the real rating of the master!
Sort:
[COMMENT DELETED]
Dec 24, 2009
0
#23
costelus wrote:
In order to get above 2600, one must defeat along the way many programs. So the answer to your question is the following: if the top players here were not be cheaters, they could easily outplay even the strongest GM's in the world. Ussually they make zero mistakes in hundreds of games and compared to them even the correspondence champions from the past look like kinderdargen kids. In reality however the top players here did not even hear of "ELO" and their OTB playing strength is well below 2000.
However, you have to admit that the competition here is much lower than at the international level. It takes a lot of money to go around to all the tournaments around the world, whereas almost everyone has access to the internet. Therefore, those people who are not going to do ultra-well in those tournaments will probably not go. The super GMs that are left will have much harder competition, and Carlsen who might make 4000 on this site would only make it into the 2800's in the real world. So not necessarily are people who are not titled cheating on this site, its because they only have to play against patzers like me, rather than Carlsen, Topalov, and the other 2800 giants.
In response to the general question, I would say that there is no constant proportion between chess.com and FIDE. 2800-2900 players here are probably 1800-2400 FIDE, leaving a whopping gap of up to 1100 rating points. I am rated 1600 here, but only 1150 USCF, probably corresponding to 1000 FIDE, with a gap of up to 600 rating points. In the extremely low rated sector, I would say that chess.com rating here would be lower than FIDE, (seriously, who could be rated 0), so I will draw the conclusion that the rate of increase on chess.com is steeper than in FIDE.
ori0 wrote:
philidor_position wrote:
ori0 wrote:
philidor_position wrote:
ori0 wrote:
a nother factor is the time out. a player winning as a resolt of a time out (or loosing) has a rating that is not only based on how well they play and as above the snow ball starts roling affecting every one playing agenst him and any one who played agenst those players who where affected by the time out.
This may not really be the case. I don't see any difference between a Time out and blundering into a mate in 1. If a player loses frequently because of time-outs, his/her rating will reflect it, exactly the same way the rating of a player who frequently makes game losing blunders reflects their blunder rate.
your right if were talking about a profesional game but most of us play here chess for fun fiting it in ower schedual so time outs here arent a resolt of thinking slowly, it is just a resolt of peole having more to life then just making the game time at chess.com . therfor the person loosing isnt thinjking slower and isnt loosing because of being less good he just has other things to do.
that affects the rating sistem badly
that does NOT affect the rating system badly, because it's not the players that are rated, it's the players' performance that are rated. it makes all the difference.
true! exactly! it shows how well the player fits the time of the chess games in tohis life instead of just how well he thinks. if the loose is because of thinking to slowly i whould agree but here the rating of the players a lot of times is affected by people just not coming back on time loosing because they didnt fit chess in to there schedual nothing to do with thinking slowly fast or fiting the moves in to the time just plane not coming back!! to the game
Well, you have to consider two sides: true, your rating will not look right if you lose a game on time, however, you have to assume that approximately as any people people you are going to time out. It is not very accurate, but it does show that your rating might be closer than you might think.
You can't really make any blanket statements about someone based on their rating. All you can say is that in the same environment, the person who is rated higher should win more often than someone who is rated lower. Also, as the ratings go up, the number of draws usually increases as well, so it is even less clear who will actually "win" a given game.
Someone who is rated 2900 is more than 3 standard deviations above normal, which is exceptional for the population. But you couldn't say that this person's strength at Chess is any better than Kasparov, for example. If Garry himself were playing a prolific number of games on Chess.com (and not being a professional Chess player), he would likely have a rating similar to that. And on the flip side, if the 2900 player was playing in FIDE/USCF tournaments, because of the smaller pool of players, his rating would be similar to the highest players there.
It requires a very large pool of players, and a large number of games, to push a rating this high, and remain there. Also, since games against lower-ranked players do not help your rating as much, once you are close to the top ranks you need to continue to beat (and not just draw against) highly-ranked players to increase. This is just not practical.
This is assuming, of course, that neither player is computer-assisted. There are programs on the market right now that tout a "3000-4000 Elo", but they would require some actual games against rated players to see where they stand. There is a nice article on computer Chess engines (and their ratings) at:
http://en.wikipedia.org/wiki/Chess_engine
costelus wrote:
In order to get above 2600, one must defeat along the way many programs. So the answer to your question is the following: if the top players here were not be cheaters, they could easily outplay even the strongest GM's in the world. Ussually they make zero mistakes in hundreds of games and compared to them even the correspondence champions from the past look like kinderdargen kids. In reality however the top players here did not even hear of "ELO" and their OTB playing strength is well below 2000.
No, this is not true. They could still not be cheaters and the top players would still crush them. There is more to high-level play than just numbers. Also, at near-perfect play, it becomes much harder to determine what is a "mistake". You can win the numbers battle and still lose the game.
Sambirder wrote:
...you have to admit that the competition here is much lower than at the international level. ...So not necessarily are people who are not titled cheating on this site, its because they only have to play against patzers like me, rather than Carlsen, Topalov, and the other 2800 giants.
Accurate statement, and I would add that the population of casual players here likely far exceeds the number of serious-to-professional players worldwide. This makes it much easier for someone here to have a rating far above the site average.
HO!HO!HO! MERRRY CHRISTMAS; I WAS RELEASED FROM CHESSWORLD.NET BECAUSE I ACCUSED OF WINNERS USING AN ENGINE I REACHED 1795 ON MY OWN, AND HERE I'M JUST1329 ; THOMAS,
OUTSIDE OF USING AN ENGINE, I BELIVE CHESS GAMES ARE DETERMINED BY 50% STRADEGY AND 50% LUCK;
my FIDE rating was about 2200 when I quit playing congresses and tournaments, yet my number on this site seems to have levelled out at about 2100. Admittedly I've been out of the game for some time which may be a contributory factor, but I suspect that the main reason is that I was very good at managing my clock across the board and putting my oppo under pressure (I've beaten a few GMs and a lot of IMs). With the almost limitless time available here on the 'turn-based' my speed of thought is negated and my impatience is frequently my undoing. This is one reason why I believe it futile to compare any two different forms of chess.
And as regards time-outs - I've had 2 so far - both completely unavoidable (one came about because a hurricane knocked out the net for the whole area for more than 24 hours) - I don't believe they really affect your number in the long-term unless it is a habitual occurrence 
Addenda: I just checked my stats, an' both my best win an' worst loss were time-outs !! - it all comes out in the wash !!
tonydal wrote:
Niven42 wrote: No, this is not true. They could still not be cheaters and the top players would still crush them. There is more to high-level play than just numbers. Also, at near-perfect play, it becomes much harder to determine what is a "mistake". You can win the numbers battle and still lose the game.
How in the world would you--or indeed anyone--know this?
My point is that they aren't titled or professional players. Their rating or performance doesn't mean much outside of Chess.com.
Funny how they are not titled or professional players, yet some of them easily crush professional players (see for instance what happens when GMs like Becerra or Khachiyan "dare" to play against the anonymous geniuses in the top here).
We had the same problem in World of Warcraft. Arena ratings over 2000 were common, yet since the ratings were supposed to follow a bell curve, you would expect that most players would be right around average...
Except that what was happening was that teams were selling spots to those could afford to pay for them, and once you had enough wins, you could retire from the team, and sit on your 2000+ rating.
By going to a rewards-based system, some of these issues could be solved. Rewards-based systems are more intuitive, too, since most people expect their rating to just continue to increase over time, and think there has to be something wrong with the system if their rating falls. I've had to explain the Glicko/Elo system hundreds of times to players, but most people just cannot visualize it in their heads. After all, we wouldn't ever think that a 1200 provisonal player is "average", even though everyone enters the pool at that point.
The bottom line is, a 2900-rated player is so far to the right on the curve, it's really unlikely that they could achieve that rating without artificial help, yet it's also difficult to understand why, if they were really that good at Chess, they weren't playing professionally. Since very few professional or titled players exhibit this type of performance, and despite their prowess often lose to the highest players here, you would have to surmise that anything over 2800 here is engine use.
Although rewards-based systems (weighting the score based on the number of games completed) have their flaws, this may be the only way to truly weed out artificially inflated ratings.
IF A PERSON USES AN ENGINE , THEIR PROFILE WILL SHOW IT; SORRY, 50% LUCK 50% STRATEGY THATS JUST THE WAY IT IS; TOM
tom my friend how are u? :)
my answer is simple it does show on there profile
140 wins and 1 loose?? vs gm's and fm's ??
Niven42 wrote:
We had the same problem in World of Warcraft. Arena ratings over 2000 were common, yet since the ratings were supposed to follow a bell curve, you would expect that most players would be right around average...
Except that what was happening was that teams were selling spots to those could afford to pay for them, and once you had enough wins, you could retire from the team, and sit on your 2000+ rating.
By going to a rewards-based system, some of these issues could be solved. Rewards-based systems are more intuitive, too, since most people expect their rating to just continue to increase over time, and think there has to be something wrong with the system if their rating falls. I've had to explain the Glicko/Elo system hundreds of times to players, but most people just cannot visualize it in their heads. After all, we wouldn't ever think that a 1200 provisonal player is "average", even though everyone enters the pool at that point.
The bottom line is, a 2900-rated player is so far to the right on the curve, it's really unlikely that they could achieve that rating without artificial help, yet it's also difficult to understand why, if they were really that good at Chess, they weren't playing professionally. Since very few professional or titled players exhibit this type of performance, and despite their prowess often lose to the highest players here, you would have to surmise that anything over 2800 here is engine use.
Although rewards-based systems (weighting the score based on the number of games completed) have their flaws, this may be the only way to truly weed out artificially inflated ratings.
hi first of all thx for your comment! :)
can you please tell as more about this system and the exact way it works becaus
it sounds good, and with all the information about this sistem we may be able do turn to the stuf and ask tham to consider the usege of this new system combind with there regelar system.
thanks agen!
NOW I'VE BEEN PLAYING FOR OVER 40 YRS. , BUT I'M NO MASTER, LEAVE THE MASTERS TO THEIR ENGINES. AND I'LL JUST KEEP PLAYING;THOMAS
o and agen :) does any one know what whould be the difrence in the actual game playing between a player with a (real) 2500+ player to a 2200-2300 (real) player ( talking fida)
i know threre will be a big difrence but in what spacifec aspect? what does a player 2200-2400 miss?? i mean they play briliantly to! but will almost always loose to a player as KASPAROV or more accurite will always loose to such a uge player! but what does kasparov see that a player that is realy good 2200-2400 does not see??
bringing games as an example my help as see how good is realy a true master!!
Try playing against chess computers in that range and you will see that the tactical range of a computer and a master is very large. Computers calculate every single response before making a move. Grandmasters can only calculate up to a certain depth. Computers on the other hand can go much deeper.
Ex: Deep Blue was capable of calculating over 1 billion moves a sec.
http://www.chess.com/echess/players.html