Monday, March 16, 2009

Answering Wormwood

Wormwood said:

I wonder how motion patterns fit into all this. I have no idea what the english term for them are, but I mean the procedural muscle memory related to any complex physical movement. like throwing a ball. - studies show those patterns develop roughly before you turn 14 or so, after which it becomes very difficult to acquiring new patterns. which is the reason why a right-handed adult has extreme difficulties learning to throw with his LEFT hand.

so, when we work these chess things right into muscle memory, we're actually offloading the task from our conscious mind, much like what you're talking about, but into a different type of unconscious memory.

now, the question is, are visualisation drills neglecting this resource completely, as the processing of problems happens without physically moving any pieces. or, are the drills still having the same effect motion pattern-wise, due to simply THINKING of moving the pieces.
I'd be inclined to think the latter is true, as we do know simply even watching someone throwing a ball excites the very same areas of brain as if we threw the ball ourselves. but it's a question to be asked, and a distinction to be made. there is a difference, and in some type of training it might matter.

loosely related to that, I went on ICC today, and tried mating KQ vs KR. it's been months since I've seriously drilled it the last time, maybe even half a year. so I thought I'd probably fail miserably, since I can't really say I have it in my active conscious mind anymore.
but the results were quite surprising: straight from the start I realized I couldn't really visualize much any of the important squares anymore. I couldn't see which squares were controlled etc, instead it was a sea of square confusion with the pieces floating in the middle. just like when you're really tired and just can't force yourself to see the board properly. -but oddly enough, the moves actually came to me without much trying. I won more than half of the endings, which is quite close to when I'd been drilling the ending daily for a couple of months. and the weird thing is, I mostly couldn't rationalize WHY I made the moves, but they just came out correct. I couldn't see properly, nor give reasoning for moves, yet I kept mating the computer. it was almost like my hand was possessed, and something outside me made the moves. and I think it must've been muscle memory from drilling the mate hundreds (thousands?) of times.

As usual you penetrate to the core of the matter right away. The ultimate consequence of my reasoning is that there is no such thing as automatic learning by repetition. There are:
  • Training. The application of the exercise at hand.
  • Feedback. The correction of errors with consciousness of sufficient intensity.
  • The Miracle. The mysterious unconscious hence automatic adaptation of your brain to the task. It happens complete out of sight.
If you look at my improvement "career" of the past years I consider it to be proven that unconscious automatic feedback doesn't work. Since it doesn't correct anything. No matter the kind of exercise. The fact that bishops are much harder to visualize than rooks even after thousand or more games illustrates that automatic feedback isn't feedback at all. How could it be?

The fact that you can learn how to drive a car in just 40 - 100 hours shows that little repetition is needed for feedback. And indeed, if feedback is conscious, why should it be more?

So where does multiple repetition fit in the story? There are the following possibilities:
  • Adding precision. Feedback itself needs little repetition when performed with consciousness of sufficient intensity. Maybe 7 - 10 times or so. But there are all kinds of details to be discovered. And every detail needs seperate training. You must follow the right melody. You must sing the right pitch. You must sing the right pitch with each of your 3 voices. The overtones must resonate. Your tone must blend it with your neighbour. Your tones must blend with the choir. You must adjust your vowels. You must adjust your breath etc. etc.. Good technique is infinite. So repetitions are infinite. But not more than 7-10 times per subject.
  • Partly obsolete. When you train on autopilot, every now and then you will train conscious by accident. Especially when matters are new to you. But overtime the repetitions will become more and more obsolete when they become more automatic. But since you act on autopilot you don't notice that.
  • Multiple repetitions are necessary to work miracles. I don't know how the miracle of unconscious adaptation is worked, but since people claim that multiple repetitions are a necessity maybe it's here where they do their mysterious work.
Please pick your choices!

You probably already sense that I think that multiple repetitions (100+) are pretty obsolete. In my reasoning there doesn't emerge a necessity for multiple repetitions. But can that be true? When I played Troyis I got better without much conscious effort. Is there a self organizing capability that adds intelligence to unconscious brain activity? If you hit a ball with tennis which you not can see because it moves too fast the brains seem to perform unconscious goniometric calculations at lightning speed. Yet there is no unconscious discrimination added. If you do a useless exercise, your unconscious brain will adapt to perform a silly task in an intelligent way. Discrimination can only be added by consciousness. I never have thought about how to play Troyis well. I didn't develop a strategy for it. So I learned to move the knight in my mind in a pretty random way at high speed. That high speed was what brought me points. The intelligence of unconscious brain activity seems to be based on simulation and imitation. Yet it is miraculous!

Information is lacking for a final verdict on multiple repetitions. But I have little to show for the multiple repetitions I have done in the past years.


  1. "I never have thought about how to play Troyis well. I didn't develop a strategy for it. So I learned to move the knight in my mind in a pretty random way at high speed. That high speed was what brought me points. "

    Multiple repetition, unconsious or consious has the benefit that it learns you paterns. Due to this paterns you play on autopilot since you know what and how.

    For your Troyis game, you knew after awhile the weak points of the game and where to move with your knight so it became easier for you to pass certain passages/levels.

    With other words, i think repetition can help but the problem is that if one doesn't continue to use it the sharpness of knowing what to do becomes lesser and lesser until you only have a fague memory of what to do in such and such situation.

  2. I think the question about multiple repetitions hinges a lot on how the brain chemistry actually works relative to the intensity of the experience. which we don't really know, except in a very idealized way. but we DO know the practical side of it, which is: the more intense the experience, the less repetition needed.

    we've all seen the unfocused training is next to useless, no matter how many repetitions. 'luna' from CTS is an example I've brought up many times before, as he was a 'regular' with 20K+ tries already when I started, solving at around 1350 with 50% success rate. and now, 3½ years later at 60K tries he hasn't really improved. I know a LOT of similar 1200-players, some of which have played chess for decades already.

    then there are us 'focusers', who consciously work on focus, and get steady but slow improvement. it takes a lot of effort, constant fighting againg the demon of laziness, and even physically hurts. the reason why kasparov called chess 'mental torture'. because it IS hard, and takes a special kind of a bonehead to push through. no amount of chess talent will get you ahead if you can't force yourself through the painful bits.

    and then the other extreme: traumatic experiences, where one single experience marks you for life. no repetition required, and you can't unlearn the stuff even if you wanted to.

    I doubt we can achieve the focusing power of a life threatening experience at will, but it obviously is a factor in some of the learning environments. such as learning to drive a car, where you're actually quite aware of the possibility of crashing the car, and potentially even killing yourself. and for that reason I don't think it's fair to compare the repetition requirements of driving to chess. but I guess the idea is still valid.

    I wonder how electric shocks would improve the training effect? :) a small, safe but slightly painful shock every time you get something wrong. the intensity of conscious focus would probably be massive, and the traumatic memory imprint close to perfect.

    I think it might actually work, but even I'm not crazy enough to try it. -anybody else up for it? :D

    that would separate boys from men in a hurry. maybe even women from men, as they can handle more pain. :)

  3. The greatest enemy is the learning itself. If you learn to do more things automatic you shouldn't be surprised that you do more things automatic. Plateauing is the logical result. That we at a certain moment will try to learn automatic, which is impossible, might be a protection to prevent us from automating everything in life and to become some kind of zombies or robots.

    Just realizing our inclination to learn on autopilot is already enough to wake us up and to get focussed. The extra tension is only painful because we are unused to it. But overtime it should become easier. Allthough the autopilot is always lurking.

    What I'm trying to say is that pain isn't necessarily a good guide in the end. Allthough it is when you start.

    Consciousness alone is not enough. You should add intelligence too. It is perfectly possible to automate silly or unnecessary skills. But why should you? It takes intelligence and discrimination to decide on what you are learning.

  4. yeah, obviously you need to learn the right things. it's just too costly (in time) to reprogram your brain off from bad habits. better to do things right the first time through.