Rating Problems

[tacto]

It seems that the more problems someone does the better his rating gets. I am not sure why this is but if you check out a graph of any active player it will most likely be progressing upwards... This makes ratings somewhat uncomparable if the players haven't each done the same amount of problems.

Why is this? This isn't as apparant on my graphs though.

[drahacikfm]

Two possible reasons that people's ratings go up as they do more problems:

1) You get more used to the site, and you get more used to the types of problems that are here, and learn that you have to make a great move, not just a good move.

2) Maybe you get better by doing problems

[tacto]

I agree with "You get more used to the website"

that is very true. But I have been using the website for waaaaay too long for that to be an issue, I only recently started on standard ratings...

look at my graph: http://chesstempo.com/chess-statistics/tacto

or Tamas graph: http://chesstempo.com/chess-statistics/tama

[Richard Says: I just edited the above links, usernames are lowercase only]

If infact you are correct and this has nothing to do with the website... we still have to face fact. That is, it takes way too long to get an accurate rating.

I just wan't to make it clear that I am not at all bothered by this I am just curious as to WHY this happens and trying to help out.

I guess this could be normal in some way. On cts it is well known that when you use the website for the first time for a while you get improvement easily then it stops.

I have no idea how to explain this. I doubt it's that important anway but if someone has some interesting ideas let us know lol.

[richard]

Hi tacto,

I think what you are seeing is probably related to the problems having relatively high RD which means it can take a long time to settle towards your natural rating.  If you sort the users list by number of standard problems done and look at the top 10-20 users then you can see that there is not a common trend on the ratings graph, some are up, some are down.  I've recently slowed down the rate at which problem's RD increase over time which should mean users can settle more quickly towards their natural rating (high problem RD == lower rating adjustments for user).

Regards,
Richard.

[tacto]

That does make sence.

Though I still think there is something else maybe aswell.

Maybe it's just my imagination.

[drahacikfm]

I think as you are here longer, you learn to "psych out" the problems better.

For example, I've only recently figured out that if you see a move which will result in you being up a knight for a pawn at the end, then that is NOT going to be the right move.  I look for another move.  Because so far I've never seen a problem where you are only up 2 points in material at the end.

Another trick is, if you see two different moves which leave you up a rook, and those two moves look almost exactly equal in value, then the answer is NOT going to be one of those two moves.  You must have something better that wins a queen or mates.

 As you learn these little tricks more and more, your rating goes up.

These tricks are useful here, but not so useful in tournament chess.  In a tournament, if I see two moves that win a whole rook with no counterplay for the opponent, then I just play one of them.  I don't think to myself "that can't be correct, there must be a mate".

[tama]

Im sorry Drahacikfm i have to disagry on your method on solving tactic. The point of every problem is to fined the BEST move in the possition.

Let me give u a real hint ( If u c a good move, Look for a Better one!) Its stupid to have the mind set thinking -Oh ON this cant be the right move because ill only be up 2 pawns! i need to win 3 at least.....- this is just rediculas.

[tacto]

tama, a fide master would know all that.

He is just making a point. I agree that there is maybe some pattern but I think the big problem set would take care of it.

[drahacikfm]

tama:  read my post again, and maybe you will understand it this time.  Your post shows you didn't understand what I wrote.

By the way your hint "if you see a good move, look for a better one" is about the most common piece of chess advice.  It's nothing new.  I first heard it 40 years ago.  And that's exactly what my post says:  look for a better move in the situations I described.  Try reading it again.

[tama]

lol sorry i didnt know u were a fide master. When u fined a good move look for a better one, i made that up even when i know its common scence.40 years?!?! i wish i hade your expireants.

[richard]

Tama,

I think the point most people make about the "When you see a good move--wait--look for a better one." [a quote attributed to Lasker] is that it usually applies to situations where you haven't found a move you are 100% sure wins.  If you've found a move you are 100% certain will win then there is little point in looking further when in over the board play.  There might be an argument that for this site, having two winning lines isn't that bad as usually the "best" winning line is better in evaluation (I've tried to run this argument myself in the past :-) ).  The problem with this argument is however that it ignores the use'rs experience.  If the user works hard to improve their rating then gets a rating drop after choosing moves that would have won them the game in OTB play , then it can be quite frustrating (and I imagine demotivating).

Some users don't mind the alternative winning lines and don't seem that bothered when they lose points on positions with alternatives.  However enough users are annoyed by this, so the next set will definitely have a lot less of these issues (where alternative winning moves are much more reliably detected and will not lead to a problem being marked incorrect - when choosing an alternative winning line the user will be asked to keep looking for another move).

Regards,
Richard.

[tacto]

Yeah, but its still good to learn to see the best first even if it's just a move quicker or whatever lol.

[tama]

I actualy feel that not changing the way problems are right now is better then making every winning line over the board not drop the users rating because it will seriosly effect all ratings with a huge increase. LIke right now i worked and am working hard to improve and get my rating up, but once people can simply c other winning lined that arent the original ones i just feel like its abit to cheap and easer.

 I like how it is now because its tough and i dont want some easy to get high rating i want to earn it!

[tacto]

Tama, your logic is off.

If the ratings do adjust becuase of...whatever then becuase its a RATING system it will still be equallty hard to earn ratings. Since if a problem becomes easy it won't give the solvers much rating anyway.

[tama]

Yes i know but all the problems will become easyer for everyone, u no what i mean.