Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Questions re the Classic Shredder 6 GUI with Shredder 8 UCI engine

Author: Geert van der Wulp

Date: 01:10:50 03/04/04

Go up one level in this thread


On March 04, 2004 at 04:01:29, Ingo Bauer wrote:

>Hi
>
>>I have now answered my own question regarding the strange analyses of Shredder
>>8. They still exist with the Classic GUI, same as with the CB GUI.
>
>Yep, I suspected that allready.
>
>>And I beg to differ on that point... I consider it a full-fledged "bug" because
>>it makes the Shredder 8 engine worthless for analysis. Just the first move
>>cannot be called an analysis. Many like me do not have any interest in playing
>>chess with a computer, and only use it as an assistant in analysis.
>
>I would consider a feature related to a priciple not a Bug. (e.g. Diesel engines
>- usually very efficient but noisy, is that a bug?)
>
>But be aware that there a several engines out there that are collecting the PV
>with the same principle. That means that these engines do have the same "bug".
>(With the small difference that these engines are not as good as Shredder :-) )
>
>Second. Usually you only have this within one PV. The next or the previous PV
>(especially at long time controlls) are showing the right line (okok there migt
>be a very small chance that two following PV are wrong). If you suspect
>something weired check these two PV. If anaylsing a game and you have something
>wrong in a NEW PV you should wait for the next PV anyhow.
>
>Interesting experiment: Take Fritz (or any other engine you like). Make a 30 min
>anylsis of an unclear position. Lets asume you have a 8 move (16 ply) variant
>then. Go to the 15 ply and see if Fritz is still prefering the 16 Ply as the
>"best" move. I have not checked this but I think the possibility is very high
>that Fritz is now thinking about a different move. Question: do you trust Fritz
>PV? (I have to admit that the last move may not a bogus move but it can be a bad
>one as well.)
>I think for an analysis you NEVER should just trust the PV. The PV should be
>played move by move and then if you reach a certain depth you have to trust the
>eval or you have to analyse the game allways to the complete end.
>
>Bye and have fun with S8
>Ingo

Ingo,

You are trivializing the problem here. You are right that this problem cannot be
rightly called a BUG. It is a negative aspect of the engine. And I see very well
that this is a reason not to buy it. No matter how strong the engine might be.

But you say that one should never trust blind on the PV, that one always needs
manual assist for the PV in order to get a good analysis. I agree, but do you
not see the problem? When a human player sees the PV AS IT SHOULD BE, then it is
easy to decide whether or not one wants to have a closer look at it. If the PV
shows some crazy moves, then this will cost a lot of extra time.

Regards,

Geert



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.