"What's bothering me is, the computer analysis doesn't see Be5 as a dead draw, but rather gives black a .93 edge. "
Black has almost a pawn advantage because he is a pawn up.
Not if Black can't convert it into a win. Like I said, I played this position out against the computer, and eventually it was showing 0.00 even though black was "up a pawn".
"What's bothering me is, the computer analysis doesn't see Be5 as a dead draw, but rather gives black a .93 edge. "
Black has almost a pawn advantage because he is a pawn up.
Not if Black can't convert it into a win. Like I said, I played this position out against the computer, and eventually it was showing 0.00 even though black was "up a pawn".
Exactly..."if" Black has the material and winning chances here.
That is exactly my question, how is it possible for Black to force a win. I don't see any way, and every engine I play the game out against by just moving the bishop along the diagonal eventually settles on 0.00.
So what is Black's winning possibility?
"What's bothering me is, the computer analysis doesn't see Be5 as a dead draw, but rather gives black a .93 edge. "
Black has almost a pawn advantage because he is a pawn up.
Not if Black can't convert it into a win. Like I said, I played this position out against the computer, and eventually it was showing 0.00 even though black was "up a pawn".
Exactly..."if" Black has the material and winning chances here.
That is exactly my question, how is it possible for Black to force a win. I don't see any way, and every engine I play the game out against by just moving the bishop along the diagonal eventually settles on 0.00.
So what is Black's winning possibility?
Easy...people make mistakes. You're looking at this strictly from a chess engine POV.
I don't understand what you are saying. The engine evaluates assuming best play.
The drawing idea is simply to move the bishop back and forth on the diagonal. Are you saying the engine thinks white will move off the diagonal, and finally decides that it won't after 30 moves or so?
What the author of the Original Post does not realize is that computers are not chess gods. Computers will not miss 5-move forcing sequences that win outright that humans may not find. However, they are terrible at evaluating positions. They overrate material. For example, I've seen computers give a +3 in scenarios where White has a Rook and Knight, Black has a Rook, neither side has an immediate tactic to win material, and neither side has a pawn. We all know that's a draw, but computers overuse material count to evaluate positions.
Computers can't evaluate Openings or Endings properly. It thinks the King's Indian Defense is +1. It does not realize until moves later that White's advantage is nowhere near that large!
Too many of the younger generation think that the computer is the end all, be all of chess. They are not! They will find forcing lines far better than we ever will. A 2020 computer will calculate far faster than we ever do and would not lose on time against a human. But they also evaluate positions incorrectly.
If you take a Correspondence Chess site that allows use of artificial intelligence, like ICCF, and you take a Correspondence master and put him up against a 1700 player that uses artificial intelligence exclusively to find his moves, the Master will always win. Yes, he will use the bot for blunder checks, forced sequences, etc, but he will also be able to tell false evaluations better than the 1700 player. 22.Ne5 says +0.83 while 22.Nd2 says +0.58, but the Correspondence Master understands the positional aspects better than the artificial intelligence, and figures out long term that 22.Nd2 is stronger than 22.Ne5.
If you are strictly relying on computers to do all the analysis work for you, you are making a huge mistake!
Yep, as ThrillerFan says. Software works by calculating millions of lines, and low grade evaluation. A Human evalates as sac bishop for h-pawn, then wrong colour rook pawn draw, end off.
You are missing something: the fact that a computer gives black a small advantage in a drawn position happens because the computer you are using hasn't god enough power to go to the required depth or you haven't given the engine enough time to evaluate. Computers don't have limits, and are the best at chess. It's the technology we are using that limits them
With correct play, the game should end in a draw. However, checkmate by a series of legal moves is possible, so it is not a draw by the rules of chess (it doesn't meet the criteria for insufficient material).
This is a very easy draw. The a-pawn can't win alone, because black's bishop has the wrong color. Therefore all white has to do is sacrifice the bishop for the h-pawn and the king can just move to a1-b2.
The engine evaluation is overrated. It can even evaluate drawn rook endgames as +1.5. That does not mean it can't find the best moves though.
If an engine sees a forced win, the evaluation isn't going to be <+1.0. It will be much higher. An engine evaluation is just a number an engine uses to sort the available moves. Two different positions may have the same engine evaluation, while one is winning while the other is not.
Ok, so I was white this game, and this was the position heading into move 48:
My first thought was to play Be5, as after Black takes the pawn I didn't see any way for him to make progress, but then I realized that I could protect the pawn with Bg1, and I got careless and made the move quickly without realizing the bishop would be trapped after ...Kg2. So Bg1 lost on the spot.
What's bothering me is, the computer analysis doesn't see Be5 as a dead draw, but rather gives black a .93 edge. I can't figure out for the life of me how it wouldn't be a draw after 48. Be5, I even continued the game against the computer on the highest level and by around move 100 it was showing 0.00. All I was doing was moving the bishop along the diagonal.
How does Black have a path to win after 48. Be5?
Full game below: