|
|
|
Topic: Opponent's rating used to calculate rating adjustment
| |
|
|
Author
| Message |
|---|
|
|
Blutigeroo wrote:
| The rating we use is quite volatile and contains very little history. At best one only knows that a player that is rated at 2000 was between 1680 and 2320 ten games back. |
In practice, it's not nearly that volatile. It's possible for a player to lose a lot of points in ten games from time forfeits but it's very unlikely that a player will win ten consecutive games against players rated much higher than themselves.
| Much better to have something like a 100 Max number that averages your 100 best ratings. |
Basing a rating or qualification system on players' peak performance doesn't work. A player's strength can fall, either because of age or because they don't have as much time to spend on chess any more. Is sandbagging really that much of a problem on this site that we need to take action against it?
Dave.
|
|
|
Sorry, I should have qualified that. It is volatile for those of us that are playing large numbers of games and perhaps at a faster pace than others. In such cases either of two players starting a game may finish 20 - 40 games or more before that game finishes. If that is not a potentially volatile rating change, I don't know what is! 
As for sandbagging, I'm not saying that it's always the case. Sometimes it really is someone that hasn't been able to play their games for one reason or another. I play exclusively in Advanced tournaments and I'm rated near the top limit. However, trying to win such tournaments is almost hopeless (zero for me) as invariably there is one or more advanced players in that grouping. Then you have the prolific numbers of "new" players moving up to their rating (I have no answer for that).
Yes, the 100 Max number implementation would be tricky for the reasons richerby has given. Perhaps having that number displayed on tournament sign up pages would be enough however that too may cause a problem with the tournament hanging with no sign-ups! 
|
|
|
I think any variation on a peak- or average-based system would have unintended consequences. One would be the fact that once someone falls from a high level they will have a very difficult time getting back up there. This is because just as the loser's rating would only drop a little, the winner's rating would only rise a little.
|
| Previous 1 2 3 |
|
|
|
