Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: That's funny

Author: martin fierz

Date: 22:05:46 10/16/02

Go up one level in this thread


On October 17, 2002 at 00:49:56, Robert Hyatt wrote:

>On October 17, 2002 at 00:18:02, martin fierz wrote:
>
>>On October 16, 2002 at 22:51:14, Robert Hyatt wrote:
>>
>>>On October 16, 2002 at 20:35:49, martin fierz wrote:
>>>
>>>>On October 16, 2002 at 18:03:10, Johan De Bock wrote:
>>>>
>>>>>Last line of the live analysis of game 6 of Kramnik-DF:
>>>>>
>>>>>Mig: Btw, if white takes the bishop on a6, then b2 is a winning shot.
>>>>>
>>>>>
>>>>>Isn't that funny :-)
>>>>
>>>>i'm not sure why that should be funny. i have seen no analysis which proves that
>>>>the position after b2 is in fact a draw. besides, it looks completely lost,
>>>>which is also the reason why kramnik resigned - it's quite natural to assume
>>>>that it is lost.
>>>>if it is in fact a draw, it is still very doubtful that kramnik would have been
>>>>able to hold the endgame as obviously a lot of precision on the white side is
>>>>required (unlike the famous DB-kasparov game where kasparov missed a perpetual
>>>>check).
>>>>
>>>>aloha
>>>>  martin
>>>
>>>Why is the DB/Kasparov draw _easier_?  It took a _bunch_ of us, working all
>>>night, to
>>>prove that Re8 led to a draw, and the line was very precise.  One wrong move and
>>>the
>>>draw turned into a loss instantly...
>>>
>>>It doesn't seem that "easy_ to me...
>>>
>>>It wasn't so easy for Kasparov either.  :)
>>
>>i guess i'm wrong then :-)
>>all i remembered was that people said "kasparov missed a perpetual". and i
>>thought i remembered that kasparov was disgusted that he missed it - but my
>>memory is of course not very reliable :-)
>>a perpetual is something you can calculate. in the kramnik DF potential draw,
>>*if* it were possible for white to force that RPP-QP drawn endgame, then it
>>would be "easy" in the sense that you can find it at some point, and once you
>>have found it it's game over again. however, if white cannot force that endgame,
>>then he has to defend RR-QN+passed-pawn, with great accuracy.
>>why is one easier than another? because once you see the perpetual, you see it
>>and it's a draw - and you cannot lose any more.
>>defending a possibly tenable but inferior position on the other hand is *never*
>>over. the computer will torture you for 50 or 100 moves (assuming you cannot get
>>into this drawn RPP-QP endgame), and any slip will lose the game.
>>
>>aloha
>>  martin
>
>
>
>
>The DB perpetual was very deep.  IE there were many "set-up" moves that had to
>be
>played perfectly, before the position with the perpetual was reached...  Many of
>the moves
>are one-only type moves, but they are very difficult to find if you don' t know
>they are
>there, which is why it took almost 24 hours to convince everyone that this was
>drawn.
>Everytime a draw was disproven, another was found, until it finally could not be
>refuted...

i looked at the position of DB-kasparov again, and i stay with my conclusion:
the DB-kasparov draw is easier. two reasons:
1) it seems the kasparov-DB draw can be proven. there are not that many lines
that white can try alternatively. the kramnik-DF potential draw will be very
hard to prove if it is provable at all.
2) much more important: in the DB-kasparov position every decent chess player
will see that Qe3 is *potentially* a perpetual. i'm not saying he will see a
perpetual (probably he won't), but he will easily recognize the "drawing
potential" in this position. in the kramnik-DF game it takes quite a conceptual
leap to realize that you can draw with 2 rooks against Q+N+passed pawn. normal
"chess common sense" just tells you that this is dead lost, while the same
normal chess common sense tells you that the kasparov position might be a
perpetual.
so even if the kramnik position does turn out to be provably a draw (and
therefore computationally approximately as easy as the kasparov draw), it is
conceptually much harder to find the kramnik draw.

aloha
  martin



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.