Sunday, May 29, 2016

The remedy

The diagnosis of our disease looks precise, and I'm happy with it. The past 63 days I've been on a streak at Chess Tempo, and I have made no progress in the area of the remedy at all. It is time to have a fresh look at it. What do we know?

Prof. Dr. Temposchlucker
From a documentary I learned that Susan Polgar has hijacked her Fusiform Face Area (FFA), a brain area that is dedicated to face recognition,  in order to use it for the recognition of chess positions.

There are some scientific papers that point in the same direction (the FFA). Yet I'm reluctant to make much use of these papers. I'm no scientist, and if I'm going to try to emulate one, I might very well limit myself due to misinterpreting scientific data. In stead I will use my regular reasoning, which is good enough, and if it leads to wrong conclusions, we will falsify them along the way.


Pattern recognition
There are about 30 tactical themes that are commonly used when talking about tactics. These themes appear in an infinite amount of different shapes on the chessboard. Pattern recognition is the ability to recognize those shapes, and map them to one of the 30 themes. How did we learn that?

Most of us have learned the tactical themes from a book, and by doing a few hundred exercises with related positions.

When we look at the problems at CT, it is easy to see that there are much more tactical themes. Themes of maintaining the initiative, themes that are subthemes from the 30 commonly known themes, themes that connect two different themes, themes around captures etcetera. It is impossible to say how many unknown themes there are. For the time being I say 70, since that gives a nice total of 100. I intend to explore the CT database for the most frequent occurring new themes. After that, only a few hundred exercises with the themes will do.

Example

Below you see an example of a subtheme of the removal of the guard theme. Often it is not sufficient to exorcize a defender from a square, but you must deny him a whole rank or diagonal.
Harassing the defender. White to move

3r3k/p1r3pp/1p1pnb2/3Qp3/1PP1B1qP/P3B1P1/2R2P2/2R3K1 w - - 1 1
 solution

5 comments:

  1. When people are just learning chess, I always recommend they go through a book that covers the most common tactical themes before doing a chess server like Chess Tempo (which I love). I think it's good to have a mental template of these themes. I think besides tactics, you can do this for strategic elements - especially once you tie them into opening structures. Similarly, certain tactics flow from different structures - although a particular tactic can flow from any structure due to blunders - certain structures lend themselves to certain tactical themes.

    As an example in fianchetto structures, you will see more often piece sacrifice tactics to break up the structure, perhaps combined with flank pawn advances. Of course, chess is being able to not only recognize, but combine and utilize these patterns to create strong moves.

    Great post, looking forward to more Tempo.

    ReplyDelete
  2. Curious:

    I "saw" the CT line 1. Bf3 Qf5 2. g4 winning the Black Knight after 2. ... Qg6 3. Qxe6 almost immediately, and then - rejected it as "too simplistic!!" So, I began looking at 1. Bf3 Qh3 instead. The "amusing" (sad, really) thing is that I did NOT "see" the correct followup of 2. Qd1!, threatening to trap the Black Queen. If 2. ... Qf5 (forced) then 3. Bg4 skewering the Black Queen to the Black Knight SHOULD win after 3. ... Qg6 4. Bxe6. Downright side-splitting "hilarious" is that 1. ... Qh3 was played in the source game (according to a comment) AND Fritz Stockfish DD64 rates 1. ... Qh3 slightly higher [+-(+5.03)] than 1. ... Qf5 [+-(+5.17)]. Not that the difference of (0.14) should matter very much when one is "winning" by the equivalent of a Rook. . . but my fantastic ability to overlook the "obvious" merely drives home the point that I have NOT learned my lessons yet. Sigh. . .

    Rhetorical question: Anyone care for a nice game of tic-tac-toe or tiddlywinks?!?

    ReplyDelete
    Replies
    1. http://chessconfessions.blogspot.nl/2007/04/letter-to-knights-from-michael-de-la.html

      Delete
    2. It would appear that playing Tic-tac-toe is much more lucrative than chess, at least for MDLM! Thanks for the laugh!

      Delete
  3. An interesting Problem:
    The ratings are:

    Problem Blitz Rating:1572.4
    Blitz Av Seconds:00:54
    Blitz Attempts:1652
    Blitz Success Rate:68.16%
    ________________________________________

    Problem Standard Rating:1480.6
    Standard Av Seconds:02:58
    Standard Attempts:10132
    Standard Success Rate:59.25%
    ________________________________________

    Problem Mixed Rating:1478.1
    Mixed Av Seconds:02:11
    Mixed Attempts:139
    Mixed Success Rate:61.15%

    The Standard Rating is ~100 points lower with ~3 times as much thinking time. With 3 times as much thinking time the difference in rating is at average higher.
    The mixed rating is ~100 points lower too but only wirh 2 times as much thinking.

    The mixed mode include ~~50% defencive puzzles. So a mixed mode player is usually checking for any dangers from the opponent first and will see the Qg4 as possible source of trouble first. A piece at the own territory has to be attacked and puhed back (thats common "positional" knowledge),.. with the consequence of a hanging knight.
    A good positional move is a good move "by chance", it increases the chances of our play. So , by chance a good positional move is "often" a good tactical move too.
    Positional understanding can! help in tactics.

    Thats one reason why i do some Convekta Strategy 3.0 training at the moment. ( And its fun too ;)
    Here some SciFi about Convekta Software
    http://www.ez-net.com/~mephisto/Become%20Master%20Rated.html
    http://chessok.com/?p=21207

    ReplyDelete