Deep Thought.. What's this rules of Bishops?
Aravind Anumala
COMP did not know??
--Pinto.
Robert Hyatt
That they are almost all draws....
--
Robert Hyatt Computer and Information Sciences
hy...@cis.uab.edu University of Alabama at Birmingham
(205) 934-2213 115A Campbell Hall, UAB Station
(205) 934-5473 FAX Birmingham, AL 35294-1170
John P DeMastri
even with a P advantage, it's impossible to win with only Bs of
opposite color on the table. Apparently the computer did not
know this and traded down from a possibly won position to this
dead drawn one.
Feng-Hsiung Hsu
It did know the rule that you are refering to. There are
other rules though, and even the chess expert on the team did
not know about one of the additional rules.
Robert Hyatt
John P DeMastri <7117...@CompuServe.COM> wrote:
>I can only imagnie that the "rule of opposite bishops" is that
>even with a P advantage, it's impossible to win with only Bs of
>opposite color on the table. Apparently the computer did not
>know this and traded down from a possibly won position to this
>dead drawn one.
There are lots of KB+KBPP that are drawn with opposite color
bishops, even if the pawns are connected. The bishop can blockade
them very effectively, and with no help from the wrong-colored
bishop, a king can't drive the bishop away from its blockade...
I think the worst case is KBP vs K where the bishop is not
the right color (doesn't attack the queening square) so that
+4pawns ahead, it's a draw... I've seen crafty fall into
this, winning one last pawn, but trading into a completely
drawn ending. Cray Blitz understands these things well, but
I haven't ported that knowledge to Crafty yet... soon, of
course (just like everything else is "soon"...)
Robert Hyatt
Feng-Hsiung Hsu <fh...@cs.cmu.edu> wrote:
>In article <3mj875$i1q$1...@mhadg.production.compuserve.com>,
>John P DeMastri <7117...@CompuServe.COM> wrote:
>>I can only imagnie that the "rule of opposite bishops" is that
>>even with a P advantage, it's impossible to win with only Bs of
>>opposite color on the table. Apparently the computer did not
>>know this and traded down from a possibly won position to this
>>dead drawn one.
>
>other rules though, and even the chess expert on the team did
>not know about one of the additional rules.
One that Cray Blitz got burned by years ago was a bishops of
opposite color where the other side had three pawns in the center
that were passed (CB had pawns elsewhere that were passed) but one
of the opponent's center pawns was doubled. This became important
as that extra pawn behaved somewhat like another bishop since it
could be used to defend a square that the bishop couldn't... We
lost that game, and CB became "smarter" the next day... :^)
Steven J. Edwards
>I think the worst case is KBP vs K where the bishop is not
>the right color (doesn't attack the queening square) so that
>+4pawns ahead, it's a draw... I've seen crafty fall into
>this, winning one last pawn, but trading into a completely
>drawn ending. Cray Blitz understands these things well, but
>I haven't ported that knowledge to Crafty yet... soon, of
>course (just like everything else is "soon"...)
Armed with the KBPK tablebase pair, Spector can play it perfectly for
both sides. Alas, the search only probes the tablebase set for ply
one positions, so the omniscence is not available deeper in the tree
where it would be most helpful. The difficulty is that each probe to
disk can cost a whopping 10 milliseconds and this adds up when there
are millions of nodes to examine.
My thought is to permit probes at nodes up to N/2 levels deep where N
is the nominal search depth. Another idea is permit probes at any
level as long as the cumulative probe time does not exceed (say) 10
percent of the nominal search time. Some experiments are needed to be
sure. Perhaps other researchers could comment on their probe
strategies where plenty of endgame data is available but unlimited RAM
is not.
-- Steven (s...@mv.mv.com)
Thorsten Heedt
: What's this rules of Bishops of Opposite colors that the
: COMP did not know??
:
:
: --Pinto.
In most of that endgame positions, if white has for example a bishop on
the white squares and black one on the black squares, most of the endings
are draw, even if one of the opponents has one or two pawns more. So
there were some games of Deep Thought in which he had a pawn more but he
couldn't win because of that. The Grandmaster knew that rule (and every
patzer, too) but not the Computer. In an easy won position that machine
exchanged queens (with them on the board such an endgame is mostly won
for the attacking side) and the game resulted in a draw.
Greetings
Thorsten
Peter Claus Noehrenberg
the program to FICS from IC$, but so far no show. What's the deal here?
Shortly after the AICS=>IC$ announcement, you declared that as soon as
you worked out problems with the Zippy interface you were moving Crafty
to FICS; meanwhile, many other computers have made the transition
without difficulty--what's the holdup with Crafty?
More saliently, you had expressed outrage at the "hijacking" of AICS
and the transition to IC$, but recently your tone has muted somewhat,
even to the point of grudging approval over the "Master Lessons" and
other show-pieces at IC$. Second thoughts? Have you been given an
"offer you couldn't refuse" from IC$??
The bottom line is: are you sticking by your convictions and moving
Crafty, or are there "other forces at work"?
PCN
Martin Borriss
|> In article <3mj875$i1q$1...@mhadg.production.compuserve.com>,
|> John P DeMastri <7117...@CompuServe.COM> wrote:
|> >I can only imagnie that the "rule of opposite bishops" is that
|> >even with a P advantage, it's impossible to win with only Bs of
|> >opposite color on the table. Apparently the computer did not
|> >know this and traded down from a possibly won position to this
|> >dead drawn one.
|>
|> It did know the rule that you are refering to. There are
|> other rules though, and even the chess expert on the team did
|> not know about one of the additional rules.
No flame intended, but it would worry me if the chess expert does not
know all rules ;)
Martin
--
--------------------------------------------------------------------------
Martin Borriss bor...@rpi.edu
http://www.rpi.edu/~borrim/ mb...@irz.inf.tu-dresden.de
Randell Jesup
>There are lots of KB+KBPP that are drawn with opposite color
>bishops, even if the pawns are connected.
I recently drew KB vs KBPPP, with all his pawns connected. Both king
and bishop got in front of them, and there was no way for his king to
penetrate. (Opposite color bishops, of course.) The only trick was to get the
king placed to hold them for a sec while the bishop snagged his pawn on the
other side in exchange for my last pawn.
--
Randell Jesup, Scala US R&D
Randel...@scala.com
Ex-Commodore-Amiga Engineer, class of '94
#include <std/disclaimer>
Feng-Hsiung Hsu
>patzer, too) but not the Computer. In an easy won position that machine
>exchanged queens (with them on the board such an endgame is mostly won
>for the attacking side) and the game resulted in a draw.
You got it all wrong. It was a won ending if the queens WERE exchanged.
The unlike bishop ending is winnable, but with the queens on, it became
drawn because the machine was trying to avoid the unlike bishop ending.
Check out a good endgame book. The rules are far more complicated than
you think.
Steven Rix
In article <3mocrf$k...@watnews1.watson.ibm.com>, f...@sawmill.watson.ibm.com (Feng-Hsiung Hsu) writes:
->
->You got it all wrong. It was a won ending if the queens WERE exchanged.
->The unlike bishop ending is winnable, but with the queens on, it became
->drawn because the machine was trying to avoid the unlike bishop ending.
->Check out a good endgame book. The rules are far more complicated than
->you think.
They're not rules, they're guidelines which are generally-observed to
hold rather than anything such as a "rule" which "should never be broken".
--
Steve Rix
S....@ed.ac.uk http://www.chemeng.ed.ac.uk/people/steve/