Vote chess improvements

Daws74

Unfortunately, you can only see the last 10 games.  When you click "show all" it is missing the opponents group name and the result.

ponz111

Youi do realize that there can be objections to each of your posts?  How about giving pros and cons for each of your ideas.  I am not saying you are wrong but there are pros and cons...

1. how can you create a vote chess rating--give an example?

2. this will not work as there are players on some teams who do not want to play rated games for various reasons.

3.There are problems with this one also. One problem is why?

4. There are some teams who only play a certain opening and do not want to play any other opening

5. this is probably a good idea!

6. no comment

7. This would ruin teams where all members participate.  There is nothing wrong with having a low rated player on your team as there should be a discussion of each candidate move with the pros and cons.

8. dont understand this--why should someone be selected out of vote chess to play a challenge game?

9.   good idea!

May comment on the other numbers/ideas later.

Daws74

1. I'd love to see chess.com implement Glicko ratings for vote chess teams just like they do for individual accounts. In the meantime, you can check out the vote chess team Elo ratings published in the Vote Chess Elo Rankings group.

rigamagician
Martin0 wrote:

I believe the group names and the results can be seen at the list of completed games (example).

As Daws74 says, when you click on "Show all" though, there are no group names nor results.  This is also true of the list found under Play - Finished Games - Completed games. 

rigamagician
ponz111 wrote:

1. how can you create a vote chess rating--give an example?

I think the hope is that chess.com will just use the Glicko system that they have in place for one-on-one Online Games, and apply it to group vote games.

Daws74 uses Elo, and recalculates the ratings for all groups for each new list, but if they just used the existing Glicko system, each group would get a new rating on the completion of a game, just as people do for their Online games.

Daws74

Yes, that would be great if chess.com could just use their existing Glicko ratings for vote chess teams. 

ponz111

I do not know about Glicko ratings but you would be rating a team by its results with other Glicko ratings?

There are a lot of players who have no rating, including myself, who participate in vote chess--how will you rate those players? Or do you have to rate those players?

ponz111

11. being able to change your vote--sound ok with limitations which would have to be spelled out?

12. In case of a tie--last move voted wins--the problem is this could slow down voting as a few players would like to make the last vote.

I don't like first vote breaks tie either for obvious reasons--suggest a coin flip by adminstrator

15. not allowing a vote for first half can be a problem as quite often the move is obvious and then why wait to vote--many will want to vote right away so it is not hanging over their head?

20 and 22 seem like good ideas [to me]

rigamagician
ponz111 wrote:

I do not know about Glicko ratings but you would be rating a team by its results with other Glicko ratings?

There are a lot of players who have no rating, including myself, who participate in vote chess--how will you rate those players? Or do you have to rate those players?

Right now if you play a one-on-one "Online" game or two against another player on chess.com, you will receive a Glicko rating based on your performance.  What we are suggesting is that groups be rated in the same way, based on their wins and losses in vote chess.  You only need to know which groups the group played against, and whether they won or lost those games to calculate the rating.  You can read more about how Daws74 does his Elo ratings in his Vote Chess Elo Ranking group.

Daws74
ponz111 wrote:

I do not know about Glicko ratings but you would be rating a team by its results with other Glicko ratings?

There are a lot of players who have no rating, including myself, who participate in vote chess--how will you rate those players? Or do you have to rate those players?

We are rating the teams by their vote chess performance, not the players on the team.

Here's an overview of the Glicko rating system:

http://en.wikipedia.org/wiki/Glicko_rating_system

rigamagician
ponz111 wrote:

12. In case of a tie--last move voted wins--the problem is this could slow down voting as a few players would like to make the last vote.

I don't like first vote breaks tie either for obvious reasons--suggest a coin flip by adminstrator

People actually taking time to read the whole discussion, and then duelling against each other to be the last vote?  That sounds a bit like heaven compared with the current situation.

wormrose
ponz111 wrote:
make the last vote.

I don't like first vote breaks tie either for obvious reasons--suggest a coin flip by adminstrator

Why not use dice? Or a roulette wheel? How about rock, paper, scissors. Laughing

Daws74

I'll take the pass, black and good old dependable rock.

ponz111

Thanks for information on Glicko Rating, Yes, it would probably be a good idea to rate each team by that rating system. 

There are some large groups which have many different teams and probably teams change who is in each team at times--how would you handle that situation?

By the way, I hear a lot of talk about how teams having problems as there is little discussion before voting. The team I am on has much discussion before each vote and the discussion includes all members who wish to discuss and that includes lower rated players. We examine all suggestions and give the good and bad points for each and quite often we come to a consensus.  About the only time we might not come to a consensus is when we are easily winning and any of several moves will maintain the win.

Generally, we are asked to wait until all discussion is over and then the captain calls for a vote--often only one move is suggested and sometimes there are two candidate moves.

The reason that often there is often only one move suggested is that a consensus is reached via the discussions.

In the discussions there is a learning process as subjects such as good and bad bishops come up. On one endgame which was a little unclear one of our members came up with a very nice essay on good and bad bishops and good and bad knights and this helped us to find the right move and futher sequences.

So, at least with our team, vote chess is a great learning process...

DoctorWho

This sounds like a lot of work. But I'm sure it will have some great benefits.

~DW

wormrose

All of the groups I manage operate as ponz111 has described. That's the way to do it. Cool

rigamagician
ponz111 wrote:

There are some large groups which have many different teams and probably teams change who is in each team at times--how would you handle that situation?

Arpad Elo said rating chess players is like trying to measure the height of a cork bouncing up and down on the surface of a stormy ocean.  Everyone's performance fluctuates from day-to-day, hour-to-hour.  That is the nature of the beast.  A rating is just a summary of an individual or a team's past performance, and doesn't always correlate with their actual skill or potential.  If a team has trouble getting their best players to join the games, that would have an impact on their performance and thus their rating.

wormrose

From one perspective it can seem ridiculous to rate vote chess teams when individual members will change from time to time. On the other hand, a VC team is not so different than in amateur and professional sports. If you consider the Miami Dolphins - how many of them are really from Miami? Is that really a reflection of Miami's ability to play football or the quality of it's citizens? Players get drafted and traded and that changes individual members and thus the performance of the team.

Ratings are an approximation. That's all they can or ever should be and yet everyone is obsessed about 'more acurate' rating systems, bringing in things like the RD factor which rewards activity.

The c.c method rewards VC teams for activity as much or more than for performance. In sports - they rank teams according to performance, regardless of which team members may have missed the most recent game.

In sports all teams play the same number of games during the season. In VC that's not the case. So I think a rating system like Elo is a better indicator than the c.c method and over time will display the team's overall performance level.

Daws74
ponz111 wrote:

There are some large groups which have many different teams and probably teams change who is in each team at times--how would you handle that situation?

 

It is a rating for the team, so it doesn't matter who plays for them in any particular game.  The results of the team are what matters for the rating. 

ponz111

Makes sense...