is a situation wherein CC would be suing itself.

However, the CC Dispatch is an integral part of the CC Community. As such, we have added a link to your periodical in the OP. We hope that this action will deter any further need for hostilities between our clients.

Very truly yours,

Dewey, Cheatum, & Howe

[laughingcavalier]

LOL - thanks Cheme
Glad you are offering these explanations of the rankings here.

[Qwert]

""Chuuuuck wrote:
I think Qwert's ELO rankings are interesting. I see some value in having them just to look at but I feel they are impossible to actually look at and give any credit to them as an actual ranking of the clans. If credit was given to them and that is what clans used to prove themselves then there would be too many top clans just trying to take on smaller clans in order to run their win% up and move themselves up the rankings. Obviously this would not be a showing of how good a clan really is but instead just who can manipulate the points the best. """
Hmm,i dont understand why people constantly repeating win%, in mine ranking i dont use any kind of win%.

[FarangDemon]

Hi Chemefreak/Leehar,

Please change the link in your OP for the Algorithm of 400 method to the below post, as it contains the clan rankings based on the complete clan war and clan league data set (Mar 29) on page 2.

http://www.conquerclub.com/forum/viewto ... 5#p3088157

If you want, you can include the following description (or a summary of it) and the feedback I have gathered in your OP:

The reason I wrote the script that calculates rankings according to the Algorithm of 400 is because I read about what qwert was doing, did some research on how ELO is implemented, learned about the Algorithm of 400 / remembered about how GameKnot (online chess site) implements it, wrote several comments in qwert's thread and then just decided it might be interesting to implement it myself. I love mathematics, I'm a systems and information engineer, so I have a lot of experience developing and analyzing measurement systems. Our is a branch of engineering that does not deal with any particular physical science - only how to interpret numbers that might represent anything or nothing in the real world.

I realize the method is very similar to what jpcloet has done before and you can view my work as a continuation of that effort. My script takes a clan war results file as input and automatically:

1) Generates a ranking based on clan performance
2) Graphs performance over time
3) Provides information on how accurate the model is (how often a higher ranked clan beat a lower ranked)
4) Provides information on how the result of each clan war influences the overall rating of a clan

The best part is that it's all automatic, so you can change around the parameters (there are many) and instantly have it all recalculated.

Some feedback:

I have to say, this is a very credible model from what I'm seeing so far. For one, it clearly shows what effect your input data has. Secondly, it removes all incentive for "farming" a clan. Lastly, it penalizes a clan (but not too severely) for inactivity, due to the decay factor. I'd like to hear jpcloet's take on it since he has put a lot of work into a lot of different models over the years to try and come up with a mathematical model.

Agreed, someone finally put some colorful pictures up so us peasants can understand... It does seem like a pretty close representation of what the clan landscape is IMO. Where the hell have you been in the CLA the past year while we've been bickering over 40 different ranking systems?

I very like this system of calculation, results are very reliable I think. decay idea is great.