chess online

chess online
GameKnot related: GK Analysis - funny mistake!
« Back to forum
FromMessage
black_cat_hamlet
13-Jun-11, 10:17

GK Analysis - funny mistake!
game

It was an interesting game overall - a lesson on not giving up that I learnt the hard way - but
the interesting part's at move 22. Obviously a blunder that I should have seized on...but then
look ahead to GK's alternate line for White's move 24...  
jstevens1
13-Jun-11, 12:07

Indeed, Alex .....
Black can play 24. ....... Bxa1!
myrydin
14-Jun-11, 08:27

Not sure I understand. Bxa1 is unavailable on move 24. I think these takes are sometimes missed when you have to squeeze diagonally between pawns. The alternative line suggested seems strange to me.
sccadams
22-Jun-11, 12:36

I have one of those
From a recent win: game

On move 24 I miss mate in 4, but my move still forces mate in 6. However, the GK continuation misses the winning move 27. Rd4. Try to get black out of that.

On the mate in 6, black gains a move by 24...e5. The GK continuation leads to mate in 5.
black_cat_hamlet
22-Jun-11, 13:14

Guess there are still a few bugs there...
But still, on the whole I find the GK analysis useful indeed!  
Gameknot.com
22-Jun-11, 15:14

Both of the situations described above are not bugs. A "bug" is when something is behaving incorrectly, or doing something wrong. In this case, this is how the computer analysis actually works. A chess engine will never "think" or produce the same analysis as a human player would.

x-machine, in your game, the chess engine saw that your opponent repeated the same position by moving the Rook to its previous square and it assumed that the game can be force into a threefold repetition draw. It's a well-established shortcut in the computer analysis that if a position was repeated twice, it will be repeated 3 times (simply because if a particular move was chosen previously, why wouldn't it be chosen again given the exact same position as before?), which means the game will be drawn. However, humans don't work that way, and a human player might have simply overlooked something the first time, and will choose a different move the next time in the same position, because they found a potentially better move.

sccadams, in your game, the chess engine simply stopped looking any deeper once it found the mate in 4. It doesn't mean it missed the mate in 6, it just wasn't even trying to find it any more. Yes, in an actual game, it doesn't really matter if you win in 4 moves or in 6 moves. However, that's not what the post-game computer analysis is for -- it shows you where you (or your opponent) missed a better move. And a move leading to a mate in 4 is definitely better than a move leading to a mate in 6, which is what the computer analysis is designed to find.

The bottom line is that every tool has its purpose, and it's unfair to blame a microscope for not working too well for looking at the stars, or for hammering nails.  
black_cat_hamlet
22-Jun-11, 15:19

You know, what I like the most about this particular website is that the owners aren't just some
invisible hand running the whole thing and never replying to anyone who uses the website!  

Maybe I didn't explain that well, but you get the gist!  
thereaper1
24-Jun-11, 00:24

Just out of curiosity, what is the engines view horizon? I have had post game analysis that has listed moves as blunders and yet the lines it provides actually lead to mating sequences.