To Fly, You Need to Hit the Ground and Miss

Temposchlucker, whose many contributions to chess learning theory have contributed greatly to my inspiration for the Chunky Rook project, has commented on my previous post and makes an important point that bears repeating:

The danger [of the animated flashcard system] is obvious: it invites to learn on autopilot, which doesn’t work. If we could only find a way to stimulate a more conscious approach and combine it with your charming idea we could make progress. Maybe more text will help since that isn’t procedural and can provide helpful cues.

I couldn’t agree more with the assertion that “learning on autopilot doesn’t work” (although I also disagree with this notion, but I’ll come back to that). In fact, learning on autopilot, i.e. passive learning, is almost a contradiction in terms. Temposchlucker suggests learning needs a  “more conscious approach”, and I think he hits the nail on the head. However, we need to pin down precisely what we mean by conscious learning, and to that end I would like to make a couple of suggestions to prompt a debate.

When you hear “conscious” learning, you may think of a learner who’s alert rather than tired, who does things thoughtfully rather than mechanically, whose mind is present rather than absent. This is certainly a crucial, if very basic, requirement to learning: do not let yourself be distracted — focus, concentrate, switch the TV off.

Having said that, conscious learning requires much more than mere attention. In two recent posts (here and here), Temposchlucker suggests that the chief characteristic of conscious learning is what he calls “feedback”. He defines “feedback” as “the correction of errors with consciousness of sufficient intensity”. While I strongly agree that feedback is key to successful learning, I think our definition of feedback must encompass much more than “correction of errors with consciousness”, and I shall argue this point on the basis of the following premise: learning is about achievement and recall, that is to say, learning means acquiring a skill and being able to remember how to use that skill at a later point. Without recall, acquiring a skill is useless, and recall of something other than skill (e.g. “I totally remember this endgame position from Silman, but I don’t remember how to play it) is also useless. With this in mind, let’s see if we can get a better grip on the concept of feedback.

First: how do we achieve the acquisition of skill? The single most important thing is, perhaps, being capable of articulating what the skill is. That’s why teachers are so useful: not because they are good at something, but because they know what exactly they are good at and are able to tell you. (That’s also why grandmasters can be lousy teachers.) Of course, not all skills are equally easy to articulate, especially at a higher level when skill becomes a matter of experience more than anything else. Nonetheless, for successful learning it is paramount that the learner gets a basic idea of what it is he or she is learning, and a simple label (e.g. “I’m learning tactics”, “I’m learning the endgame”) doesn’t suffice. “Idea” is the key word here: learners have to understand how and why people who have already mastered a skill use it in practice, and they have to have that understanding from the start. Having been introduced to a new skill, learners ought to ask themselves, “Do I get the basic idea?”, and if they feel confident in answering yes, they can proceed. If not, they should either return to the initial explanation or try to figure out the idea from the material that’s presented next. But they should not leave an exercise or chapter or game without getting the idea in the mistaken belief they have learned something.

Second: how do we achieve recall? To answer this question, let’s consider the difference between adult learners and infant or adolescent learners. Temposchlucker posits that adult learners must use feedback to compensate for their lack of the “higher consciousness” that’s found in younger learners.  I’m no expert on child learning, but I think Temposchlucker is on to something here. Part of the reason why children are better learners, I believe, has to do with the fact that their brains aren’t as cluttered yet with conscious worries about survival and social interaction; they have an intuitive grasp of these things, no doubt, but they are not necessarily aware of them on a conscious level. This frees resources for learning. Secondly, children have no “expert structure” in place to sieve new information. On the one hand, this prevents them from learning more complex and advanced material. But on the other hand, they encounter the material that’s within their reach as genuinely new and genuinely fascinating, and this makes the learning experience much more memorable. In addition, they do not impose an expert structure on the material they are learning; their structure is much more malleable (perhaps right down to the actual plasticity of their brains) and responds much more flexibly and receptively to new learning matter. (Of course, this is mostly speculation on my part. I have no actual research to back this up.)

So to return to our question: how can adult learners achieve and improve recall? A simple answer is: repetition. No matter how you learn something, if you do not use it repeatedly or do not recall it for an extended period of time, chances are you will forget. That’s life. So repetition is key to learning, but it isn’t enough. You have to be able to recall a skill when it is needed. In other words, you need to know when a particular skill is useful. This is why knowing your skill is so important: you have to be able to contextualise it while you’re learning it, otherwise you won’t know when to use it once you’re actually playing a game, or worse, you won’t think of the skill at all during your game because it doesn’t occur to you to use it to begin with.

Also, we should ask ourselves: how can we facilitate recall? I think Temposchlucker’s “correction of errors with consciousness” is one way. However, the problem with merely correcting errors with consciousness is that it itself can become mechanical. As soon as you turn conscious feedback into a mechanic procedure, you are likely to lose your consciousness of the skill you’re trying to learn again because the process loses its uniqueness. In my opinion, conscious feedback, in order to be a successful long-term strategy, has to be coupled with something that anchors the idea you’re learning in a concrete, unique and tactile experience. Anecdotes, for example, are great for that purpose. If you do a post-mortem with a friend and he explains an idea to you, it’s much easier to recall that idea later because, in addition to understanding it, you are able to connect the idea to the social encounter, to your friend’s looks, his demeanour, the way he explained it to you, or to the game you’ve just played. This makes your learning experience unique and much more memorable. Of course, we have to be able to acquire new ideas on our own because our friends are not always willing to come over just so we can gaze at them while we learn a new endgame technique. But there are other ways to achieve the same effect. Temposchlucker suggested that “maybe more text will help since that isn’t procedural and can provide helpful cues”. This is spot on, I believe, not only in regard to explanations, which I think was what he had in mind, but with respect to anecdotal cues integrated into the explanations. I was trying to do something to that end in my previous post when I mentioned Worms and called the smothered mate the “bitch-slap of mates”. I am convinced that this innocuous and seemingly irrelevant comment will help me recall the smothered mate discovery in the future. It ties the idea to an anecdote, and makes the learning experience unique. Needless to say, different learner types respond to different cues, but the basic point remains the same: contextualise the skill and the learning experience to facilitate recall at a a later point.

One last comment on Temposchlucker’s verdict that learning on autopilot doesn’t work: I said in the beginning that I couldn’t agree more. But I couldn’t disagree more, either. I think learning on autopilot is extremely useful once you have learned the idea behind a skill. Once you have absorbed an idea, and once you’ve grasped what the goal of a procedure is, going through it on autopilot can become an effective tool to ingrain the skill in your brain. Your learning at that stage may no longer be fully conscious, you may not give real feedback, but your learning is still active in that you know what you’re learning. When you learn a poem by heart, you first have to learn it consciously, remember the words, remember the sentences; but once you’re past that stage, the act of reciting the poem to yourself no longer happens consciously; you’re acting on autopilot. And yet, the recital is extremely useful as a form of practice, until you reach the final stage, the master stage, at which you have made the poem your own and no longer need to practise it: the moment when you no longer try to do, but simply do.

To conclude and sum up, learning has to be active. This means you need to grasp the idea of the skill you’re learning and what it is you’ll achieve once you’ve mastered the skill. You don’t want to lose a skill you’ve learned, so you need to be able to recall it. This happens through repetition and by making the learning experience unique. Once you’ve achieved this and are able to explain and recall the skill you’ve acquired, learning on autopilot can help you make the skill a part of your unconscious skill set.


12 Responses to To Fly, You Need to Hit the Ground and Miss

  1. Lawrence says:

    You have some good points here. If I may give a suggestion, perhaps it would be a good idea to slightly modify the diagrams so that the viewer can input the moves after they’ve been played a couple of times. I’m not sure whether this improves significantly the “consciousness” of the learning process and it might be technically difficult to implement, but it should make the viewer more alert to the situation.
    My 0.02$.

  2. chunkyrook says:

    Your two cents are most welcome, Lawrence, and your suggestion makes sense, but as you rightly point out, implementation is difficult. I, for one, have no clue how to go about it; my “expertise” is limited to gifs, I’m afraid. Also, I share your doubts about the extent to which such a feedback would improve consciousness, and whether that’s necessary even. As I said, the gifs are mainly mnemonic devices used after you’ve consciously pondered and digested the idea behind the moves, so there’s no need to add a conscious element other than watching the gif thoughtfully (that, at least, is my hypothesis). Another possibility, perhaps, is to link to games or puzzles featuring the ideas in question; this would help contextualising them further. Thanks for the input, much appreciated.

  3. najade says:

    Just some more thoughts.

    What is the function of the animation?

    My goal is to recognize all kinds of characteristics of a position. If there are 10 characteristics of a position, I recognize only 7 of them, usually. I need a teacher (in the form of a book, in my case) to tell me the other 3 characteristics.

    But some characteristics aren’t in the position yet since they appear only in the future, after a few moves. That is where the animation can be helpful, to show future positions.

    Yes, repetition is important. But I just haven’t found the past years any proof that unconscious repetition bears any fruit.


  4. chunkyrook says:

    The function of the animation is simple: a lot of chess ideas concern move sequences rather than static positions, and the animation (unlike the static position) can illustrate this. So the purpose of the animation is two-fold: it (a) illustrates the idea and (b) helps you remember the idea as a move sequence. As I mentioned, of course, (b) is hypothetical.
    I assume when you speak of the “unconscious repetition” you did in the past few years, you’re referring to solving puzzles. I myself am quite sceptical of training by puzzles (I don’t deny that they’re useful, but I don’t think they’re primarily useful for learning pattern recognition). The difference between puzzles and animations is that animations are bare-bone idea skeletons; there’s no need to tease out the idea again as is the case with a puzzle, because the idea’s already in plain sight. And I believe that’s an advantage the animations have over the puzzle. Puzzles have other advantages, of course; they contextualise ideas, they are great at teaching you calculation and visualisation. But as far as illustrating ideas are concerned, I think the animations are better. Not the least, too, because of the loop: you don’t have to do any clicking, you are not distracted by other activities; you can simply watch the move sequence unfold and concentrate on the idea at hand.

  5. chunkyrook says:

    “My goal is to recognize all kinds of characteristics of a position.”

    This has given me some thought for food. I am wondering if the goal, as it stands, isn’t too abstract, at least for my level. At this stage, I need ideas that can be expressed in a couple of moves. What I hope to achieve is break down positional analysis into key ideas for move sequences that, when taken together, allow me to correctly assess the potential and character of a position. I shall post an example shortly. Thanks for your thoughts, Tempo.

  6. chesstiger says:

    This sound as if one needs to get positions at which is not said what to do and lots of them.

    Each puzzlebook has section of mate in two, gaining pieces by double attack, … . So why not make a book with puzzles where it just says white to move, or black to move? Even sneak in a few positional puzzles to make the person who tries to solve them mad if he cannot find a tactical solution. :-)

  7. chunkyrook says:

    That’s essentially what you get at ChessTempo. In my opinion, both kinds of puzzles are helpful for various purposes, but the reason I like the gifs is that they’re more efficient in illustrating the idea. Much of the conscious feedback involving puzzle-solving deals with calculation and visualisation, and that’s not needed to understand an idea. And since the position you’re going to encounter on the board will most likely differ from the particular position you’re looking at in the puzzle, solving puzzles alone might actually prevent you from spotting the idea in the game later. The gif is all about the idea, not about calculation or visualisation. It shuts all of that out, allowing you to focus on the idea in its “pure” form. Once you got the idea from the gif, puzzles can help you cement the idea by presenting it in a “realistic” position with all sorts of pieces. But having to solve puzzles in order to get at an idea seems inefficient, and only looking at an idea once in a book and then go straight to puzzles seems to put too little emphasis on learning the idea as such in my opinion. That’s why I think the gif makes sense as an intermediary step.

  8. I don’t have anything to add of an epistemological nature, but I did find this post most interesting, and helpful.

    If you’d like to follow your animations with a solvable problem Palview will do that — that’s the software used to host your game. Here’s 2 demo pages:

  9. Also, Aquarium has some publishing features. Thei-book is pretty promising and it’ll do some kind of export to html from i-book. Don’t know how robust it is at this point, but worth investigating. Once again I’ll host if it will help.

  10. Heisman has a recent article on related topics here called ‘The improvement feedback loop’ (that link will probably stop working around April 17th). Based on that article, he might suggest, after looking at the gif, put up a series of five or so questions that use similar themes, to test you and make sure you understand it.

    My guess is the act of making the gifs is probably very helpful.

    It is up to the viewer of the animations to decide how passive to be. He can just stare at them and not think, or he can look them over, explain them in N different ways, think about how changing the original position would influence things.

    Interesting stuff. I modified my approach to learning tactical puzzles to aid my recall, as I described here. I was sick of remembering the problem but not its solution. My method helped, and was fairly exhausting (a good sign).

  11. chunkyrook says:

    I absolutely agree with the notion that the solution is more important than the problem, and that tactics puzzles can lead to all sorts of distracting feedback, or even wrong feedback, as you say in your post.

    As far as the learning loop is concerned: I think the most important stage is indeed the articulation/identification of what’s being learned (as I suggested previously many times), and this statement by Heisman hits the bull’s eye:

    “The key to any improvement loop is the awareness and identification of the problem. Many players think they can identify their own problems, but it is difficult to be objective and easy to rationalize that the causes are under your control. That is why almost every good player at some point spent considerable time around other good players who pulled them up to a high level.”

    This is exactly what the “gif”-improvement plan is trying to achieve, although not for problems, but for ideas. That is to say, not for retrospective feedback, but prospective plans.

    The articulation of a chess idea happens in words (a human player explains the gist of an idea) and in terms of moves (you learn to associate certain positions and piece setups with an idea). The Chunky Rook posts are one attempt to combine the two efficiently. Adding puzzles or positions that put the knowledge acquired through the gifs to the test are certainly a great idea, but at this stage beyond the blog’s capacities.

  12. Munich says:

    Hello chunky rook, awesome blog. I intend to read it all.
    This post is now 5 years old.

    I want to show that learning on autopilot is possible.
    Take the Fritz board vision exercise “defense training”, where you click on all undefended pieces.
    First I managed maybe 10 undefended pieces per minute.
    Soon I was at 20 undefended pieces per minute.
    After some days I was at 30 undefended pieces per minute, and my record stands now around 60 undefended pieces per minute.

    I think that if you click on pieces with 30 pieces per minute, then you are very much on autopilot.
    If learning on autopilot was not possible – how is it possible that I could still improve?

    In your post you are spot on: repetition helps to ingrain your knowledge.
    I did not learn anything new, but I “learned” to recall it faster.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: