Friday, July 07, 2006

This little post of mine

Today I had a look at an old post of mine.

That yielded a few figures:

#probs rat % recognized
35,000 1520 32
50,000 1540 85


What use is it to repeat the same problems over and over again?
It are pretty simple problems.
Because CTS only presents me with problems that I can solve within 10 seconds at average.
What can be the beneficial effect of trying to solve problems that you can already do in 10 seconds so fast that you can do them within 3 seconds?

When I put it this way it looks like nonsense.
But the strange fact is there that (grand)masters at CTS score much better than me.
Which means that they DO recognize most positions within 3 seconds.
The question arises, is this a mere side effect of their training which contributes little to their performance as grandmaster (except for blitz and simuls) or is it the core of their skill?

I don't know but I intend to find out.
The meaning of the rating improvement from 1520 to 1540 is that I stored another 1000 problems in my system so that I can do them within 3 seconds in stead of 10.

12 comments:

  1. I think you shouldnt look at rating and skill as being so closely connected....

    ReplyDelete
  2. That is an interesting question. In fact, I got sidetracked myself and spend now time reading some psychology articles instead of doing chess problems. That won't improve my chess but I do it for the fun of it anyway. In my opinion, 50000 problems are too many to efectively learn at once because you will have to solve an awful lotof them to get CTS to repeat them for you and repetition is necessary to memorize it. I am not sure, how much memorizing a few positions help to become a better chess player, though. I am speculating that it is not so much the complete position that is recognized but merely the position of a few key pieces relative to each other. Therefore, knowing a few hundred simple positions may be good enough with progress slowing down if you memorize more. After all, even Superman wouldn't be able to memorize all possible chess positions. It seems that even GMs know "only" about 50000 positions (see this article in my blog). Therefore, I think you may be better off restricting yourself to a smaller set of problems.

    ReplyDelete
  3. Studying Polgar's mate-in-one problems five times has increased the speed at which I find mating threats. I scored 303/305 in 115 minutes on my first try and 305/305 in 43 minutes on my last try. My CTS rating jumped 36 points and I spotted a mate-in-four in a real game so I agree learning small problem sets well can be fast and effective.

    ReplyDelete
  4. Yasser,
    would you be so kind to explain your comment a little more?

    ReplyDelete
  5. Sciurus,
    It not 50,000 that I'm repeating but the problemwindow of 10,000 that I'm trying to master. That is still way too much, but that's why I work with such rabies theologorum. In fact already 85% of the problems look familiar. I read the article at your blog, thx.
    Restricting to a smaller problemset would be wiser indeed.

    But now I'm so far I will try to finish it.

    ReplyDelete
  6. The super high rated CTS users are all cheaters, aren't they?

    ReplyDelete
  7. crptone2 is a cheater (James Sun from Brookline Massachusetts). Uraleech is a master who has reached the point he knows all problems a tempo. At the edge of the bell curve his rating is no longer reliable. Chromefeather is a cheater. The rest I don't know.

    ReplyDelete
  8. I'd have to agree with sciurus that 50K is too many, the fact GM's have around this many stored away is due to the ridiculously large number of games they have played, studyed and watched I suspect.
    The one question about this learning process is;
    If driving becomes a repetitive process and becomes controlled buy the lower brain then can chess tactics become the same? If you drill yourself enough can you shift the part of the brain that does the work?

    ReplyDelete
  9. DrMunky,
    I'm convinced that that is the way it works.

    ReplyDelete
  10. Space,
    I appreciate the great pains you take to read my posts.
    It's true that I often omit the word "estimated average" when I talk about estimated average rating at CTS. I do this on purpose for a greater readibility. When I talk about a subject for the first time, I try to say things correct, but after a few posts, when everybody seems to understand and to agree what we are talking about, I use more convenient language.
    I want to prevent it becomes unreadable.

    The lowering of my highest rating ever from 1567 to 1564 isn't so strange as it may look like.
    My highest rating ever was 1567 indeed. But then CTS started to record the highest rating. And that was lower. So the correct formulation "should not be "highest rating ever" but "highest rating since (date forgotten)". For convenience and readibility I continued to use "the highest rating ever".
    I don't think the core of my statements is lost by this free formulation of things. Asking for clearity is always welcome ofcourse.

    Spacecowboy, I invite you to try to understand the essence and the intension of my words rather than to take things to the letter. I would appreciate that.

    ReplyDelete
  11. Space,
    I don't feel offended or impatient since I don't have the feeling that you are trying to pester me.
    But I find your comments hard to follow. You have to know that English is not my native language, so subtleties are easy spilled to me.

    I can't see what's wrong with my conclusion when you replace "rating" by "estimated average rating".

    ReplyDelete
  12. Space,
    now I see what you mean. Yes, most (if not all) of my conclusions are weak. That is because I make a lot of assumptions. I even like to call them hypothesisses (?plural) which gives a taste of a scientific approach. But my posts are not scientific of course. They are practical. I make assumptions, I draw an (extreme) conclusion and put it to the test. These tests are for instance to measure how the figures develop overtime, to do experiments based on the conclusions or to carefully read comments of readers on my conclusions.

    I like to take the things to the extremes. Because there is shown the best if assumptions are right or wrong. That's why I do an extreme amount of problems at CTS. In the end I hope to come to a definite conclusion about CTS and its method.

    Space, you have (quite)a different look at the world compared to me. That's why I have difficulty to follow your comments. We both have a blind spot in our looking to the world. But your spot and my spot is different. Because I want information about what is in my blind spot, I always try to listen carefully when I don't understand somebody. Because it can give me circumstancial information about the things I miss. What you do with your blind spot is not my business of course.

    One thing I hate btw is discussion only for the sake of discussion.

    ReplyDelete