Wednesday, September 21, 2005

Rabbits in the clouds.

In a comment on the last post, Mouse warned against the phenomenon of memorization.
In the past we had a lot of discussions about this subject. For example here and here.

But now I am working with CTS, I ask myself: Is it possible to memorize a problem without recognizing the pattern? And further: is a problem really stored in two ways in the brain and in two places? First as a memorized problem and second as a stored pattern?
That doesn't look very likely. Because the brain is very efficient.

Mouse supposed: "the issue is recognizing patterns in unknown positions".
I have seen no convincing prove yet that it is actual possible to recognize a pattern in an unknown position.
In my post no presents given I tried to tell about this.

Take for instance the two following diagrams:

Black to move
This one costed me 10 seconds

White to move
This one I could not find in 3 minutes

Both diagrams contain the same theme.
A queen sacrifice followed by a knightfork.
The first diagram costed me 10 seconds.
For the second diagram I had to look at the solution after 3 minutes thinking and finding no clue.
I encountered far better examples than these, but this will give you the main idea.
Why does my brain think that these two patterns have nothing to do with each other?

Of course we can recognize a rabbit in a cloud. Even if we have never seen that cloud before. But here the patterns seem to have to be far much closer to each other, before the brains recognize it as the same pattern.


  1. Sometimes I missed these N forks because I see how my N can put the King in check on one square but miss the alternate checking square. When I am playing a game I look for in a N Move can threaten a king or queen then check if I can force the other piece into the correct Square. I look for a piece of the pattern than see I can fulfill the second part.

    One tactic I repeatedly missed in Ct-art was with the K&Q next to each other on the back row. The position called for Moving a rook next to the queen forcing the rooks capture. The King and Queen are now seperated by one square and now fall victim to the Nfork.

  2. I think, if you memorize on purpose of course you won't get it or improve much... But the analytical ability you develop while doing tactics (i.e. This queen is overloaded, if i take this, then it'll leave me that..) will improve.

    Both took me about 5-10 seconds, the second taking slightly longer. Maybe in the first you easily saw the idea because you thought it was more forced.

  3. I don't think you can fully memorise a position - you never remember the position of all the pieces. Without looking back at the first example, could you tell me if there was a black pawn on h5? To a certain extent you are remembering a pattern. Sometimes a very specific one, but a pattern none the less. The more you learn, the more your brain is geared up to recognising more general ones. After months of CT ART I am finally beginning to recognise most of these patterns now though. Annoyingly there are quite a few different patterns still to go ...

  4. Druss, that is a very good point!!
    So there are probably a few dozens of positions that are close enough to diagram 1 that the mind recognize it as being the same pattern. For instance with or without a bishop on c2. But diagram 2 is too far away from diagram 1 to be recognized as the same pattern.

  5. Your examples are exactly what I mean. There may be hundreds of positions with this one pattern: Q sac followed by N fork. BTW I see a camouflage that may have disturbed you in the second pattern: the B is protected twice, and this may have provoked a silent inner voice to tell you: not to be taken because twice protected, once attacked. Just a possible explanation.

    I often get a problem that I remember together with clues of the solution, e.g. aha, I had this one before, and chasing the King was successful then ...

    But the easy points I get then are not because of tactical skills, but only because of memory.

  6. I solved those pretty easily. Now, show me some problems with fianchettoed bishops and watch me stumble!!!

  7. I solved both in about 10-15 seconds. Now regarding the patterns - I didnt pay any attention to any of the pieces. All I saw was the knight, and the forking square. Just that simple idea. I saw that in about 3 seconds. The other 7-12 seconds was for a quick checking of other characteristics of the position.. i.e. doubly protected bishop.
    So this means that I 1.) See the basic stuff first and 2.) Check if the basic stuff works..

    How do you see the pattern tempo? Are you trying to see the pattern of the *whole* board in a glance?

  8. Thx for your question Nezha,
    I look at the board and my eye is catched by something. That is so strong that it is impossible to look further. A knightfork is simply never the first thing what I see. Only if the combination that I see fails, I look further for other eye-catching patterns. I don't like to admit it, but I don't control this process.

    It is so important that what you see at the first glance is indeed the key pattern on the board.
    I assume that this process can only be controlled in an indirect way. For instance by doing tons and tons of knightforks.

    That is why the time constraints of CTS are so good. It reveals what you see in the first 3 seconds of the process. And that proves to be alarming inadequate.
    Sometimes I haven't even figured out which color is mine. . .

    This 3 seconds have a special meaning in the work of prof. Adriaan de Groot. In these first seconds you make use of a part of the brain that is special designed for processing visual data. It works unconscious, and it plays an important role for grandmasters who look at the board.

  9. Training to see forks became really easy when I realized that mostly knights *cant* do any other tactic. I mean forks are it. There are no possibilities for pins or skewers or backrank mates,etc.. just the fork. So when I see a knight, I think of a fork and nothing else. More, I assume that there *must* be a fork somewhere.

    Same with bishops.. Bishops can fork too right.. but when I think of bishops, I think of pins and skewers first.. thats why I usually miss forks by bishops but rarely miss skewers..

    Well, thats my thinking process, or at least I think it is.

  10. There is a lot in this 'first 3 seconds' thing. When solving the easiest, level 10 problems in CT Art, I tend to either see the answer in the first 3 seconds, or I struggle a lot.

    Also, it is very unconscious when I do. The answer just springs out.

    When it does not spring out I end up going through a list of things to check: unprotected pieces? deflection into knight forks? all out sac against the king? overloaded pieces? But manually going through these just isn't the same. It is best when the answer just springs out at you. If only it would happen all the time though!