Forums

What is the computer analysis trying to pull?

Sort:
Avatar of JeffMGreg
JeffMGreg
Apr 5, 2009
0
#1

So I played a game and won by resignation.  I sent in for the analysis of the game.  Looking at the last "play might have continued" line I see, well it looks like to me, that the analysis set my opponent up for a nice mate on me.  Doesn't the analysis have ~2500 rating?  How could this happen?  Has this happened to anyone else?

Avatar of spherenine
spherenine
Apr 5, 2009
0
#2

Yes.

Avatar of DBucky8
DBucky8
Apr 5, 2009
0
#3

Yes, and thats pretty funny:)

Avatar of TNziggy0076
TNziggy0076
Apr 5, 2009
0
#4

wow... how about that... i guess its just not meant to see more than that many moves no mater what level its programed at

Avatar of lastwarrior2010
lastwarrior2010
Apr 5, 2009
0
#5

yes this happends a lot, it looks for a lower depth and stops, it would see the mate if it looks 1 more move into the variation.

Avatar of bSlaguitarro
bSlaguitarro
Apr 6, 2009
0
#6

22...Rd1#

Avatar of theturtlemoves
FM theturtlemoves
Apr 6, 2009
0
#7

This is called the computers horizon so it "only" looks eight moves ahead, in this case.

Avatar of Subrosian
Subrosian
Apr 6, 2009
0
#8

Are you sure the computer didn't play 21. Rcxc7, instead of Rdxc7?

Avatar of RC_Woods
RC_Woods
Apr 15, 2009
0
#9

Chess.com computer analysis, it turns out, is unusually weak. I have a class example of this as well. See the following game. Move list shows Chess.Com engine (CCE) responding with exd5 which is very weak: