Why is Stockfish obsessed with the h7 pawn? Nf3-Ng5-Nxh7

Laskersnephew

In #37, White didn't complete tripling because Black gave him an opportunity to win decisive material, That's how chess works: When your opponent moves have to rethink your plans

GambitShift

GambitShift
Laskersnephew wrote:

In #37, White didn't complete tripling because Black gave him an opportunity to win decisive material, That's how chess works: When your opponent moves have to rethink your plans

 

Yes, you could triple or you could double. 

blueemu

Just a general remark:

I'm not anti-computer. I worked for 20 years as a computer tech making cartoons, including Hollywood feature films. I've also worked as a computer systems analyst for the military. I like computers.

But a lot of the people on this site seem to act as though Stockfish (or Leela, or AlphaZero) was God's own oracle, and they waste countless hours trying to interpret the "revealed wisdom".

If you are using a computer to analyze your chess game, it is essential to bear in mind that what a chess engine does is not "thinking", as humans understand the word. They examine hundreds of millions of positions, grade each position according to some internal parameters, and decide on a path through the search tree based on those grades. The highest scoring path indicates the move to play.

This has NOTHING to do with the way humans analyze a position and decide on a move. No connection at all.

So trying to guess "what Stockfish was thinking" is automatically an exercise in futility. That's why I urged you to learn to play like a human, rather than trying (and of course, failing) to learn to play like a computer.

Thoroen

The computer dosen't care about h7. It wants to win. Fastest way of winning seems to be a knight on e6. However in blacks desperado attempts to save materiale gives up a pawn. White dosen't care about a treath on h2 because it does not exist, but when blacks Best opportunity is to offer material Black is going to do just that, and White deviates from the plans to gobble up material.

You Are applying human conciderations to machines. 

Laskersnephew

But a lot of the people on this site seem to act as though Stockfish (or Leela, or AlphaZero) was God's own oracle, and they waste countless hours trying to interpret the "revealed wisdom".

That is right on! But the computers are awfully good at calculating relatively short tactical sequences, so you have to pay them attention. Chess is a strategic game, but it's also a concrete, tactical game. In fact, it's the blend of those two qualities that make it such a great game

k00m00d00

 

blueemu
Laskersnephew wrote:

But a lot of the people on this site seem to act as though Stockfish (or Leela, or AlphaZero) was God's own oracle, and they waste countless hours trying to interpret the "revealed wisdom".

That is right on! But the computers are awfully good at calculating relatively short tactical sequences, so you have to pay them attention. Chess is a strategic game, but it's also a concrete, tactical game. In fact, it's the blend of those two qualities that make it such a great game

Yes, certainly. A tactical sequence is one thing. Or using an engine to blunder-check. That's fine.

But wondering what it's thinking?

It's thinking "01100001 10100110 11101001 00111110".

pfren

White is ahead in development, also two healthy pawns up, he has absolutely no weknesses, and Black's king is still stuck in the center. This means that you should not bother that much about what Stockfish is claiming. Ng5, Qe3, Re3, all of them are more than enough to win the game.

Laskersnephew

But wondering what it's thinking?

It's thinking "01100001 10100110 11101001 00111110".

 

Well, when I say something like "What is Stockfish thinking?" it's just a kind of sloppy way of saying "what does the computer see that I don't?" In fact, it' a good way to improve your tactical vision