Learning chess as research: Functionality of the pieces
When I was young, to learn opening moves, we’d learn variants by heart from the Encyclopedia of Chess Openings (or the “Yugoslavian Encyclopedia” as we used to call it). While chess openings books at the time provided variants, they didn’t explain why White or Black was supposed to play these moves. While learning by heart is an integral part of learning chess in particular and learning in general, blind rote is an undesirable way to understand the demands of the position and the objectives of the played moves.
Even Carlsen found it difficult to comprehend the objectives of the moves Kasparov recommended he make when playing Kramnik, so how could a lesser player than Carlsen be expected to? So no, rote learning is not the main way to learn chess. True learning means to understand the relevant ideas of a chess position and how the good moves attempt to realize these ideas. It’s a deeper understanding of the precise combination of stratagems and general principles as they’re manifested in an actual chess position. This is the true, effective way to learn chess and to enjoy the beauty that lies in the masters’ games.
DecodeChess is a wonderful tool for active learning. Here’s an example: The following position (Kramnik-Leko, Budapest Rapid 2001(3)) is very interesting and replete with ideas.
Chess programs reveal to us in a matter of seconds that the best moves here are 21.Qa1 and 21.Qa3. So why are these good moves, and why is Qa1 better than Qa3? In other articles in this series, we’ll see how DecodeChess helps us understand the answers to this question. Right now we’ll focus on the question of pieces’ functionalities; we’ll see how learning the functionalities using DecodeChess helps us to understand the complexity and the beauty of the position, as well as improve our “chess-specific” understanding. A quick glance at the pieces’ functionalities shown by DecodeChess shows both clear and hidden elements.
While it’s obvious that the role of the arrow from the black pawn at e6 to Bd5 (the pawn at e6 defends the d5 bishop), Rf8’s functionality at defending square f7 is not.
After all, not only is no white piece threatening the f7 square, but it’s defended by two black pieces. Is this a bug in DecodeChess’s algorithm? Clicking the arrow reveals what the algorithm came up with: “The f8 rook protects its king by guarding square f7 – 21.Bxd5Pxd5 22.Qb7Nf6 23.Qxf7+Rxf7.”
Well, square f7 is indeed vulnerable, and it’s important to keep the rook on f8 to defend it. Can White do something about it? It certainly seems so, as playing Bd6 threatens the rook. Let’s try: 21.Bxd5Pxd5 22.Bd6Re8 23.Qb7, the black knight appears lost as square f7 is now vulnerable. However, clicking the “What if?” button reveals that 23…Qf6 is the recommended move. Why? Because of the lovely variation: 26.Qxd7Rcd8 …and the bishop at d6 is lost!
So, did square f7’s vulnerability play a major part in the final decision of what to play? Ultimately, square f7’s vulnerability is not the decisive factor of the position. On the other hand, should we not have looked at it? Certainly so! Does this deepen our chess understanding and make us better chess players? Definitely. This is the kind of observing and understanding that was somewhat lost with the chess engines, and can be enhanced using DecodeChess’s smart algorithms.