Monday, July 10, 2006

Inside the mind of Temposchlucker part I

CTS helps to improve your OTB skills.
For me that is a proven fact.

But it is clearly not a very effective method.
You have to work like a frantic dog, doing 50,000 problems in 600 hours to get some measurable results. There must be a better way.
In an attempt to try and find it I do some brainstorming.
I'm just thinking out loud.

My startpoint is an article from about a year ago:

A new study discloses the fact that it is not only lots of study and practice that is essential to master chess, but experts and amateur players use different parts of their brains during matches.

Research by scientists in Germany involved scanning the brains of 20 men as they played against computers. Half were grandmasters the loftiest ranking in the chess world while the other half were merely good amateurs who'd practiced and played for at least 10-years.

It was noticed that for a few seconds after each player made a move, tiny bits of energy called "focal gamma bursts'' appeared in his brain.

Grandmasters had relatively calm medial temporal lobes - an area thought to be crucial for performing new tasks and establishing short-term memory - during the games when compared to amateurs.

In amateurs, "focal gamma bursts" were most prominently detected in the medial temporal lobe and in grandmasters researchers measured the bursts most often in their frontal and parietal cortices, parts of the brain linked with long-term memory and the ability to perform complex motor skills.

The results of this study, which was carried on by the University of Konstanz, indicated that experts and amateurs use their brains in fundamentally different ways during a match.

The lead researcher of this study says that precisely how and why brains are used in different ways remains to be explained. He however explained that with each new move, amateurs must focus on analysing what essentially is new information. While experts are able to retrieve "chunks" of long-term memory while relying on brain circuitry outside the medial temporal lobe.

Grandmasters while studying chess get to see hundreds of thousands of different (chess board) positions. They store these 'chunks' in their brain. When they see a similar position, they retrieve the data from their memory.

Lead researcher Ongjen Amidzic said that amateurs never encode in long-term memory the information they cull over the years from a chessboard. Why remains a subject for additional study, he said.

A neuroscientist at the University of California, Los Angeles said that this research appeared to be the first one to distinguish between experts and amateurs. And the results themselves were not entirely unpredictable.

Reading in Wikipedia revealed the following:

Depending on storage time the memory is divided in:

  • Sensory memory (milliseconds - secs)
  • Short term memory (seconds - minutes)
  • Long term memory (days - decades)

The long term memory can be divided in:
  • Declaritive (explicit) memory
  • Procedural (implicit)memory

What's in declaritive memory can be explicit put into words. It requires conscious recall.

In the procedural memory there is implicit "how to"-knowledge stored. Recall of information works unconscious. It works more on a stimuli - response basis. You can't put it into words.
Complex motor skills are handled by this memory.

The article seems to suggest that grandmasters record their knowledge the same way as one learns new complex motor skills, that is in their procedural memory. While an amateur stores his knowledge explicit in his declaritive memory, where it is of no use when playing chess.(They only can talk about chess with this knowledge:)

I think this is what is meant by DLM when he talks about the difference between chess knowledge and chess ability. I remember lively discussions amongst the Knights about the difference between memorization of problems vs. plain pattern recognition. The base of the discussion was that most of us felt that these two "tasted" different.

I have studied SOPE in an intellectual way half a year ago. Now it is put to the test by pawn ending problems at PCT all this study proves to be of no use. I still score terribly on pawn endings.

So I think we do best when we look upon learning chess as acquiring complex motor skills. (Now I lose the romantic part of my readers:)
J'adoube describes it beautiful in an old post of his called Tetelestai:

Blue Devil Knight asked about percentages. If I take my time, then 100% correct. All of them. All the time. How could they not be? I've done these problems so many times it really is like playing a video game. I can do Level 10 in 18 minutes. That's smoking.

Exercise until you walk thru the problems like playing a video game. That's the way to go. I believe that the much critisized time constraints of CTS to solve a problem within 3 seconds actually is its strongest point. You can only solve a problem within 3 seconds when it is handled by your brain as a complex motor skill. Stimuli - response. A tempo:)

What's the way to acquire new motor skills?
I found two things: repetition and visualisation.
I experimented with visualisation by playing thru the solutions in my mind with my eyes closed.
At the time it seemed to work. Looking back in retrospect I don't think so any longer.
On repetition: it's not clear if the method of spaced repetition is equally suitable for procedural memory as it is for declaritive memory.

(to be continued. . . .)


  1. Wow. . .you had to really reach back into the past to dig up that old post of mine. . . [grin]

    Anecdotally, I find repetition is the only way to retain long term. In my kendo training, we do thousands and thousands of repetitions of the same cut over and over - building "muscle memory" - the synaptic pathways are being established and reinforced with each repetition. For me, pattern recognition is no different in chess, hence I do the same in chess training.

    The problem I have is that even though I have long term retention of tactical themes, if I come upon new tactical problems, I don't solve them quickly a la CTS, but methodically. Sometimes what is entirely obvious to another player takes me longer to identify.

    I have noticed, however, that the more games I play the quicker solutions come to me.

  2. I have thought for quite some time that the reason older players really struggle to improve like their younger counterparts is the time factor. Kids who reach IM or GM level must have played and seen thousands upon thousands of games, the average adult beginner is way behind on that front. If you see 50 games on the Ruy Lopez berlin defence then you can usually knock out the first 10 moves OK, if you reach 1000 then probably around 15-20 I would guess. It becomes a repetitive task, and moves elsewhere in the brain like you say.

  3. i like the analysis and the scientific way of approaching the problem of chess improvement. i must say, your argument is pretty convincing, at least to me. I may give CTS another go. :)

  4. Jim, you would be surprised to know what's in the caverns of my mind. . .:)

  5. hey tempo. i was curious what your method is for cts. do you try to get every problem right, no matter how long it takes? or do you try to get the answer in three seconds? i try to get the answer within three seconds, but i'm wondering if that's how you do it to. if i want to emulate you, i have to do it right. :)

  6. GK,
    it's a bit of a middle way, as my success rate (81%) indicates. Actually I do better now (86%) but it takes a lot of time before the original success rate gets there from its original 79% if you have a great amount of problems done. I have to feel pretty sure about the answer before I execute the moves.

    Since yesterday I started a new experiment. Because I look upon the learning process now as if it were a complex motor skil. Because of that I do now every problem 3 times in a row before continuing.

  7. i'm curious tempo, do you think it is possible that you know chess strategy and openings well enough that all it took for you to improve was to work hard at tactics? you said previously that you read strategy books by nunn, etc., but your game didn't improve. is it possible that at the time your lack of tactical prowess was holding you back? it seems to me that this is highly possible, and i'm curious what you think.

    and btw, i'm posting again if anyone is still reading my blog. :)

  8. GK,
    no that's not the case at all. MDLM was right on the money in his article when he said that all the positional, strategic openings knowledge in the world is worth less then the ability to think one ply deeper.

    The problem is to convert knowledge into skill. Otherwise it is useless.

  9. an interesting post.

    RE: Lead researcher Ongjen Amidzic said that amateurs never encode in long-term memory the information they cull over the years from a chessboard. Why remains a subject for additional study, he said.

    I imagine us us amateurs who play often must encode SOME positions in this way. Like the first few moves of an opening. But far fewer than a GM. I wonder if some of the problem is amateur when playing a position haven't studied to a point of knowing what should be done in certain positions. A GM sees a position and knows he has studied it before. I sometimes recognizes positions but I don't am uncertain what is best play. This requires me to calculate.

  10. The question we have to answer is: how can we train so that we DO encode the information in the procedural part of the brains? All methods that don't adress this are going to lead us nowhere.

  11. Hey Tempo, it looks like we are online at CTS at the same time. i'm curious, what is your rating for OTB chess? because if you are in the 1800 range, I would have to say that you are probably much stronger in the endgames and strategy than I. We have a similar CTS rating, but I think that you are probably a better overall chess player. I am guessing that you were already skilled at the other parts of the game and needed just a dose of tactical skill to help you out.

  12. GK,
    I have a dutch rating OTB of 1751, which is comparable with a USCF rating. Strategic I was allready a strong player, my endgame skills are still very weak. But I have learned how to avoid endgames.

  13. So you were good strategically, and endgames never came up so it didn't matter if you weren't good at, since you were weak tactically, doesn't it make sense that all it took for you to start crushing people was to master tactics?

  14. GK, if you want to put it that way, yes. What are you up to?

  15. well, i have just started to actually train again. for the longest time i was just playing chess. but i'm doing some training on CTS. i'm shooting for at least 100 a day. my rating is pretty good right now and my RD is low, so it could take a while for my rating to change. but i enjoy it (it's addictive). i'm impressed by the number of problems you've done there. hopefully i'll come close to that number someday!

  16. hi tempo,check this site
    maybe you will find there something interesting.

  17. Wouldn't it be easier to develop some sort of focal gamma burst suppression pill? Or maybe some sort of anti-gamma ray death helmet? This would reduce all the neutrons and cut down on the voices in your head telling you to play 2.f4

    Just kidding. Interesting post and it very much brings us back to the true question/importance of pattern recognition.

    I have not hit cts in some time. My focus has been on....well fishing. But when I have sat down, it's been for endgame study. PCT's repetition style is very nice and I think helps with pattern recognition.

  18. hello again. we're both on cts at the same time again! do you get paid for doing the CTS problems? :) how long are you on there every day? just curious.

  19. oh, i just got the meaning of active on everyone's profile! i guess i just got lucky this morning when you actually were there. i'm looking forward to part II of your inside the mind series!

  20. Reminds me of an older post. There, I said:
    Typically, they break memory up into 'procedural memory' (e.g., remembering how to ride a bike), and 'declarative memory' (e.g., being able to consciously recollect and describe facts such as the color of your first bike). These two types of memory correspond to two different types of knowledge, knowing how to do something (e.g., I know how to throw a baseball or do a Judo throw), or knowing that something is the case (e.g., knowing that your first bike was a Schwinn).

    I like de la Maza because it is clear that for experts, much of chess is know-how, tapping into the vast and unconscious procedural memory system that is built up via extensive experience. It is not a special ability [experts have] to think faster.

    I think everyone would agree (and this is what the Russians knew for many years) that the key is repetition. Tons and tons of practice: chess is like a huge, complicated, bicycle. It is just really, really hard to learn to ride it well, and some ways of practicing are better than others. E.g., spending two years drilling on weak squares is probably not as smart for a beginner as drilling on tactics.

    From people at ICC who have played thousands of blitz games and still suck, apparently playing lots of games really fast isn't the way to do it. However, I bet that playing tons of very slow games would do wonders. If we only had the time.

  21. Pale,
    did you catch anything worthwile to mention?

    at this moment I estimate about 2 hours a day.

  22. Blue,
    now I understand your comment on my old post better.
    If I understand you well then if you train a rat to do something, you train its procedural memory. How do they do that? With punishment or reward or something? Can we people learn something from it? I.e. electroshocks after each wrong answer or something like that?

  23. Ivan,
    thx for the link. I read a few articles there, which strenghten the idea that what and how we are trying to accomplish things has a scientific base. That's reassuring.

  24. "Pale,
    did you catch anything worthwile to mention?"

    What? You didn't find the idea of a gamma burst suppression pill important?! Tempo, I was just trying to be funny. Maybe it didn't come across that way.

    Now for a legitimate contribution to the discussion:

    IM Josh Waitzkin dropped chess and took up Tai Chi to become a high level competitor in that sport as well. In an interview he described how similar the two disciplines were as far as how he processed things in his mind. He spoke of Tai Chi and chess as though they were languages. Indeed, he said while in the middle of a simul (a chess excercise that no doubt requires one to use blocks of information) he felt he was thinking in the same way as when he performed Tai Chi. Jim has already said how his kendo training involve repetition to gain muscle memory.

    I think it is striking that an IM player can also excel at a sport that requires similar memory processes. I guess the real mystery is if you can "rewire" your brain to approach chess in a more language/pattern recognition way. Does repetition really succeed at that? A 10 year old chess prodigy hasn't had, by virtue of their age, the same amount of exposure to the game as a forty year old serious amateur, yet they somehow take in the information differently. You could do all the repetion you want, but if the information isn't being stored the way it should, you'll lose it. That's a bleak outlook but possibly relevant.

    Anyway, I need to get back to my shop and work on my anti-focal gamma burst helmet. ;)

  25. Pale,
    LOL "did you catch anything worthwile to mention?" at riverside, I meant.

    But to your other catch, if DLM can do it as an adult, I can do it as an adult. I'm going to find out how.

  26. Tempo: we train rats with the tools of operant conditioning, which indeed involves reward in our case (we deprive them of water and give them drops of water when they do what we want). Punishment (negative reward) is also possible, but a less effective training tool.

    We could market a new chess training device. When you get the tactic right, you are shown an erotic picture. When you get it wrong, you get a little electric shock to your John Thomas. That would be pretty motivating :)

    PMD: your question is great. Clearly repetition isn't all there is. I was speaking more about building up pattern recognition. The way to do that is lots and lots of repitition, but some people are just better at picking the stuff up than others. This is probably partly genetic, partly a result of how your brain was shaped during your development. But if we could fix all those things in multiple subjects, the quality of the repetition (I predict) would be key (i.e., the set of patterns you chose to learn in the first place).

    Again, what we need is funding for a 10,000 person study where we try a bunch of different techniques to teach them chess. One group starts with 100 blitz games a day, another does tactical training. Each month, the different groups play each other in a Round Robin. :)

  27. For Convekta the smiley faces and red X's are good rewards and punishments :)

  28. dirk, very nice post. thank you. i love your stuff. you help us all. my comment concerns this VERY provocative statement by you:

    "After 3 years of doing tactical exercises, my calculation ability has reached its top. I can't imagine there is something to gain for me along the road of calculation improvement anymore."

    i agree but a few qualifiications:

    this is really not to say you CANNOT IMPROVE your calculating abilities, even if it could be 90% correct, to say so. this really has to do with "deminishing returns." a system can get so large or complex or ladden with previous history, that the cost of changing it goes up exponentially.

    i recollect bovinniks famous statement in One Hundred Selected games quoted by timman in his Art of Chess Analysis, about going into the country to breath the air. in this great passage, the mighty bottvinnik recounted how he became a candidate master, a master who beat other masters, a grandmaster, a master who beat other grandmasters, then a grandmaster WCC contender.

    we all have had situations where we took two or three or even four months off from internet rapid, only to come back, and play STRONGER, or SHARPER or DEEPER.

    so the idea of switching contexts entirely is one we all can use... from tactics to endings, from fast to slow OBP, from overexertion to rest! thx, david

  29. Trans,
    It's not very relevant if my calculation skills are at 90% or 100%.
    The message I try to bring forward is that at a certain moment the improvement must come from a different source than gain in calculation skills. Scientific research revealed no significant difference in calculation skills between grandmasters and able amateurs. So their advantage has another source.