VOTE CHESS RATING INSTEAD OF FIXED SCORING SYSTEM

I wonder if it wouldn't be possible to give a rating system to the vote chess teams, instead of the current fixed score system.

At the moment, the vote leaderboard doesn't give you a clear idea of the vote chess groups potential.

Team	# Matches	Win %	Points
Fast Players Group	317	75%	994
THE POWER OF CHESS	173	91%	756
CHRISTIAN CHESS WORLD	69	87%	267
Chess Unlimited	140	62%	215
Funtastic Chessmates!	53	89%	215
THE CHESS PLAYERS TEAM	43	88%	190
☸ People of Islam ☸	54	59%	140
Black Shield Assassins	31	100%	140
Rapid Chess	56	59%	125
Team Romania	25	100%	120
♕ K⋅TEAM ♕	26	73%	90
The Dark Knights	21	86%	89
The Chess Champions!	31	55%	89
WE ARE ADMIN	35	57%	87
Team Italia	18	94%	82
* The Dream Team *	16	100%	70
€uropean Union Team	16	100%	70
Chess Warriors	19	74%	65
The Cool Boys & Girls Chess Group	49	55%	60
We Chat Global	13	100%	60
Indian Marauders	12	100%	60
☺☺BLACK оя WHITE☻☻	12	100%	55
Vote chess and fun group	39	28%	55
WORLD CHESS GROUP	55	22%	52
SnowWhite	30	37%	50
Narz's People	9	100%	45
Team Canada	9	100%	45
The Best	14	57%	44
Power of the Attack	27	37%	42
Vietnam Group	28	29%	40
Light Force	18	44%	40
The Winning Team	52	25%	40
Team Bangladesh	8	100%	40
Team England	13	62%	40
Quick Players	9	89%	37
Arab chess players	10	70%	35
O'zbekiston Jamosi	14	64%	35
Cheers! Vote Chess Café	7	100%	35
Learning with Fun	8	88%	35
French Defense Fanatics	7	100%	35
The 2K Club!!	8	88%	35
Chess Friends for Friends	10	60%	32
Chess Masterz	18	44%	30
Team Bosnia and Herzegovina	23	35%	30
Team Britain	14	50%	30
Fanboys & Bookworms & Readers	9	67%	30
Chess Association	9	100%	30
Chess Club Finland	9	67%	30
The Twin Ports Fianchetto --- Team Cold Climate	7	86%	30
The Learning Group	12	50%	30

I am not saying that the groups on top are not really good, (or maybe one of the best)... but they are there mostly because they play much more often than others.

I wouldn't like to remove completely this leaderboard, but maybe it would be really interesting and useful to add a line with the Elo rating for vote chess groups (or according to glicko system as used in chess.com) with the same features as for individual players. The vote chess groups starts with 1200 and then gets more or less rating depending the wins and against which teams they play...

I even think it would possible to create it already with the current information of games played in the past...it would give a clear picture on what the vote chess groups potential is.... not just because they play more or less oftenly and or because they play always against weaker groups or recently created.

Please let me know your thoughts! :)

And chess.com staff please let me know whether there would be any constrains creating this. Thanks in advance!

Agreed  & supported 100%.

My team has beaten Fast Players four times (no losses) but they are high ranked because they simply play so many games.  May as put Aniko as one of chess.com's highest rated players because she has so many wins (no offense Aniko, props for your dedication!).

i'm not saying no to this, but honestly, most teams will never finish enough games to even get an accurate rating, which is why we didn't do this at the start.

 Yes, this is happening because new groups are accepting challenges with less than 1 day...(clearly they are not prepared)

Anyway, I think that this happens in all the chess games modes...what about adding also the time out ratio too? (yes, I know that I am asking too much..:) but I spend most of my time playing vote chess in all of my groups... and I am tired of 2 things.... winning on time, and playing against really unprepared groups or really strong groups... with the leaderboard points... you don't know what to expect so you isssue the challenge to the wrong team always..

It would be really nice to have something like the ELO rating available.

Thanks in advance!

BTW, congratulations Erik for this fantastic site... I love it! :)

Apart from that, as a win is worth the same if it is against a strong team or a newly created group... all the groups are "tempted" to challenge those "weaker" groups... and that's the main reason why there are too many loses on time...

If they were afraid of losing a big amount of points on their ratings... they would focus onto the strongest groups instead...

I agree, a rating system would be great.

I'm not gonna mention names, because technically our move only happens in 2 days,  but in one vote chess game we are voting for a mate in one after 8 moves against one of the top 5 teams on this list. Clearly they don't belong in this position.

I clearly think this is a fantastic and correct approah. lots of groups are at the top simply because they play far more games than smaller but much better votechess groups. and there is no reason why this form of rating cant reside next to the ususal one.

A great idea! It would be much better than the current one! :)

Well, if there are not enough games for accurate ratings (good point, I didn't think of that), what about a ladder of sorts,

 Yes erik you are absolutely right that it was the right way to start with. As with the passage of time new features are being encorporated into function, I think there might be some new VOTE CHESS teams could perform very well but it would take them ages to reach the top since they will have to match the number of games played by very seasoned and old group teams.

There might be different analyses to scrutinize which team is playing better and IS actually better. Afterall it is CHESS WIN that is important for a team's rating as well as the TEAM WORK. I reckon the no of games played or under process should be secondary.

Therefore, I agree with kokino on this one that NOW might be the right time to give new teams fair chances to stand higher (or lower) mostly because HOW WELL they are performing and not because HOW MANY they are playing.

Thanx for listening

Regards

Nabeel.......DaSorcerer

I understand that many groups (including mine) are impatient to see themselves at the top, and existing point-system will always keep groups playing the maximum number of games (not maximum % wins) at the top...

But we must realise that it is very difficult to implement conditional moves in vote chess... that means the games will take very long time to end anyway.

Now, for the sake of argument, let us assume that Glicko system is implemented; will it be possible to assign points based on all past games played, or will every group has to start with 1200? This is important, as without the stats of previous games played, it will be like starting vote chess in chess.com afresh, and it will take years to get some decent stats for each group. We will find more groups objecting than the number of groups agreeing... otherwise it is definitely a very sound logic...

Instead of the Glicko, is it possible to do simpler algorithms based on existing stats, like % of Wins and the number of points/matches? This will not have all the advantages of Glicko, but something that can at least be implemented right away, and we dont have to start afresh... Some of us might be able to come up with some brilliant algorithms...

We must realise that a few voiced opinions will not help Erik to reach a decision - any implementation will have to get the vote of the majority...

I know it is a hard job, more over when Erik's staff is working hard in other priorities, but the info is there so it should be possible to implement the glicko system using all the games archive, so rating would be the real ones... anyway, this is an idea and my personal opinion, of course is Erik's final decision if they want / can implement it... or not.

 You are right, and as I mentioned before, one of the problems because of some many games lost on time, is because strong teams keep challenging newly created or weaker groups (with not enough members)... at the end, a win is a win with the current system, so if you want to jump on the leaderboard, you don't risk playing the best... is easy.

As kokino has said, if Glicko is possible base on all games previously played, I am for it. If everyone has to start from 1200 from scratch, then I am against it.

"strong teams keep challenging newly created or weaker groups (with not enough members" - I know exactly what you are talking about - that is one of the abuse potential of the current system... 

Some teams have completed a lot of games indeed, so a Glicko rating for them based on all the games they've played so far would probably give a pretty good idea of their strength.  The current points system on the leaderboard doesn't really tell you too much about a group's strength.  Excellent suggestion anyway.

I'm in favour of some kind of rating system for Vote Chess, as the current Vote Chess leaderboard doesn't make much sense, indeed.

If Glicko is not appropriate (I don't know if it is or not, just a thought), then maybe simply use a points system taking into account the % of wins of opposing teams : if I win against a team who has 100%; it's 10 pts, if I win against a team which has 50%, it's 5 pts, and so on. This way, winning plenty of games against weak teams doesn't give you many points.

I am not sure, simply guessing... If Glicko or similar system is implemented, all groups will have to start with a clean slate, ie. 1200 or whatever[It is extremely difficult to calculate based on past games - e.g. we dont know the rating of group B at a particular moment of time when group A played with it]. As conditional moves cannot be implemented (I sure wish there was some way a group admin could use up the remaining time if he/she was satisfied that sufficient votes have been cast), games will take a very long time to finish (each 3 days/move game will take 6 days/move or 4 months for 20 moves - if games take 20 moves on an avg to end, we will complete 3 games / year.) That means it will be nearly another 5-10 years before all the groups will have any meaningful rating.

Are all of you guys and gals saying "yay" OK with that?

I think I've said this before, somewhere...

What is needed is a RANKING system instead of a RATING system. Winning moves your position up, losing moves it down. The size of the movement depends on the start rankings of the two teams.

 rooperi, excellent idea!

But how do you implement it?

(1) How do you rank them initially?

(2) What happens if you play someone below you, and win?

(3) What happens if you play someone above you, and lose?
 

I believe what we are ALL looking for is a brilliant algorithm based on current stats of points, %wins, games played, to give a more meaningful "ranking" based on actual playing strength rather than a slewed ranking based on total number of games played.

I think we have so many members, some brilliant statisticians, that it should not be impossible for someone to come up with an algorithm to satisfy everybody!

------