|
|
|
Topic: Should Miguel change the rating system?
| |
|
|
Author
| Message |
|---|
|
|
This question arises from some short remarks made by 'mateintwo' in another thread. Personally, I am completely open-minded about this and I would simply like to start a debate about the pros and cons of our system and the various alternatives.
All the best,
C
|
|
|
We can suggest many things but never changes because the site is poorly managed.
|
|
|
I don't think it is too bad given that M isn't a full-time webmaster; but if we could at least start discussing (in the hope of reaching a consensus) we might eventually have a concrete proposal to put to him.
Although I implied that I had no suggestions, my best idea (with regard to games between 2 players with established ratings) is that a player should be prohibited from gaining points by winning a game against an opponent more than 400 points below him (or her). The higher rated player could still lose points (if the game was drawn or lost) and the lower rated player could still gain points, but 'meaningless' accumulation would be precluded.
|
|
|
ClivetheBeard
I agree with you. But let's remember that the rating system in Queen Alice is based on United States Chess Federation system.
There are some special cases. For example, rating differences above 350 points are calculated as 350 points, fixed.
For players that have rating higher than 2100 points, the equation changes. We calculate 3 or 2% of the rating difference, instead of 4% in the general formula. The K factor usually is 16, but in special cases it is 12 or 8.
Well, you can see further information here:
http://main.uschess.org/content/view/7520/393/
I hope that I help you. Greetings! 
|
|
|
About the difference, let's suppose this:
John: 2500 points
Jack: 1600 points
John's equation: Rn = Ro + 0.02(Dif) +/- 8
Jack's equation: Rn = Ro + 0.04(Dif) +/- 16
Rating difference: exceed 350 points, so we take 350 as the difference. And, the difference is Opponent - Yours. So, to John the difference is negative and to Jack is positive.
1) John beats Jack:
John = 2500 + 0.02(-350) + 8 = 2501
Jack = 1600 + 0.04(350) - 16 = 1598
It makes sense. John is stronger than Jack, and their new ratings is similar to the old ones.
2) Jack beats John:
John = 2500 + 0.02(-350) - 8 = 2485
Jack = 1600 + 0.04(350) + 16 = 1630
Wow, Jack won 30 points and the "penalty" to John was 15 points.
3) John draw with Jack:
John = 2500 + 0.02(-350) = 2493
Jack = 1600 + 0.04(350) = 1614
Well, we can see when a player has higher rating, the formula helps him, protecting his rating (the factor is 2% and the K factor is 8). But, a weaker player (that has 4% and K = 16) may up very fast, if he can beat a good player.
|
|
|
Thanks rbivanov.
The problem is (of course) that the internet can turn a correspondence game into a blitz game and a good player can accumulate hundreds of points (one at a time) very easily. Not sure what the answer is, other than what I proposed.
I think that one USCF rule that we should definitely adopt is the one that limits ratings gained by defaults. I know that I've benefited several times from people not turning up in the past, but I'd be quite willing to play under this rule in the future.
|
| 1 2 3 4 Next |
|
|
|
