Monday, March 28, 2016

Two methods of learning

The past few months have been very clarifying. I was in doubt between two different methods of improvement, with each their own logical arguments. Which one to choose?

The first method is based on slow conscious learning, while the second has become known as "the salt mines". Which is based on massive repetitions at high speed, a form of unconscious learning. Conscious learning is intelligent, while unconscious learning aims at speed and precision.

The lesson learned, is that the two methods aren't mutually exclusive. I don't have to choose. Both methods of learning have their own application and complement each other. In order to understand how the two methods interact, we should extend the "learn how to drive analogy" (hat tip to Aox).

The learn how to drive analogy.
Conscious learning will learn a person how to drive a car, and to become an experienced average driver. Your performance will be limited by the natural speed and precision of your unconscious processes. If you want to become a racing driver, you need more. You need to train subtasks like shifting gears in order to do it faster and with more precision. On the other hand, you need strategies to improve the main task too, which is done by additional conscious training. Both methods of learning are needed to optimize the task as a whole.

In general we can state: conscious learning is needed when we need to add intelligence, unconscious learning is needed when we lack speed and precision.

The experienced average driver.
The big question is: what does an experienced average driver mean in chess? For me, that is a master level player. If I look at the problems at CT where I fail by making an error or using too much time, it is for 98% caused by a failure to apply plain common sense to the problem, and for 2% it is caused by a lack of visualization skills. I have solved a lot of 2000 - 2400 rated problems in the past, and that was about the same. May be a 95 : 5 % ratio or 90 : 10 %. But almost every time I solved such high rated problem, I slapped my head, after seeing that the solution was very simple in essence. Meaning, that I need to apply more intelligence to the problem, not speed or precision. Meaning, conscious training must be applied. Above 2400, the problems became more esoteric, and I was hampered by a lack of imagination or visualization skills. For which salt mines could be necessary.

What you need, is a realistic vision concerning your failures.

How I feel when solving problems

12 comments:

  1. No doubt: thinking process is important and its possible to improve here.
    But it is interesting that we can measure the tactical skills with a m1-e. In my eyes the limiting factor for chessimprovement is the patternrecognition/vision. Good player can play good without (much) thinking. They play bullet better than we can play OTB.
    The positive effect of a strict thinking process is, that it is a saltmine by itself. You repeat the same things over and over again. But i am afraid that these saltmines are not efficient enough to improve an adult in fast processes.

    ReplyDelete
    Replies
    1. If you analyse your failures and time consumers at CT and fail because you don't apply chess intelligence or it takes too long to apply chess intelligence, than you need conscious training. Otherwise, you banned yourself to the salt mines.

      You must realize that a narrative (like a thought process) enhances visualization. Both in speed and quality. See http://temposchlucker.blogspot.nl/2011/09/mentalization.html

      Delete
    2. I have the same impression. I have been playing chess for over 15 years. Up to now I have played about 120K chess games (95% of these - bullets and blitzes). Most often I do not know what to play due to the lack of specific KNOWLEDGE and understanding the core of the position. While I have been playing bullets - the matter of pattern recognition comes to the top. However when I am playing 30 minutes game (G30 or slower) it looks like my lack of knowledge and understanding is a key factor of failures.

      What I can notice more? I am a really fast M1e and FAC solver. In my case - it is probably not possible to improve more than 10% from the high scores I have achieved so far. However my lack of knowledge is really visible when I am playing against stronger players. They simply beat me not because they can "play better", but due to the better understanding and possesing more important concepts (pawn breaks, plans, mating atttacks, creating weaknesses, etc.).

      Recently (actually it was a few days ago) I played bullets against a player rated 1800-1900 (at bullets) and I have had problems to beat him. It was clearly visible I play weaker moves and I lose a lot of time to understand the core of the positions.

      I think if I start learning more concepts plus replay and analyse the master games - I can improve my chess much more. Continuing salt mines leads to nowhere in my case. Of course there are some drawbacks with the visualization and noticing which pieces are defended and under attack, but this flaw is very low important against the lack of concepts and specific knowledge.

      What do you think about it? Would it be any importance if I learn chess with extending my knowledge - and after that I could improve? Would it refute or validate any of the hypothesis of ours?

      Delete
    3. Your sucess in chess is based on KSA = Knowledge, Skill and Ability
      You need them all. While improvement in knowledge and abilitys is still easily possible at higher ages, the skills are not that easy to improve with age.
      But if you learn new concepts you still need to put them at the board. With bulletplay you can only manage to bring concepts at the board which you had already deeply learned ( in slower games ).
      I would guess that if you learn something you will need to deepen it at plenty of slow games before it can show effects in bulletgames.
      The problem with knowledge is to get this knowledge on the board. I experienced that i could tell some low rated player over and over again that they made erros in the opening according to the opening rules which they did "know" perfectly but at the board that was all gone and they made the same errors again and again and again and again and again...
      It is necessary to implement this knowledge into a thinkingprocess. Munich did use such new knowledge as a Mantra, he told himself during a gaim over and over again: "to take is a mistake ", "to take is a mistake ", "to take is a mistake " and did overcome this way the common error of taking first.

      Delete
  2. @Tomasz, not too fast. Although my ideas are applicable to "higher" chess knowledge as well, that is not what I am talking about here. I'm talking about the chess logic within tactics here. The chess logic that combines the familiar chess motifs into combinations.

    ReplyDelete
  3. Interesting site here. Do you happen to know whether the "salt mine" speed ratings are correlated with general tactics ability? Or are say you and Tomasz on the same level? It seems you like hypotheses...here are two contrary ones. Experience with lots of chess exposes everyone to the same sorts of problems in the same proportions, so different kind of chess abilities advance hand in hand.

    Alternatively, maybe the tactical "rating" is a crude instrument that hides all sorts of interesting variations in how different brains tackle chess. For example I have the impression my visualization / memory /calculation abilities are really weak - I can't reproduce even the position I'm playing at the moment on a separate board. I do best in problems that have very little branching on the tree of plausible moves. Whereas some people may be able to hold a position five moves deep in their heads but have a lot of trouble figuring out what are good moves to look at, so they have to calculate a lot.

    By the way I've spent some time on Chess Tempo. So I know from the CT ratings Temposchlucker mentions I am way worse than you guys. More room to improve I guess.
    mfardal

    ReplyDelete
    Replies
    1. I don't know the tactical ability of Tomasz (nor his nickname at CT, for that matter).

      You are right. Behind the rating is hidden how people get that rating. Everybody makes quite different mistakes. That is highly personal.

      Delete
    2. Anonymus asked: "Do you happen to know whether the "salt mine" speed ratings are correlated with general tactics ability? "

      There are some scientific papers about that:

      "experts are superior to novices in picking up information from a board position. They clearly perceive faster all kinds of chess-specific perceptual cues. If chess players’ are, for example, asked to detect as fast as possible, whether one of the kings is checked or not, masters are clearly superior in speed as well as in accuracy (Saariluoma, 1984, 1985). The same superiority can be also be found when chess players assess if a mate in one possible (Saariluoma, 1984). ...." ( http://www.uv.es/revispsi/articulos1.01/SAAmo.pdf )

      Delete
  4. Hi Aox,
    Thanks for the link. It's kind of interesting. However, I have zero doubt that there's a strong correlation between ability and these perceptual/visualization/memory exercises. The questions for me are (1) how much dispersion there is about the correlation. (2) If there are many dimensions that dominate this dispersion, or if it can be explained by a small number of extra variables (corresponding to good/bad board memory, fast/slow calculation speed, etc). There aren't any plots or stats in the paper that address this (I didn't look up the references, however).

    I've seen your palindromic handle in the forum at CT, are you a user on there too? Care to provide your rating? (Don't worry, I'm not trying to compare myself to you, just wonder how it compares to salt mining speed.)
    mfardal

    ReplyDelete
    Replies
    1. The references of the paper provide some ( small ) statistic. I know of no statistic analysing the effects of the different (sub)skills of tactics on the general tactical skill. My personal breakthrough in understanding of the problem of improve of adults in tactics is : that it is hard to improve in m1-e ( but not that hard to improve in some subtasks called board vision. )
      Board vision and tactical vision are not substantially different. Many board vision exercises are just adressing the tactic: "hanging piece" for example

      If we cant improve in something that simple as mate in 1 to the level of masters, then how can we improve in tactics as a whole to master level?

      At empirical rabbits blog we did diskuss the effects of speed/time on the quality of a move: as longer we think as better the move. Its no problem for any class tactician to perform in tactics like an IM if they simply think long enough. I once did push my standard rating up to 2300 in standard by just.. thinking for hours at one problem. A friend of mine, rated 1600+ did push his standard rating above 2400 by thinking many hours or even days about a single problem. Of course its nesessary to have a good visualisation and memory ... to benefit of such long sessions. A master can do it quicker, they have superior tactical vision. But such visions are virtually not improvable as an adult ( it seems ).

      If a tatctician does think twice as long at a set of problems their performance will be K elopoints higher where K is an individual number between 100 and 200. ( read more about it at empirical rabbits blog )
      So if we could improve at simple tactics we would automatically improve in all types of tactics . Implementation of a thinking process, better visualisation and so on would estimatingly improve K.



      You can find statistics here http://chesstempo.com/chess-statistics/aoxomoxoa
      http://www.chessgym.net/usr_ranking_att.php
      http://www.chessgym.net/usr_ranking_def.php

      Delete
  5. @mfardal

    You can compare the ratings, places and the level of commitement (or sacrifice if you will).

    You can find statistics:
    A) ChessGYM

    http://www.chessgym.net/usr_profile.php?user=thinkerteacher
    http://www.chessgym.net/usr_ranking_att.php
    http://www.chessgym.net/usr_ranking_def.php

    B) ChessTEMPO
    http://chesstempo.com/chess-statistics/demon_szybkosci

    One of the argument we cannot refute is the simplest one: in PRACTICE (a chess game played OTB) we cannot think forever. The time limit is the factor NOBODY can avoid or omit. we can argue how much time is "much" or "little".

    To compare Aox's statemet: I am able to solve most high school maths tasks (puzzles), but the amount or time and energy is HUGE (very big level of time consumption). In contrast - I can solve 95% tasks at the primary school level with just a fraction of time (very fast).

    From my point of view - the time factor is one of the key ones, but it should not be overestimated. Why? It is easy to explain what I mean: If two people find the same (best) move in 2 minutes, it does not mean they are 'equal chess solvers'. Unless we measure their energy consumption - we will draw wrong conclusions. One of the player could use just 5% of eneregy whereas the other one - 85%. The difference is HUGE. That's my point of view.

    ReplyDelete
  6. Thanks Aoxomoxoa. I didn't know you could just reference the user profile, and for some reason you didn't pop up in the alphabetized list of users. As expected, I am not as strong as an Aox.

    ReplyDelete