Scotch Game; why a-pawn here?

Sort:
Avatar of xman720

Computers don't differentiate between equal positions in analysis, so if it sees it can draw with 16...Ne4, it will play that move. Unlike humans, there is no differentiation between simply draws and complicated/close draws as long as it sees a draw is forced. I can't replicate Arrakis09's computer recommendation of 16: ...Ne4, but my computer does say black can draw after 16: ...Ne4. Therefore, it makes sense that depending on how the computer chooses its candidate moves, Ne4 would be the first choice and it would play that move. It just depends on the computer Arrakis90 was using.

Specifically, the two relevant processes minimax computers use here are Alpha/Beta pruning and node reordering. The computer uses heuristics to guess at the value of a position (this is the same heuristics it uses to evaluate positions for normal usage) and then reorders nodes so that positions it t guesses are winning go first. This is to maximize the efficiency of Alpha/Beta pruning. It makes sense that Ne4 is one of the first moves the computer checks because without calculation, it looks great. Winning the pawn, centralizing the knight, you have to see Bxf7 to see the flaw. Obviously the computer then checks this move, calculates a while, and sees that Ne4 is drawing.

After this, it checks all of the other candidate moves, but with an Alpha value of 0, which is essentially black's best alternative. It uses this alpha value to prove nodes by saying that if it ever gets to a position where white is guaranteed a draw or better, it stops calculating and refers to 16: ...Ne4

Therefore, if the position is drawn and Ne4 draws, it makes sense that it heuristically chose Ne4 and then used that drawn position as the Alpha value to compare to other positions. After pruning all other positions and determining they are at most a draw, it referred back to the earliest draw move in its move order and gave Ne4 as the answer.

Avatar of pfren
xman720 έγραψε:

Computers don't differentiate between equal positions in analysis, so if it sees it can draw with 16...Ne4, it will play that move. Unlike humans, there is no differentiation between simply draws and complicated/close draws as long as it sees a draw is forced. I can't replicate Arrakis09's computer recommendation of 16: ...Ne4, but my computer does say black can draw after 16: ...Ne4. Therefore, it makes sense that depending on how the computer chooses its candidate moves, Ne4 would be the first choice and it would play that move. It just depends on the computer Arrakis90 was using.

Specifically, the two relevant processes minimax computers use here are Alpha/Beta pruning and node reordering. The computer uses heuristics to guess at the value of a position (this is the same heuristics it uses to evaluate positions for normal usage) and then reorders nodes so that positions it t guesses are winning go first. This is to maximize the efficiency of Alpha/Beta pruning. It makes sense that Ne4 is one of the first moves the computer checks because without calculation, it looks great. Winning the pawn, centralizing the knight, you have to see Bxf7 to see the flaw. Obviously the computer then checks this move, calculates a while, and sees that Ne4 is drawing.

After this, it checks all of the other candidate moves, but with an Alpha value of 0, which is essentially black's best alternative. It uses this alpha value to prove nodes by saying that if it ever gets to a position where white is guaranteed a draw or better, it stops calculating and refers to 16: ...Ne4

Therefore, if the position is drawn and Ne4 draws, it makes sense that it heuristically chose Ne4 and then used that drawn position as the Alpha value to compare to other positions. After pruning all other positions and determining they are at most a draw, it referred back to the earliest draw move in its move order and gave Ne4 as the answer.

 

Your computer says that the pawn-dropping move 16...Ne4 is Ok because it sees Black getting back the pawn with ...Re8 in the end of the line.

But: Come on! White is under pressure in any human vs human game because of the d6 pawn, and Black should try to work towards picking that pawn. Who cares if the computer can save that pawn folowing a long variation? 16...Ne4 might not put Black at a disadvanage, but in regular terms, it is a bad move.

Avatar of MaxLange-simulator

Lmao these computertards are delusional. 

Avatar of TimothyScottPuente
 
Scotch Gambit 1
Scotch Gambit 2

 

Avatar of Chesslover0_0

Well I mean maybe the computer is looking at an early half open C-file for black, other then that, I don't see why, axb6 accomplishes the same thing for the Rook, so I'm not sure. 

Avatar of Arrakis09
pfren wrote:
xman720 έγραψε:

How is this position equal? Are you really indifferent to either color here? Or are you saying it's equal because the engine says 0.00? There's a difference between 0.00 because black has a series of only-moves that lead to a close draw, or 0.00 because the position is equal and nothing white does will win. White has a clear advantage in this position. The straw d pawn will never be captured except by black's king if white's rook move. If black isn't careful, he will lose this position either by missing a tactic that allows white to play Re7 (which is devastating due to the d6 pawn) or missing a  tactic that allows white to sacrifice the d6 pawn in exchange for extraordinary piece activity. I don't buy this position is equal, I want the white pieces.

 

Any good reason for Black to drop a pawn with 16...Ne4?

White doesn't drop the pawn. In the computer lines Black immediately takes the knight with the bishop, plays his rook to the e-file, and runs the king over to capture the d-pawn. White can hold the pawn, but is busy doing that and can not make progress.