Saturday, November 26, 2011


I organized my list with questions. The result is: tactical scenario's.
For instance for the position of the previous post:
  • Hunt the King.
  • Prevent him to flee.
  • Win a piece that he has to sacrifice to prevent mate.
  • Add pieces with tempo to the equation.
Mr. Z gave me an article with citations of 49 grandmasters who gave advice on how to improve in chess. I counted their advice for you.
  • Study annotated games 19
  • Study openings 11
  • Don't study openings 6
  • Study endgames 10
  • Don't study endgames, it's a hoax 1
  • Work hard 8
  • Play games 8
  • Play stronger players 7
  • Mental preparation 6
  • Tactics 6
  • Study theory books 6
  • Analyse your own games 5
  • It's hopeless if you are not good by now 4
  • Get a coach 3
  • Be patient during play 3
  • Continuously beware 2
  • Drink beer 2
  • Visualization 2
  • Base your plans on pawns 1
  • Physical preparation 1
  • Don't underestimate your opponent 1
  • Love the game 1
  • No sex during tournaments 1
The grandmasters were totally divided about the importance of memory.


  1. Me Munich: I like these 2 most:
    - It's hopeless if you are not good by now
    - Drink beer

    I guess most strong players dont really know what helped them most?

    Reminds me of the question of my son Joey who speaks 4 languages fluently: "Mummy, how comes that I actually know polish?"

    Well, he doesnt know. But since my wife talks to him mostly in polish and I talk mostly to him in german, and we live during winter in the Caribbean (speaking spanish) and in summer in England - it is natural he speaks them all. But how did it exactly happen that he acquired these languages? Good question.
    Part of the truth is, that he memorizes whole sentences. He is simply copying sentences. I studied spanish, too, but I am missing the ability to hear a sentence once and then repeat the whole sentence immediatly after I heard it. Rather I need to ask for each word I did not understand, and probably need to ask again for the word I did not know. It makes me believe, that children can grasp complex information much better then we do. They have especially a more detailed STM? Then again, aox says that children can only memorize 5 items at once, while adults can memorize about 7 items ("magic 7") at once. So STM needs to be split in two different kinds of STM?

    Fortunately, a child is using its LTM. It stores the whole sentence. And I remember, that at the brink of his 4th year it took my son ages to memorize the desirable word "helado" which means "ice-cream". So there are cleary differences in memory, and somehow children have on one hand a very poor memory, and on the other hand a very good memory if it comes down in grasping complexity.
    Without knowing about it, like these 49 GMs dont know about it.
    I believe, that if we want to have a chance in improving, one key aspect is to break down complexity of the chess game.
    How? Not so easy to say.
    Your new method of writing down questions might be such a step into the direction to break down complexity.
    Then again, with language, at a certain age it is very unlikely for us to become accent free. We can try by having headsets on, and repeat again and again phrases. Get a native coach who is correcting us.
    I am sure we can learn a language decently, but getting rid of the last bit of accent - this is difficult to probably impossible?
    What could that mean for chess? We can learn it to some degree, but there might be a limit? If we cant become GMs, then maybe hopefully we all can become IMs?

    But asking my son in 20 years to come how he did acquire languages is going to be as useless as asking 49 GM's. If they didnt go deep into the matter like we do, they can just guess around, delivering a list like you presented here. But most probably have not a clue how they did it. And a lot of them dont improve once they matured, because most of them dont have a good training method, and hence most cant improve after they became 30 years old.

  2. An international master is a grandmaster with an accent. I like that:)

    Studying annoted games and building tactical and positional scenario's might actually be the same. What do you remember best? What makes sense and what makes impression. I'm especiaaly interested in scenario's my opponent dosn't know.

    Yesterday at the club I looked at a drawn endgame. The only thing white had to do was playing his bishop between e1 and d2. Black tried to find an entrance point for his king which he knew wasn't there. So he played his knight around. On a certain moment white lost his patience and wanted to DO something. So he played h3. Black and I IMMEDIATELY knew he would loose a pawn about 10 moves later and another 20 moves later the game. We recognized the scenario while white didn't.

    Now that is a scenario worth remembering!

    In the mean time we simulated that we looked 30 moves in the future. That is how to visualize without overloading STM.

  3. Me Munich: For the hard puzzle where you formulated the concept, I have just come across this easy puzzle. It is a puzzle I did slow, so it was on my "repeat slow solved puzzles".
    But when I saw it, I could not help but remembered your very hard puzzle.
    Honestly: Do you see the similarity in the pattern? The main actors white queen and white bishop are there, the environment black King, black queen and some other pieces around the black king are there to.
    The main difference is, that my easy puzzle is much simpler (CT Blitz rated at 1225 versus your hard puzzle which is CT Blitz rated at 2708).
    But if it had been better known by me (beeing able to solve it in very view seconds), I might have had a chance when I tried your very hard puzzle?

  4. @Munich,

    From the scenario, 2 of the 4 elements are present in the easy problem. Probably it doesn't matter if you derive the elements from 2 simple or 1 dificult problem. As long as you obtain them. When you have all elements the difficult puzzle stops being difficult.

    That's why the title is "its not rocket science".

  5. What you see is that pattern recognition is not the problem. Even in a totally different position we cannot help to recognize the theme. That is simply how our mind works. Obtaining the themes, that's the problem.

    Once obtained, visualization is no problem either. The theme is your chunk, so to speak.

  6. Me Munich: I am still currently doing my set of "puzzles I did slow (13 seconds or more needed". Amongst them this easy puzzle (CT Blitz rated at 1280):

    2+2 = 4?
    I think the hard puzzle suddenly becomes easy if we were able to recall these two easy puzzles instantly.
    But I was not able. Both puzzles are in my set of "not well known patterns/themes/concepts". For both I previously needed 13 seconds or more. (I dont know how much more. but often - if I cant find them quick, we are most often looking at a solving time of 20-30 seconds.)
    So here we have a good excuse why I did not solve the hard puzzle: Important themes I could not recall in my STM prunning/calculation. I found this puzzle-theme/pattern/concept in the hard puzzle, but failed with my forth move, only considering 4...Rf8+.

    Both puzzles not know so well by me resulted in failing the hard puzzle.
    This pattern in the seconds puzzle was also responsible, why I was not able to find the solution. I needed to much time, and if I cant grasp one variation in the tree of variations within 20 seconds, chances are high I wont find the solution. Rather I skip this variation of the tree and try something else. I believe it is vital not just to be able to find the solution of the single patterns, but to have them ready/available in very short time. because deep in the tree of variations, I am not thinking many seconds about a move.
    All I thought was 4...Rf8+, and the bishop move was not candidate move by the time I spend some seconds about this part of the variation in the tree of variations.

  7. @Munich,
    You gave me a link to the difficult problem. What is the easy one you wanted to show me?

    In the scenario there is a hierarchy according to which the moves should be investigated. In itself it is very good that the move Rf8+ comes to mind. If everything else fails, add a new piece to the equation with tempo. But the magnet-move Bg3+ is higher on the ladder.

    The familiarity of the theme is often in conflict with the hierarchy of the moves you should consider. What we are doing is trying to hammer a round pole in a square hole.

    Familiarity has a lot to do with the frequency of a theme. Adding new pieces to the equation with tempo is much more common than the winning of a piece by a magnet-move.

    There is an evolution-battle going on in our minds. The fittest moves climb up the evolutionary-ladder. 'Fit' here being related to 'most frequent'

    Our automatic brain is only semi intelligent. It emulates intelligence. If you automate stupid moves, you can find stupid moves very fast.

    What I'm trying to accomplish is to find a standaard approach to a tactical problem. A checklist with a logical hierarchical structure to fill in. It should lead to right move.

    Right now I'm not worried about speed. Yet. I'm looking for a guidance to my thoughts.

    Once found, I'm going to look after automation and speed. But this is the phase to add intelligence to the guidance. Which is slow by definition.

  8. I looked at your proposal that 2+2=4 while solving problems. It is true, but in another way than we think.

    We are used to look at pins, double attacks, skewers and so on. In those cases 2+2=4 indeed. But the reason why we miss the difficult problems is the lack of common sense. The inability to formulate a logical reasoning.

    We are more than familiar enough with the classical tactical themes, but we omit the basic themes for logical reasoning. We are not familiar enough with them, no matter how logical they are.

    We have to change focus from tactical themes to the basic elements of logical reasoning untill they become familiar too.

  9. Sorry, here the correct link to the easy 2nd puzzle.
    We have no trouble in finding the solution, but it is not only about finding them. We need to find them quick in the tree of variation. Because within the tree you are not looking long at a future point. Here at move 4 I doubt I spend longer than a few seconds. All together I spend maybe 2 or 3 minutes, and in this time I covered many variations.

    So yes, we were not familiar enough with these 2 simple tactics. Actually it is enough not to be familiar with one of them to fail the problem.
    With my training I make sure next time I am more familiar with them.
    But time for logical reasoning?
    With an abiltiy to instantly notice the patterns here - I think I would not do much reasoning, but just think "ah, this is like one of the easy puzzles I learnt form CT" (This thought I have often in my real games lately - I swear to everything that is holy for me, this I really do. I really do notice them in my games and the other way round, too. When I see a puzzle, then I recall the game. Or the very hard puzzle temposchlucker gave. It works in both directions.).


    It is actually somehow simple - If you know it! (1280 CT Blitz rated only, so a bit higher then the first I have given here).

  11. Me Munich:

    An idea:
    There are two sorts of difficult puzzles:
    First, the ones with lots of calculation, and deep variations.
    Second, the ones which could have been easy if you simply had spoted it.
    The 2708 CT Blitz rated (your example) is a puzzle of the second category: the relativelz easy ones that for some strange reason almost nobody gets correct.

    I would love to have some puzzles sorted into these two categories.
    I could imagine it is worth to study the character of the "easy" puzzles, that are very difficult.

  12. So overall, GM's are OK with alcohol, but not sex as far as chess improvement.

    I was rather hoping for the opposite.

  13. Hey, is that familiar to you? Again it is a very easy puzzle, rated just 1282 in CT Blitz rating. Isnt it striking similar?

    Here I copy the one I previously posted above, so we have a collection: