Are FIDE and USCF ratings honored by Chess.com?

nate23 wrote:

ponz111 wrote:

Question: if someone rated 2500 in another venue and started at 1200 here and suspose he played an IM [rated 2300 at chess.com]  a two game match and each player won a game and lost a game--what would be the new rating for both players? [assuming they played no other games, of course]

Well, assuming the IM didn't lose a crazy amout of points, I would guess the IM would have a 2284 (if you play enough games,  16 points is most you can lose.) And I would guess the other would have about 1375

You are wrong in your assumption of max losses of rating points. I played a someone in 960 who was 500 points lower, I resigned as a gesture of kindness, after feeling bad about aboslutely crushing my new friend in the 5th straight game. I was trying to help my friend realize I valued their friendship more than wins and points. I was flabbergasted when I lost over 100 points....

If you're talking about ratings here, 16 points is definitely not the most you can lose (I know that from experience). 

ok, fine, but how DO the ratings of games work? [there must be a formula?]

ponz111 wrote:

Actually. it is a big turn off for players who have established a much higher rating--they for the most part want to play their peers.

If it takes say 10 straight wins to get to a rating you have elsewhere then that is 10 straight games wasted. And if it takes 15 straight games it is 15 straight games wasted.  And this is not reasonable to some high rated players. Why should they play on chess.com when they can play in other venues which are more realistic?

Also to have a rating of 1200 and play some unsuspecting person when your real strength is say 2200 that is unfair to your opponents.

There are more effective ways to do this than to start everyone at a rating of 1200.

I agree that it is a bit cruel to make a 1250 play a GM rated 1200. I am sure many people would say that it doesn't happen very often. Perhaps they are right, but I would like to see a system in place to place players more closely to their strength. I have been submarined by a few players in tourneys here. I looked many months lately to find they a several hundred point higher than when we played. With the current system, it creates circumstances similar to players sandbagging in USCF or FIDE to get an unfair advantage to win the tourney prizes.

I have noticed what appears to be people with multiple accounts and who are retreads. I think they do this to try getting not only their rating higher, but so that they get easier opportunities to increase it. Due to the rating system here, my rating did shoot up higher than it has been since. Due to my misplaced confidence, I took some lumps against players I would't have bothered trying to play. I am not saying coming up with a good test to really pin down everyone's true rating would be easy, but I certainly think the current system needs reworked.

chess.com uses the Glicko Rating System. The formular is quite complicated, an explanation from a staff member is in this article: http://www.chess.com/article/view/chess-ratings---how-they-work

Yes, it's complicated, but it works approximately as I stated already.  New player with 1200 rating (considered very unreliable) plays old player with 2300 rating (assuming he has many games, considered very reliable).  New player beats old player, then new player's rating will be well above 2300 (say 2500, still considered a very unreliable rating).  Old player's rating won't fall much, because we assume it was right to begin with and new player happens to be a strong player.  Now old player beats new player.  Old player gets back almost exactly what he lost in the first game and new player loses quite a lot (because his rating is considered unreliable and because if he was really 2500 he probably wouldnt lose this game).  They both wind up around 2300 which is what I said to begin with.  Now, it's possible I'm wrong, because I am not actually calculating this, but it's how I understand it to work.  There was a deathmatch where a master without a chess.com rating saw his rating fluctuate wildly while his opponent with an established rating saw little change.  I don't recall which one it was.

I said that after you play a number of games (around 250) their is a maximum of 16 rating points lost or won.

If you have played this number of games and want to test this, go to the live server, find somone with a 700 point difference, accept their challenge, and the rating change estimator will say (win +0, draw -8, loss-16.)

Yes, exactly.  People saying they've gained and lost more than 16 are still in the process of establishing a rating.  In that case you can gain and lose far more than 16.  Once the rating is established, it changes much more slowly.  This is the exact thing that allows a master strength player with a new account to surge quickly to their correct rating without killing everyone else on the way.

Interesting..... sorry Nate....I guess we were only partly right by the way we chose to present our ideas.....I guess it certainly depends....

By the way, here is the article with the math if you want it...

http://www.glicko.net/glicko.html

By the way, premium members can look up their rating derivation (RD) in the players statistics. That's the variable responsable for the amount of points given or taken at a loss, a win or a draw.

@RCMorea: Yes, you're theorie is all right, but I wanted to lead the way to an official explanation which is a little bit more precise and also explains the main difference to the ELO system used by FIDE and USCF.

Are you guys kidding? If you look at people's profiles like Anand's they made some people start with a higher TT or turn-based rating.

I think maybe they went ahead and gave him cutsies in line, Cyrus. 

yeah, I have noticed prior wcc getting either their demands met or biased treatment....maybe they were also affraid he would be insulted

Yeah, if I was a 2800 and started out here at 1200 it wouldn't exactly encourage me to play.