Chess rating system

oh ok makes sense pretty interesting
Would you say that you experienced the Glicko system to work fully symmetrically, i.e. will you end up on the same rating if playing higher rated players 10 times and beating them, let's say, only once, compared to playing lower rated players 10 times, winning 9 out of 10 times? I like to play both higher and lower rated players, but it would be interesting to know if there would exist a strategy to increase your rating by focusing on lower OR higher rated players. That would constitute a glitch, IMO...

Interesting idea, Clownfish. The formulas could be used with some sample data to determine the results. But to use that approach in real life, as a "strategy" per se, would require you to know not only the ratings of your potential opponents but also their RD values. Then you would have to run the numbers to determine the best possible sample of opponents against which to battle. Even then, there's no guarantee of better success against the lower-ranked sample group.


And what if the higher-ranked players have very high RD values and the lower-ranked players have very low RD values? Then the results for the two scenarios you describe above could be equal or even reversed!


I offer the following spin on your idea: We have no way of affecting our opponents' RD values. But we can affect our own. By definition, the RD value will gradually increase with time. So if we take extended leaves between groups of rated games, that would affectively cause our RD value to increase before each group of games. Therefore the ratings swing caused by our game results would be more dramatic (which could be a BAD thing). But we would need to know what time constant the USCF (or FIDE or whomever) uses for the RD equation. 2 months? 6 months? 1 year? 2 years? And of course we would have to stay "up on our game" during those sabaticals.


In the end, I don't think there's an easy way to minipulate the formulas or the system. There may be a way, but it's probably not very easy. Or we could try to win every game. :-)

Chess rating ELO is hard to improve.
Ratings...Shmatings...Get to playings, fools! 
But then I wouldn't get to play around with the math formulas. LOL.
LOL it is kind of easy too.
High rated is good!!!!!!!!!!!!!!!!!!
joeyson wrote:

wow ppl

just play

ratings dont even matter

 Ratings are important to people, it justifies their abilities, thought starting with a 1200 rating does not accuratley portray the skill of the individual.


i'm rated 1200 and i don't play golf

come get me

thats how u start.
I think ratings do matter to serious chess players. I have never met one that it didnt matter to. Are you a serious chess player? I consider myself a serious chess player but the only rating that matters to me is my uscf and fide otb classic chess ratings. Online ratings dont matter to me at all and neither do blitz or rapid chess ratings.


I'm pretty new to chess and I'm not familiar with the uscf & fide otb classic chess ratings.  How does one get rated under these systems and how are they different from the on line ratings?

This thread is humourous - i was puzzled at the "Why then did I go down from 1200 (to 1361)" - then i read on and realised that the poster thought ratings were like golf handicaps. Broke into a smile - (i was feeling somewhat tense for reasons not related to websites) - and feeling relaxed - it was theraputic! Thanks, MickBJ!

  (the above is meant seriously! i am *not* taking a dig at anyone. Perish the thought...)

By the way, internet-chess ratings (say fics or icc) matter in a way - it you play too casually for a long session your rating falls abnormally below your "normal" at the relevant site - so you tend not to get opponents against whom you can play a good game (until your ratings pick up again).
An interesting aspect of ratings is that you can estimate about how often one player should beat another.  If they are 100 points different, for example, the stronger player should win about 64% of the time, 200 points about 76% of the time, 300 points about 85% of the time and 400 points about 91% of the time.  Gives you a lot of respect for Kasparov's record rating of 2850.  Of course, you shouldn't be thinking about this when you are playing since even against a player rated 400 points higher you still have a chance.
justice_avocado wrote: 1361 is higher than 1200. this isn't golf.
so we're not sposda get our ratings to zero!!!???
fischer wrote: AlecKeen wrote:Becca wrote:Rating has its place but its not the most important thing. Sometimes you can lose a game on time and it will seriously affect your rating this has nothing to do with how well you play.

Oh yes it does! How well you play includes how well you manage your time. Time is as much part of Chess as it is in other games. In football you could score the greatest goal in history, but if the referee blows time before it goes in it doesn't count. Similarly in Chess if you don't get your moves in within the time, you lose, and correctly so.

 I could be wrong, but I assume she's talking about blitz games. There are lots of people who are great blitz players but terrible in long games, and vice versa.

 I try to play differently in blitz games. In long games I try to play (what looks to me through my myopic chess goggles) the best move, whereas in blitz games I try to play moves that make the opponent think

erik wrote:

everyone starts at 1200. then as you play you get a new rating. it is all based on the Glicko ratings system :) check it out - it's a fun read! 

 The problem I have after reading this is Glicko assumes the rating of a player who has been out of the game for a while is not reliable. Let's say that I play chess for 20 years and earn a 1700 rating then stop playing for 20 years. A new prodigy comes up and after 2 years of playing has a 1700 rating but his rating is earned more recently than mine was. That would make the prodigies rating after only two years of play more valid than my rating earned after 20 years of play because I haven't played in 20 years. I could have 1000 professional games under my belt and the young prodigy only 100 so how is my rating less valid than his? I would consider the newer players rating the less reliable because he has fewer games. Time should have no effect on how many points someone gains or loses from a match. If we could bring J.R. Capablanca back from the dead would anyone say that his rating is less valid than Vladimir Kramniks?  


I'd like to know what is going on with the tactics trainer rating, if anyone else is amused by how the trainer rating changes, or if it is just me

First I went up to about 1600, then suddenly plunged to 800 in the space of about 2 days, then zoomed back up to 1700 before drifting down to less than 1400, then raced up above 2000 in about 48 hours before drifting back to about 1900

The actual rating is less important than the fact that it seems to be moving all over the place. Is my skill in tactics really changing that much over the space of a couple of weeks (and in some cases a matter of days), or is the tactics trainer on drugs?

Don't get me wrong, I think that the tactics trainer is one of the best features of the site, but I would prefer to see a slow and gradual increase in my rating that reflected a real improvement in my tactics, rather than a completely bonkers, helter-skelter voyage all over the place