cdani wrote:At last Karjakin has failed like an human...
Hi George, maybe you should run a few matches. White is winning in my tests with ease.
I do NOT have to, anybody with the latest Komodo or Stockfish can take this position and let it play both sides and see that it would have ended in a DRAW
==> [D]6k1/1b3p2/p3pPpp/6P1/1p3Q1P/1Pr5/4K3/8 b - - 0 1
We can play this out if you want to George
Okay where do you want to meet, and witout using any engines?
this is an engine forum.
you said an engine could only draw from this position.
I checked your post it is incorrect
set it up and we will play it out to see if what you post is correct or not
Why don't you find out by yourself set it up at 1 or 3 minutes per side and let Komodo or Stockfish play both sides till the very end and you post it yourelf after you learn and find out how to DRAW
would someone post a draw line. all of the lines I play using my methods win for white.
There is no draw, white wins easy. George is clinging to human chess ideas, however, they are not applicable here.
George wrote:In this position which is currently playing LIVE, Karjakin has a Rook + Bishop and 2 extra pawns to Jakovenko Queen minus 2 pawns, all engines are giving Jakovenko a huge edge , but the position is a dead draw where if played precisely it should end in a DRAW. I believe that the value of the Queen Should be set to 9 in the beginning up to the middle stage, then reduced to 8 once the position enter the endgame stage, or simply to 8.5 throughout the entire game ==>
This was the original position from where if you let Komodo or Stockfish play it out by itself both sides, the position should have ended in a draw, but for some reason every engine was giving the White side with the Queen the edge or advantage
[D]6k1/1b3p2/p3pPpp/6P1/1p3Q1P/1Pr5/4K3/8 b - - 0 1 This game is still in progress Jakovenko vs Karjakin ==> http://khantymansiysk2015.fide.com/game-analysis-2/
Such claims regarding whether positions are won, lost or drawn and that engines have no idea are made here frequently.
In most cases, the poster gets left with egg on their face.
The output shows that Stockfish misses the fortress defense. The line shown by Stockfish is full of nonsensemoves.The computer didn´t understand that it is important to defend the a6 pawn. After 42.Qc7? Rxf6 43.Qxb7 the right defensemove is 43...g5!
Komodo 9 agree with this. But with a completly wrong evaluation.
The winning line shown is typical for this fortress type: collect all the pawns and then threat to go into a winning pawn endgame.
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
BeyondCritics wrote:
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1
Please read my replies. I do not believe any of mine were "harsh." I am very aware that fortress detection is undeveloped in many engines.
Perhaps you meant to reply to some other poster(s).
BeyondCritics wrote:
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1
Please read my replies. I do not believe any of mine were "harsh." I am very aware that fortress detection is undeveloped in many engines.
Perhaps you meant to reply to some other poster(s).
My deepest apologies, i think your replies are completely correct. Thats why i have answered here.
So again, sorry for your inconvenience.
BeyondCritics wrote:
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1
Please read my replies. I do not believe any of mine were "harsh." I am very aware that fortress detection is undeveloped in many engines.
Perhaps you meant to reply to some other poster(s).
My deepest apologies, i think your replies are completely correct. Thats why i have answered here.
So again, sorry for your inconvenience.
OK, no problem then.
Curious to see how Stockfish assesses the position you posted:
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1[/quote]
BeyondCritics wrote:The winning line shown is typical for this fortress type: collect all the pawns and then threat to go into a winning pawn endgame.
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1
I think it's worth 100$. Would you like to try, if you draw it I give you 100$, if you loose, you owe me 100$?
Waiting for your move .
Thomas Lagershausen wrote:The keyposition is: [D]6k1/1b3p2/p3pPpp/8/1p4PP/1P3r2/4K3/2Q5 w - - 2 42
The computers missed the strong human move 42.g5 ! and play the stupid 42.Qc7??.
After this blunder by the computers black can build a fortress with rook and pawns and black draws the game.
Are you sure about Qc7 being a blunder? Can you provide a drawing line for black? I just ask because SF, Komodo and Houdini think it's the best move (and it's not like they didn't consider g5). Also a fortress position is never reached in their evaluation and white wins with ease. I tend to trust a computer in this kind of position.
Thomas Lagershausen wrote:The keyposition is: [D]6k1/1b3p2/p3pPpp/8/1p4PP/1P3r2/4K3/2Q5 w - - 2 42
The computers missed the strong human move 42.g5 ! and play the stupid 42.Qc7??.
After this blunder by the computers black can build a fortress with rook and pawns and black draws the game.
Are you sure about Qc7 being a blunder? Can you provide a drawing line for black? I just ask because SF, Komodo and Houdini think it's the best move (and it's not like they didn't consider g5). Also a fortress position is never reached in their evaluation and white wins with ease. I tend to trust a computer in this kind of position.
BeyondCritics wrote:The winning line shown is typical for this fortress type: collect all the pawns and then threat to go into a winning pawn endgame.
Nevertheless i disapprove the harsh tone employed by some. Did you try to understand the intent of the original poster? Is it reasonable to assume that chess programs might fail in this type of positions? What do you think of this position?
[d]8/1Q3p1k/p3prpp/8/1p4PP/1P6/4K3/8 b - - 0 1
I think it's worth 100$. Would you like to try, if you draw it I give you 100$, if you loose, you owe me 100$?
Waiting for your move .
Yeah, it looks like there's no fortress here either:
.