Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: king attack position movei-frenzee

Author: Sune Fischer

Date: 14:18:35 05/21/04

Go up one level in this thread


On May 21, 2004 at 16:16:37, Gerd Isenberg wrote:

>On May 20, 2004 at 15:41:26, Sune Fischer wrote:
>
>>On May 20, 2004 at 15:37:07, Uri Blass wrote:
>>
>>>>>>>judging by the score it seems that white has a small advantage.
>>>>>>
>>>>>>Pretty even doesn't mean 0.00, it means 0.00 +- 0.5
>>>>>
>>>>>No it means 0.00+-0.25
>>>>>
>>>>>0.32 is not pretty even.
>>>>
>>>>I disagree.
>>>>Pretty even means drawish and 0.32 is definitely _not_ a convincing winning
>>>>score by any definition.
>>>
>>>there is something in the middle between convincing winnoing score and drawish.
>>>
>>>I do not consider 0.32 as convincing winning score but I also do not consider it
>>>as drawish.
>>
>>I get 0.3 scores just by analysing the opening position, 0.3 is within drawish
>>IMO.
>>
>>-S.
>
>If a strong program sacs a piece for king-attack +0.3 is probably enough to win
>most often. If my program plays The King and has +0.3 with > 2 in material it is
>probably enough to lose ;-)

Well 0.01 can be enough if it is the right move, but I don't believe the engine
sees anything concrete at 0.3. Some positional weaknesses perhaps but those can
be insignificant in the given case, ie. simply misevaluation.

Basicly either the attack works, and then it often works very quickly and score
increases ply by play, or the exchanges just lead to a rather equal position
which can be hard to assess at the leaves.

Of course if you analyse long enough you might see +2 (or 0.00), but then again
if it takes that long it to get a convincing score I'd still claim it would be
equalish if played over the board. :)

Sometimes it is tough to find "the blunder" when looking at a lost game,
I suspect most often it is a series of non-optimal moves.
g5 may be such a non-optimal move but I suspect there were other of similar
quality in that game from black.

-S.
>Gerd



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.