The 4 Levels of Chess Thinking
I have come to believe that there are different levels of chess thinking. At the very bottom level is your basic block before you punch level. This is your fundamental attack and defence. You attack and defend and anticipate your opponent's responses in turn.
The second level is more advanced and is the positional play level. Here you know how to optimize your position in one or two move steps. For example, following the positional rule of placing a rook on an open file. Optimizing the placement of your pieces etc.
The 3rd level would be judgement and planning. This is thinking in basic schemes and concepts. This entails the ability to create plans when the course of play is unclear. Perhaps attacking and defending moves do not yield any substantive progress. Perhaps you are unable to optimize your pieces in an obvious or meaningful way. This is where deep strategic thought predominates. This is where you evaluate a position according to its elements or according to its imbalances and then formulate strategic manoeuvres by co-ordinating your pieces to achieve a specific aim. This is the level that makes superior endgame players like Magnus Carlsen, Smyslov, Capablanca, Fischer, Beliavsky etc. In the endgame, these players are no longer focusing on single attacking and defensive moves or a simple improving move - they are looking at deep strategy.
The 4th level is integrated calculation. This is the step that allowed the legendary Grandmaster and prolific author Alexander Kotov to rise "meteorically" and achieve elite levels in Soviet chess. More than just attack and defense, optimizing the position, and planning prowess, one needs accurate calculation to back their deep reasoning. I do not believe that any human can master this using just their mind. This kind of thinking is better demonstrated in the elite levels of correspondence chess. In correspondence chess, a given player is allowed to analyze the position using a chessboard i.e. they can move the pieces around before making a move. Any grandmaster with the luxury of moving the pieces around, to aid their analysis, before making a move would easily beat Magnus Carlsen, naturally.
I believe the 4th level of integrated calculation was exemplified by neural network engines like AlphaZero and Leela Chess 0. These engines calculate fewer variations than traditional versions of engines like the old Stockfish versions (not the new one with NNUE technology). If memory serves, I once heard that these neural-network-based engines may only calculate a few hundred variations while traditional engines, before the introduction of AlphaZero, would calculate hundreds of millions of positions per second.
Because I originated these "levels of chess thinking," I may experience a small sense of pride and that may be dangerous. As a result, I welcome and appreciate any and all criticism of my schema/theory/model to bring me down to earth somewhat.
N.B. This was originally intended to be a chapter in a book I was working on. I decided to write a short blog about it.
