29 ( +1 | -1 ) Ratings CapI must admit I still find it odd to play people with a rating of 1800 who have never beaten anyone with a rating over 1600. Why not cap people's ratings at say 100 points more than the best opponent they have beaten in the last 50 games?
82 ( +1 | -1 ) if i understand you right.....because if a higher rated timed out, or made a blunder, or his/her pieces didnt fully load, and made a bad move from that, then an 1800 might beat a 2100 on an accident, when the 1800 might only have a strength of 1500/1600..... plus, for all the people that lucked out against brunetti , they would have an extremely high rating, and brunetti would not be able to have a higher rating than already, because he makes very few points on wins, loses many on losses, and no one would be higher than him to beat, or something like that. if rating system would be totally out of order....
55 ( +1 | -1 ) Ratings Cap ClarificationBy cap, I meant the absolute maximum, not what the rating should be. Just trying to introduce a little rationality into a system that inflates ratings dramatically. I think that it is fine for people to consistently play (and beat) players 200 to 300 points beneath them. I am just not convinced that these victories should raise their rating beyond a certain point. I assume that if we wanted an accurate rating system, we could adopt whatever it is that the Internet Chess club runs (www.chessclub.com).
95 ( +1 | -1 ) I think its fine.The rating system here works fine. I think your looking to much at the points maybe. By looking at opponents time investments on his/her moves and past games u get a pretty accurate idea what kind of a player ur dealing with. Trying to get "accurate" ratings is (maybe)just vain and not interesting anyway.Some players r more serious about their chess(or should I say outcome of game), maybe using a real board cutting down their visual analysing time. A player maybe rated 1500 with 30 games going on using 30sec/move (and playing vs strong opposition).Hes probably a much stronger player OTB real life then another guy rated 2000 here sitting with books databases analysing..analysing..but why should it bother me, im here to play some nice games and hopefully some nice chating.
13 ( +1 | -1 ) A+Ithink glorioso is better then his rating.. can u pls give me a forum +? I just gave u one and my forum rating stinks..Im a nice guy. :)
13 ( +1 | -1 ) i agreeWith mr samuelsson. and since it matters, i'll even give ya a + :), even though he beat me in a humiliating way not long ago... That's how big a person I am :P
99 ( +1 | -1 ) gunnarsamuelson response.I think it is possible to get some idea about who are you playing against by looking at the factors you suggest. Although I have played both ways--with almost instantaneous response at some times and agonizing over individual moves at others, and the ratings difference is probably only 100 points.
Given this set of understandings, I do sort of assume a plus or minus 100 points on any number that I look at--and thus agree that any rating system that assumes that it can be much more accurate than that is going to be more misleading than helpful.
I do stick to my original point of view, that it would be nice if there were some way to devise a system that prevented people's ratings from being 200 to 300 points off line. The cap idea was just off the top of my head. I am sure that the more mathematically inclined could be more systematic.
I give you a ratings plus for your comments: civil, concise, and informed.
136 ( +1 | -1 ) cappingI thought about this when I recently started using the site (I was quite surprised that the ICC rating algorithm wasn't implemented). However, I think this system gets you close enough anyway.
If you regular play and beat people 2/300 points below you you're going to be increasing your rating by 8 points quite regularly. After a time, this level of opposition is going to be 400+ points below you, so you either have to risk losing 16 points for a 1 point gain, or start playing better (higher rated) opponents. These people will then have a higher rate of success against you, so your rating will then start to diminish.
Admittedly you can create a rating for yourself a couple of hundred above your realistic ability, but you can't continue increasing like that, so there is a kind of virtual capping in place anyway. Also, checking opponents average rating clearly shows those that do this.
Ironically though, I would like to have an ICC rating, but play here as I think the site has been very well developed, and have been impressed with the other players here. Good games of chess are more important than an accurate rating in my opinion, although it would be nice to know.
58 ( +1 | -1 ) Re: FiddlerI agree that ratings do not make a chess site. In terms of the size, I am not sure the rating inflation is as limited as you say. I wonder if there is not a second order effect--i.e, if you inflate your ratings by 300 points by playing against people who basically do the same, what is a real rating.
Example: you are a player who has a rating of 1800 beating people who have an average of 1500. The players with 1500 ratings have made their ratings by beating people with an average of 1200. I am curious as to whether one could actually assess that mathematically.
22 ( +1 | -1 ) Kasparov......is rated the highest in the world (2800+), and cannot therefore play anyone higher than his own rating since he is #1. Not sure what I am getting at, but a rating simply means you have won a LOT of chess games, I guess...
34 ( +1 | -1 ) Thanks to all who contributed.Since this thread seems to be about to pass off the front page and into oblivion, I just wanted to thank all those who threw in their two cents worth. I am going to try playing on www.playchess.de, which seems to have a more accurate ratings system and see if it produces any better chess. Will let you know if it does.
27 ( +1 | -1 ) gameknot vs. otbcan you infer anything about your gameknot rating in terms of what your otb rating would be if you played otb tournaments???
just wondering about people who do play otb. is your gameknot rating close to (or above or below) your otb tournament rating?
78 ( +1 | -1 ) GAMEKNOT CHESS RATINGS?CONVERTED TO OVER THE BOARD CHESS RATINGS IS BASED ON YOUR SCORES WITH HIGHER RATED PLAYERS WHO ARE RATED OVER 200 POINTS AND LOWER RATED PLAYERS WHO ARE UNDER 200 POINTS.IN GENERAL IF YOU HAVE A PLUS SCORE AT LEAST(4)GAMES UP!WITH PLAYERS OVER 200 POINTS ABOVE YOU OR EQUAL OR WITHIN YOUR CLASS THEN YOU ADD BETWEEN(100-200)POINTS TO YOUR PRESENT GAMEKNOT CHESS RATING TO GET AAN IDEA WHAT YOUR OVER THE BOARD CHESS RATING MIGHT BE?AND IF YOUR LOOSING TO PLAYERS UNDER 200 POINTS OR ONE CLASS BELOW YOU THEN YOU SUBTRACT(100-200)POINTS FROM YOUR PRESENT GAMEKNOT CHESS RATING.BUT YOUR GAMEKNOT CHESS RATING MUST BE AN ESTABLISHED RATING AND YOU HAVE PLAYED OVER (24)GAMES!