|
|
|
Topic: rating formula needs adjustment
| |
|
|
Author
| Message |
|---|
|
|
Here is a link to a ratingsystem.
http://math.bu.edu/people/mg/glicko/glicko.doc/glicko.html
|
|
|
Ratings
The ratings list can be calculated on the first of each month and uploaded to the site on, or shortly after, that date (this is the only time you will see your rating change - it isn't recalculated after each game).
Ratings are calculated using the 'Glicko' system using the formulae described here. You will notice that after your rating there is a figure in brackets - this indicates what the Glicko system calls your 'rating deviation' and gives an indication of how reliable, or accurate, your rating is. You will notice that your RD gets smaller, the more games you play - this is because the system is more confident that it is getting close to your 'true' rating. If you don't play any games for a few months your rating won't change, but you'll notice your RD getting larger again.
There are a couple of things to stress about ratings...
Ratings are relative, not absolute - there is no reason to suppose that your rating should be the same as your rating anywhere else. An 1800 rating here does not mean the same thing as an 1800 FIDE, FICS or USCF rating.
Any rating system makes the assumption that there is a random element in player pairings - people like to play their friends, so some pairings will be far more frequent than others.
Above all, please don't take the ratings too seriously - it's really all just a bit of fun.
Please let me not explain the system. My intention is only to show a system to soneone how is clever than me. 
|
|
|
Beco, you have 59 games under your belt. Several of us have hundreds, and you are saying to negate the results of all of them. Also, this means that no ratings will be available for tournaments for a while. Finally, there is a problem in the way that the ratings grow: If you as a rated player play a provisional player and lose, it does not affect your ratings, but if you win, it does. This is one of the causes of rating inflation. Miguel has this as a matter of policy, so that rated players will not avoid playing provisional players.
|
|
|
the problem remains as it was.a novice may win a high rated player and earn a rating of 2000+ and the the problem of infalation remains as it is,may be after a month or two we have to repeat the process starting with 600 or 500 points.
This is one of the reason that why the chess is not much popular relative to other games, the novice comes, play and win some early games that boost his rating and moral but when he starts loosing he goes away saying graphes r sour.we have here eamples of that here as well.
we need to change the system not just the starting points.i think it as a insult to the members here that a newcommer, how much smart he is, play "just a game" and gets into top 10.
|
|
|
| think it as a insult to the members here that a newcommer, how much smart he is, play "just a game" and gets into top 10. |
I'm sorry, you got it wrong. Nobody gets in the charts after playing just one game, you have to play 25 games to appear in the rating list. Players with less than 25 games are provisional, their ratings mean nothing until they complete 25 games (did you see the 'p' right after the rating?).
Miguel
|
| Previous 1 2 3 Next |
|
|
|
