Wednesday, January 16, 2013

High rated tactics redux

I haven't repeated my tactical problemset with 2000+ rated problems since nov, 22th.
I used to have a succesrate of 100% and a solving time under 40 seconds for every problem by then.
Yesterday I repeated the set with a succesrate of 91%.
But the average solving time has increased to about 100 seconds. Which means that it takes much more time to recall the solution.

This seems to indicate that forgetting is preceded by an increase of recalling time. Or maybe we simply call "forgotten" the state where the recalling time has become unacceptable.

The balance of recalling by move sequence vs. by geometrical pattern (diagram) is changed from 60:40 to 50:50. This might be a weak indication that move sequences are more prone to fading a way in memory than geometrical patterns.

I will train with the set untill the succesrate is 100% and the maximum solving time is below 40 seconds again.


  1. Not related, but did you notice the total absence of media coverage for Wijk aan Zee in the Dutch Media?

    We have the World Champion and the highest ranking player playing and I haven't heard a thing!

    Tomorrow they'll play each other, I wonder if that'll make the news. I'm sure there is some football transfer rumour or speedskating marathon in some obscure village that will be deemed more newsworthy...


  2. In Holland there is no room for chauvinism and we are proud of that.
    The main problem is that there is no recent dutch worldchampion. We are still in the aftermath of Euwe's championship. It's time to get a new one. Cor van Wijgerden layed the base for that by creating a method that leads youth to mastership. Who can take matters (masters)to the next level?

  3. I would say: forgotten is a problem if the solving speed is "back to normal". A 1600+ rated player did solve 2400 rated CT standard problems just by thinking more than 1h+++ at every problem.
    The time, which is needed for "forgetting" is depending on the length, intensity and method of learning. That you can see at Empirical Rabbits blog..

    I think learned geometrical patterns "should" stay longer than move sequences. We are eye creatures. We can recognise "an animal" at a pictures in a speed, that its possible to calculate that only ~ 5 layers of neurons where used to do so.
    A sequence of moves is "abstract" if you dont combine it with something "natuaral" like a physical movement.
    Brainscans of Masters did show that masters are using parts of the brain which is usually used to recognise human faces to recognise chess positions. I think here is an important reason why adults "cant" improve. Children have a lot of unused neurons which can be used to do different things. A brain is getting rid of unused neurons by getting adult. Then its "only" possible to "reprogram" neurons which are usually used already in a different way.
    This way we "cant" add skill after skill but we can "only" replace one skill by a different one. Not AND but XOR. Something like: Either we see forks OR pins.

    I think we cant improve "realy" if we cant create layers of neurons to "analyse" positions. All chess knowledge ( memorised things ) dont help "realy" to get better. The memory has to become "hard wired" - automated - high speed using "new" non-chess-neurons.

    If i am right with these bad ideas above then the goal could only be to use the limited set of neurons best. Instead looking for many different rare tactics more common ones like forks...?

  4. That doesn't make sense to me. I don't think a grandmaster will have more neurons than we do. If he has not more neurons, your idea suggests that he lacks some skills that we have. I find that a ridiculous idea, to be honest. It's just a matter of effiency, not of brain mass. And if it is a matter of brain mass, we can acquire that too. You can use windows on certain hardware, or you can write a decent operating system. All on the same hardware.


    Generally, the number of neurons in the cerebral cortex increases until adolescence reflecting a growth of synapses.[2] A decrease in synapses is seen after adolescence reflecting synaptic pruning, and approximately 50% of neurons during development do not survive until adulthood.[3] Pruning is influenced by environmental factors and is widely thought to represent learning.

    There is much research done with "skilled" people. The brain is changing with training and the plasticity of the brain gets less with age. Example:

  6. I believe that it is clear that training up to a certain level with short term repetitions, taking a break, and then training until you get back to that certain level (again with short term repetitions) is inefficient. Short term repetitions contribute very little to long term memory. A better strategy is to have roughly equally spaced repetitions throughout the time interval. This will certainly give more long term memory retention per repetition. It also has the advantage that it weights your training more towards practicing finding solutions rather than just recalling them from memory, which should also should be good.

  7. @BK,
    it isn't quite like that. I use the training regimen as provided by CT SRS.

    There are a few interesting side effects. In the beginning there were a lot of problems I remembered solely by their move sequence. But I start to see more and more patterns in the problems, now I'm already familiar with parts of the solution. Overall there is a shift from remembering the move sequence towards remembering the geometrical patterns of the combination.

    Today I didn't recognize a certain problem, which I used to solve in the past solely by remembering the move sequence. Yet I could solve the problem pretty fast. Only at the end I recognized that it was a problem I used to be very familiar with. Now it was solved by seeing the patterns solely.

    There seems to be a relation to the degree I remember this dataset with tactical problems and my tactical sharpness OTB. But I might deceive myself here.

  8. About the changes in the brain of Susan Polgar:
    "In order to isolate the areas of her brain she uses when playing chess, Susan is given an MRI scan. There is an area at the front of the brain which deals with face recognition, allowing most people to remember a face in 100 milliseconds. Astonishingly, this is the very place where the experts find that Susan has moulded her recognition of 100,000 chess scenarios. Over years of childhood practice, Susan has hardwired these countless scenarios into her long-term memory and can recognise one in an instant – as quickly as someone might recognise the face of a friend or relative."

    The complete Video:

    About some other changes in the brain of chessmasters:
    Tracking blood flow in the brain to detect spikes of activity, researchers found that master players of shogi -- a Japanese game similar to chess -- use two regions of the brain to make critical moves.
    "Professionals are trained extensively for a long time, over 10 years, hours every day. This extensive training (may have) shifted the activity from the cerebral cortex to the caudate nucleus," the study's lead author Tanaka said.

  9. @Aox,
    I'm familiar with both the papers and the videos. I think you are too certain in your conclusions. More certain than the scientific papers themselves. The core point is: how much can an adult brain adapt. I don't know, but science isn't sure either. So I make matters simple and I'm going to find out by acting as a guinee pig myself. You can be right or you can be wrong. The only thing you can't be right now is certain of the outcome beforehand.

  10. I am not definit about this, then i would stop to play chess. This Hypothesis could explain why its so hard to improve as an adult.

  11. Here a reminder about adult improvement in chess:
    Tempo and I could improve in chess.
    I dont know if we ever can attain master level, but at least we did improve considerably. Despite this neurons-issue, this is prove enough for me that adults can improve.

  12. you are all crazy people. without any doubt about it! love to you all. dk

  13. @DK, I assume that "persistent" is the word you are looking for ;)