Rankings

[esquire]

To anyone out there. Is there a way so we can modify the current ranking system. We could maybe reward players that actually play. And reward players that play and win against higher ranked player (even after the player has reached a high games played mark). I know it is all fun, but having certain players ranked really high, and all they do is play low ranked players, and having low ranked players who beat high ranked players never become high ranked because they played too many games to get much higher with each win is kind of deflating.

Also allowing certain players to stay high-ranked, while never playing is very unfair to many.

Also, maybe you could put some modifications that allow you to get more than 1 or 2 point increases in your rank when you beat top ten (20) players--just a suggestion. Keep up the good work.

[Agresor]

hi,esq
i wrote many times about Kleier system,pls read my writing about in january.It is no doubt,system is totaly corrupt,no chance to any position changes after some 50 and more games.It is totaly unmotivated and with big logical and mathematical mistakes.
BUT the bosses hir WILL NOT CHANGE this rating system!
It was many replays on this metter ,with no respons from spion or silver,or with some funny answers on statisics from a guy named Herman Kleier who have no idea about how to do and calculate such rankings!!

i hope someday will be better,and somebody changes permanently this system and also also cancel WSC system,and start NEW age of claculation of ratings fro players!!

Many players have stoping serious play exactly from that bad Kleier ystem!

till this days
greetings from Agresor

[spion]

Hi!

i made an offer several times now:

If someone wants to have a new (better?) rating system, feel free to make suggestions.

If someone can offer me a new system, i will look through it.

Yet i added the WSC system on many players request.

The Kleier system is the official ISF system and so players from europe are able to compare their ratings on Gravon to those of real life tournaments.

But if someone can offer me a new revolutionary system, i wont trow it away

But i do not want to add ranking systems until every player is #1 somewhere.

spion

[Agresor]

Hi
Are you seriously Spion,i can suggest new ranking system!
Very efficient and very easy system,but i think your famus Herman Kleier will find something out with her bad theorieto persist on this corrupt system.
Also i am not sure whay shall i do this,after all you give me zero answer on some suggestion that i done till now(including agressor advantage rules,ecc..=).
But i am still cooperative ,so i wait on your answer!?

greetings
Agresor

[esquire]

One suggestion. The ranking could be different from the rating. You could have a rank of players--and you rank goes up with each win of somone higher than you. Kind of like the Yahoo Games ranking system. The top 20 poeple can be formally challenged, this works good if they do not play poeple that are ranked close to them, which will enable them to pass them. Also, if somone does not play often, their ranking will decrease.

This system works, and it is fair. For example. If I am ranked 20th and I beat the #1 ranked person, my new rank would be 10. And if I beat someone lower ranked, nothing happens. But if I lose to someone lower I iwll decrease, etc.

[googoomuck]

What about The Chess Express rating system?
_________________
Metaforge Message Board Moderator

[esquire]

I agree with GooGoo--that is a good system to use. I just got through reading, and it sounds very good.

[spion]

Hi

The WSC rating system is mostly the same as the Chess Express Rating. The only difference is the handling of players with less than 5 games. The rest is exactly the same.
So if you do like the chess express rating, look at your WSC rating

spion

[spion]

hi again.

I guess the major problem is the static type of the rating systems we are using. Once a player has finished a bunch of games (say 50), his rating will stop to vary much from game to game.

It may be frustrating to see that a win only gives you 1 or 2 points or even nothing.
But thats the nature of these rating systems: Show the players overall playing strength.

So what you likely wish to got is a dynamic rating system.

H.Kleier is currently working on one.
Perhaps i can find enough time to think about a system which would be quick to program.

spion

[jerry]

I ask the older generation of players here, was FairHousing Fair?, The WSC system for play was based were?

How serious do we have to be here?, I have found both systems equal here and Meta, as Spion has mentioned, be realistic with points, or just play the game.

Then again, I,m just Australian, so I don't understand the fine lines, but I do understand "Buttwheat" who just say's "Play the game and quit the BS"

If their is a better system,can it help for others to winge about what is right or wrong? Players strength has been shown in the rankings, if correct, I guess a good system would be only for the honest players?

Confused?

Ghochan, is the only person that may put this into perspective..

Cheers Jerry.

[Agresor]

Hi all
as i tell you I can show you IMPROVED CHESS(wsc) SYSTEM ,who is adapted also on some critical problems in praxis and is very symple to implement in programm.
This system us we in Slovenia(not much players..) and works excellent,while is mathematical and logical correct and very simple system!

greetings
Agresor

[aeolus]

Play the game, learn, get better and you will rise in the standings.
_________________
Aeolus

[tableplay]

Single Exponential Smoothing to give more weight to recent results is a reasonable solution in my opinion:

Step 1. New column called Smoothed Kleier Score is added
Step 2. Smoothed Score is calculated as shown below
Let
N=next smoothed Kleier Score
K=current regular Kleier Score calculated using Kleier system as if this was only your 20th game ever played (can be adjusted higher or lower than 20 but the idea is to estimate your rank as if you were a new player to Gravon and were receiving your ranking for the 1st time)
S=current smoothed Kleier Score (S=K the 1st time the equation is applied)
a=smoothing constant (the higher the value, the more weight for more recent results) where 0 < a < 1

so N=aK + (1-a)S

[ghochan]

Jerry said I could add some perspective to this. I'm not sure that I can since the only thing I can be sure of is that when I play a rated game, my rating will go down as a result Besides, I studied history and numbers give me a headache.

I'm not a chess player but I do remember a discussion about rating systems back at WSC (this is about the fourth time I've witnessed this same worn-out topic being discussed -- bottom line: no system is going to make everybody happy). In the WSC discussion, somebody (I believe it may have been llimoinc) pointed out that the chess rating system is not well-suited to Stratego because in Stratego luck plays a much greater role in the outcome of a game. In other words, there is a much greater chance in Stratego that a lower ranked player will defeat a stronger player than in chess, where the ranks of all pieces are visible and the initial set-up is always the same.

I thought that the WSC system tried to remedy that by changing the formulas a bit, but I don't pretend to understand how that works. I also have absolutely no clue how the kleier system works, although I have tried in the past to read his explanation (I would have more comprehension of the Iliad in the original ancient Greek), so I can't be of much help in determining which is more accurate or which is "better."

I do believe, however, that any system will over time distribute people on a ranking in a way that reflects their relative strengths and weaknesses. I also think there is some merit to having one system that can be used in different settings. The fact that the kleier system is used by ISF for board tournaments is for me a strong arguement for using it here. Maybe the discussion should be about whether the online environment of gravon is somehow different from the face-to-face tournament environment of ISF that makes the kleier system ineffective for us here at gravon. For example:

- How does the number of players on gravon differ from the number who play face-to-face? Does this matter?

- Does the fact that some players, notably some high-ranked players at gravon, rarely play any games have an effect on the validity of the kleier rankings?

- In a face-to-face tournament, players are assigned opponents and there is a true mix that ensures that most players will at some point have a chance to play most other players. At gravon this is not the case. There is no control over who plays whom and how many times they play. Players in different time zones may never meet other members in battle. Does this have an effect on the validity of the kleier system?

If you find that these circumstances affect the accuracy of the Kleier ratings, then you should begin talking about how these variables can be accounted for, and what system would do a better job of reflecting actual performance levels. I kind of doubt that there will ever be agreement on this. At WSC, we tried to pull together some of the best minds in online stratego to develop a rating system that would work for us, and I think it satisfied most people, but there were still some who didn't care for it.

Some people have argued for players losing points as a result of inactivity, and I have always felt that this would be a big mistake. The online community is much "looser" than a face-to-face club, and the incentives and penalties have to be different; otherwise people will just stop playing. Penalizing someone because they only play once a month or so could just push them to quit altogether. On the other hand, allowing people to sit at the top of the rankings without ever allowing themselves to be challenged isn't right either.

Maybe we need two tiers of competition. A "masters" ranking for those who play above a certain number of games per month and just a "rating" for everybody else. If a master's ranked person falls below the required number of games, they would go back down into the regular pool. The master's list could use a more aggressive point based system while the pool could, perhaps, be based on kleier (or opposite, if that is a better way to do it). Just a thought.

Well, for someone who had nothing to say on the subject, I just spewed a lot of words onto the screen. Sorry about that. I hope this helps in some small way.

ghochan

[esquire]

Spion's suggestion is the only thing I would like to change. I have played too many games now to really move up in the rankings. I usually get either 0 points or 1 point per a win. My ranking/rating is very low in the Kliers because of this. I have improved much since learning how to play this game--it took me about 150 games. If I were able to get more than 1 point per a win I would be happy.

I rank 7 on WSC, so that is pretty cool. But nobody on here even looks at that ranking system.