Improve your chess understanding using Stockfish!

How does the engine think?

When we are deciding what to play in any given chess position, we consider some of our possibilities and our opponent’s probable good replies, and then we try to judge who’s better at that point we are anticipating; maybe two, three, four moves in advance: “I want to play this, then he/she could play this, and after that I have this move and I like this resulting position”.

Unlike our human minds, which start considering a move based mainly on our chess knowledge, a chess engine can “calculate” all possibilities in any position.

The engine then determines who has the advantage using evaluation functions based on mathematical algorithms and parameters. Such as material unbalance, central pawns, exposed enemy king. It is extremely complicated, at least for me! (If you’re interested on learning more you can check out this article https://chess24.com/en/read/news/how-do-chess-engines-think)

Long story short, these evaluation parameters aren’t flawless; there are certain gaps where they simply do not apply.

When the machine gets it wrong:

The next game, played by actual people, for real, is a brilliant example of when the “machine” gets it wrong.

Black's turn, Stockfish is showing winning advantage for white, did you verify that? It is reasonable, the g3-bishop is worthless for black and b6 in the next move seems unavoidable, after that black’s position collapses. Hot to prevent b6? There’s no way, apparently. Stockfish shows Be1, Qa8, Kb7..., these all lead to an excruciating position Black. But it gets interesting, THERE IS a way to stop b6, is just that the “machine” can’t "see" it, but a human player did, in the actual game, Laszlo Hazai played

b6 is stopped!! Is this a good move? Stockfish keeps saying white is winning...

After taking the Queen, white plays h4. The only way to try to open a path for the Queen

3...gxh4 4.Qd2 h3!! was played. No open diagonal! And also a move the machine can't "see"

There's nothing better for white than a draw. A few more moves were played. I bet the player with the white pieces couldn't believe his eyes, nor could I if this had happened to me. Stockfish still doesn’t "believe" it, keeps showing winning advantage for white. But the fact is that material advantage is only a real advantage if it's effective to checkmate the opponent's King, which is never gonna happen here. This is a game a human player find the way to save, and "the machine" would have lost.

When the machine gets is super right:

Ok, so the “machine” can get it wrong a tiny percentage of the times, very unusual times, what about when it gets it super right, but it is almost impossible for a human player to emulate or to imitate its “reasoning”.

White to play. At this point g4 had just been played, and Kortschnoj faced the threat of the collapse of the king side pawn barrier, especially h3 is threatened. So he, most humanly, played fxg4??. Seems very reasonable and close to the only way to defend the king side, but it surrenders the battle entirely, black pieces find their way in all the same, and there's no way to stop it.

So, what's Stockfish's solution to this problem? Not to defend h3 at all!

Looks surreal to give up, kind of, the defense of the king side, very few players would even consider this, Kortschnoj didn't. Easy for Stockfish to "see" it, not for a human player. The justification for this is in move 25.a7!, you'll see...

The most human reply. Results in a completely lost position for black. Bxa6 (not a good idea), g3 (better), and Qg7 (actual good fighting chance) were other possibilities, but lets not get into that.

King's Indian players know that losing the c8-bishop could very well mean losing all chances to attack the king side, and in this defense, it all comes down to attacking the king. From a human point of view, to sacrifice material in exchange for that bishop makes perfect sense for white pieces, although it's hard to considerate it from 23.a6!!. For the machine, it "makes sense" simply because it's anticipating all possible outcomes, and very few humans, if any, can actually do that

The king gets ready to run to the queen side, will be safer there. Bxf1 and Qxf1 would also work, but I think running with the king is more human and simplier and the simplest answer is ussually the right one.

No way to stop white's threats, no counterplay, no defense, no nothing.

Complicated, messy, tactic positions, that can be solved with pure calculation, that's when the engines do their best work. The "machine", unlike us humans, doesn't sweat, has no expectations, no knowledge bias, and simply anticipates almost everything.

How to use Stockfish then? Some tips:

Chess games aren’t perfect, not even Carlsen’s, that's not the goal, we all make mistakes.

When you play a chess game, you’re playing with your own mind, your own ideas, not Stockfish’s. So build your own chess philosophy, study general concepts, get interested about the functioning of it all, as in what idea supports what move (rooks go in open files because they can attack/penetrate enemy territory from there, for example).

Don’t panic, chess knowledge is extensive, but each idea/concept you learn will make you improve, and that’s when Stockfish can become your biggest ally.

Use Stockfish as an advisor, don’t take the moves it is showing just because, go a few moves beyond and try to find out why it “thinks” that’s the best move.

As in Kortschnoj – Kasparov: 23.a6!! isn’t a good move without 25.a7!. And, besides all the calculations Stockfish is using to validate that , there’s also an idea behind it: strong counterplay in the queenside, and “defending” the king side by making black surrender its c8- bishop, thus disabling the attack.

Stockfish can be an extremely powerful weapon if you use it to understand chess better.

A puzzle for fun at the study! Do you dare to try and solve it? Good luck!

Can you find the move that Stockfish can't "see"? White to play, and force DRAW...

I hope you found this entertaining and useful, if you liked it, please click on the heart, thank you! I wish you the best of lucks!