3 ms·

Huh? The only thing I can think of is that the engine analysis can swing immediately once the move is played because it hasn’t thought about the new position fo

Huh? The only thing I can think of is that the engine analysis can swing immediately once the move is played because it hasn’t thought about the new position for very long. Obviously if you just watch a stream the evaluation changes immediately. It would be very rare for the computer in the first place not to have a strong move as one of its candidate entries. Obviously the evaluation can shift but the difference shouldn’t be apparent to a human player and it would be even more unlikely that a computer would reject a top move but a human would find it.

dalbasal 5y ago
This usually happens in very open position, with many <0.5cp options. The engines tend to explore totally different paths than humans. They don't care if a move is slow, forcing, etc. Humans will tend to explore exchanges, faster & more forceful lines. Engines treat these like any other line, without prejudice. I'm a pretty average player. My positional moves aren't generally reasoned beyond 2-3 moves ahead. I go here. You do something, then I go there. Where I do reason deeper, it's usually forceful lines where a lot of material is traded. I go here. We trade 3 pieces. I attack this. Opponent defends, then I move there. In a slower position, I can't reason ahead as much because I have no idea what the opponent might do. He has too many options.
- valec 5y ago
  > Engines treat these like any other line, without prejudice. i'm not sure if i'm understanding 100% what you're saying but don't most engine heuristics try more potentially favorable first, such as moving a pawn? i think most modern engines have pretty aggressive move ordering
  - codeflo 5y ago
    You’re correct. The more aggressive version of this is called “move pruning”, effectively cutting off entire branches of the search tree. Techniques like this are essential for the strength of modern chess engines, but like every heuristic, they can be wrong. Sometimes, the engine makes a mistake because it literally didn’t consider a certain response.
  - dalbasal 5y ago
    Kind of, but they don't know which one is more favourable and resources are limited. If there are lots of seemingly OK options, an engine will have to spread those resources and explore lots of moves. Chess engines tend to find good moves that seem really passive to humans. They don't create a threat or force an exchange. The purpose of the move may not be that complex, the payoff will become obvious 2-3 moves later. A human just wouldn't have thought to even explore it. There are no hints (like threats and exchanges) that it is worth exploring. That's what I meant by prejudice. Conversely, even in very option-rich positions... humans will explore the more active lines, and calculate material exchanges more deeply than other moves. If a normal player is calculating 7 moves deep in the mid game, that probably means a lot of material is being exchanged. Remember that an equally matched engine is exploring for more moves than you are. The human is better at deciding which move to explore.

dalbasal 5y ago

This usually happens in very open position, with many <0.5cp options. The engines tend to explore totally different paths than humans. They don't care if a move is slow, forcing, etc. Humans will tend to explore exchanges, faster & more forceful lines. Engines treat these like any other line, without prejudice. I'm a pretty average player. My positional moves aren't generally reasoned beyond 2-3 moves ahead. I go here. You do something, then I go there. Where I do reason deeper, it's usually forceful lines where a lot of material is traded. I go here. We trade 3 pieces. I attack this. Opponent defends, then I move there. In a slower position, I can't reason ahead as much because I have no idea what the opponent might do. He has too many options.

valec 5y ago
> Engines treat these like any other line, without prejudice. i'm not sure if i'm understanding 100% what you're saying but don't most engine heuristics try more potentially favorable first, such as moving a pawn? i think most modern engines have pretty aggressive move ordering
- codeflo 5y ago
  You’re correct. The more aggressive version of this is called “move pruning”, effectively cutting off entire branches of the search tree. Techniques like this are essential for the strength of modern chess engines, but like every heuristic, they can be wrong. Sometimes, the engine makes a mistake because it literally didn’t consider a certain response.
- dalbasal 5y ago
  Kind of, but they don't know which one is more favourable and resources are limited. If there are lots of seemingly OK options, an engine will have to spread those resources and explore lots of moves. Chess engines tend to find good moves that seem really passive to humans. They don't create a threat or force an exchange. The purpose of the move may not be that complex, the payoff will become obvious 2-3 moves later. A human just wouldn't have thought to even explore it. There are no hints (like threats and exchanges) that it is worth exploring. That's what I meant by prejudice. Conversely, even in very option-rich positions... humans will explore the more active lines, and calculate material exchanges more deeply than other moves. If a normal player is calculating 7 moves deep in the mid game, that probably means a lot of material is being exchanged. Remember that an equally matched engine is exploring for more moves than you are. The human is better at deciding which move to explore.