Kasparov misses a draw with Qe3!
Bill Doyle
Kasparov apparently missed a draw with:
Qe3 Qxd6 Re8!
After many hours of analysis by carbon and silicon,
there is no win for Deep Blue!
Wonder how many times this has happened to Garry?
Steven C. Den Beste
Instead of taking the bishop, what if White uses its queen to check the
black King?
Since the board had simplified, one must assume that DB had a very long
look into the future, since the tree branching was no longer as wicked;
surely it could have maintained the pressure on GK to prevent a
repetitive check leading to a draw.
I know that bishop looks ripe for the plucking, but it'll still be
there, and taking it loses incentive. DB proved in this game that it was
more subtle than that.
...but then I'm not a very good chess player.
--
Steven C. Den Beste
sden...@san.rr.com "Cheap, good, soon - pick any two."
sden...@qualcomm.com
Don C. Aldrich
On Sun, 04 May 1997 23:59:12 -0400, Bill Doyle <bdo...@clark.net>
wrote:
>Kasparov apparently missed a draw with:
>
>Qe3 Qxd6 Re8!
>
>After many hours of analysis by carbon and silicon,
>there is no win for Deep Blue!
Oh, really? Well, it seems to me that a timely Qb6 by white saves
all, e.g.,
45. ... Qe3 45. Qxd6 Qxe4 46. Qc7+ Kg8 47. Qxb8+ Kh7 48. Qb6++
Or 45. .... Qc1+ 46. Ke2 Qb1+ 47. Kd1 Qxc3 (what
else?) 48. Ra7+ Kg8 49. Qxb8+ Kh7 50. Qb6 game over.
>Wonder how many times this has happened to Garry?
==Dondo
Remember: A shark is just a trial lawyer on Prozac.
Paul Hodges
Bill Doyle wrote:
>
> Kasparov apparently missed a draw with:
>
> Qe3 Qxd6 Re8!
>
> After many hours of analysis by carbon and silicon,
> there is no win for Deep Blue!
>
> Wonder how many times this has happened to Garry?
I think 45...Qe3 flat out loses to 46.Qd7+ then if
46...Be7 47.d6 +-
or 46...Kg8 47.Qxd6 +-
-- Paul
Kjell Tore Sandum
Paul Hodges <hod...@smartchess.com> wrote in article
<336D8B...@smartchess.com>...
:
In my opinion, Doyle is correct: 45...Qe3! is a draw.
After 46.Qd7+ Kg8 47.Qxd6 the simple 47...Rf8 leaves white with the same
problems of avoiding perpetual check. The main difference is that now two
more plies are added, and the draw is probably behind most computers
horizon.
KT
--
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
_/ Kjell.T...@hiMolde.no / Molde College_/
_/ Tel: +47 71 21 40 00 / \ / P. Box: 308 _/
_/ Fax: +47 71 25 25 17 / \ / N-6401 Molde _/
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
brucemo
> : I think 45...Qe3 flat out loses to 46.Qd7+ then if
> :
> : 46...Be7 47.d6 +-
> :
> : or 46...Kg8 47.Qxd6 +-
>
> In my opinion, Doyle is correct: 45...Qe3! is a draw.
>
> After 46.Qd7+ Kg8 47.Qxd6 the simple 47...Rf8 leaves white with the same
> problems of avoiding perpetual check. The main difference is that now two
> more plies are added, and the draw is probably behind most computers
> horizon.
If we missed a win for DB, it's not in this variation, I think. 46. Qxd6
seems harder.
bruce
Steven C. Den Beste
Bill Doyle wrote:
>
> Kasparov apparently missed a draw with:
>
> Qe3 Qxd6 Re8!
>
> After many hours of analysis by carbon and silicon,
> there is no win for Deep Blue!
>
> Wonder how many times this has happened to Garry?
There's an online chess column associated with the NY Times which has an
analysis of the game at:
http://www.nytimes.com/library/cyber/week/050597chess.html
Regarding this, he says:
> A desperate attempt at a perpetual check with 45 . . . Qe3 would
> fail after 46 Qd6 Qc1 47 Ke2 Qb2 48 Kf3 Qc3 49 Kg4.
Bill Doyle
>
> >Kasparov apparently missed a draw with:
> >
> >Qe3 Qxd6 Re8!
> >
> >After many hours of analysis by carbon and silicon,
> >there is no win for Deep Blue!
>
> all, e.g.,
>
> 45. ... Qe3 45. Qxd6 Qxe4 46. Qc7+ Kg8 47. Qxb8+ Kh7 48. Qb6++
>
> Or 45. .... Qc1+ 46. Ke2 Qb1+ 47. Kd1 Qxc3 (what
> else?) 48. Ra7+ Kg8 49. Qxb8+ Kh7 50. Qb6 game over.
>
>
>
> Remember: A shark is just a trial lawyer on Prozac.
You weren't paying attention:
the line is: Qe3 Qxd6 Re8! not Qxe4?? this is key.
This is still a draw 24 hours and much analysis later.
Cheers
Serge Desmarais
Steven C. Den Beste wrote:
>
> Bill Doyle wrote:
> >
> > Kasparov apparently missed a draw with:
> >
> > Qe3 Qxd6 Re8!
> >
> > After many hours of analysis by carbon and silicon,
> > there is no win for Deep Blue!
> >
>
> black King?
>
> Since the board had simplified, one must assume that DB had a very long
> look into the future, since the tree branching was no longer as wicked;
> surely it could have maintained the pressure on GK to prevent a
> repetitive check leading to a draw.
>
> I know that bishop looks ripe for the plucking, but it'll still be
> there, and taking it loses incentive. DB proved in this game that it was
> more subtle than that.
>
> ...but then I'm not a very good chess player.
>
> Steven C. Den Beste
> sden...@san.rr.com "Cheap, good, soon - pick any two."
> sden...@qualcomm.com
Here is GM Gausel variation in the post-mortem comments :
45...Qe3 46.Qxd6 Qc1+ 47.Ke2 Qb2+ 48.Kf3 Qxc3+ 49.Kg4 h5+
50.Kxh5 black's gone
Serge
eric fitch
The more that strong players look at this line, they seem to think it's
a draw. This from IM Malcom Pein on the electric minds forum:
"Hi everyone, I am writing from the web room at the
Equitable. The more I look at this position the more I
think its a draw. It was actually GM Jon Tisdall who
brough this to my attention. I have checked with Joel
Benjamin and he said he thought it was a draw but he
did point out that maybe DB could play Kh1 instead of
Kf1. Frederic Friedel disagreed at first so I told him
to switch Fritz on again and lo and behold he reckons
its a draw as well. I think Garry is going to flip when
he finds out and Fred was not looking forward to
telling him !!
The interesting thing is that everyone was psyched out
by DB and assumed Qe3 lost because if it did not then
DB would not allow it !!
It seems to me we have had a case of mass hypnosis /
delusion that took in 1000s and 1000s of players.
Deep Blue did play a fantastic positional game the
funny thing is its flaw here was that its lookahead
when it played Kf1 did not extend far enough down Qe3.
It still played a beautiful Ruy Lopez as far as I am
concerned.
Anyway, its hold the front page on The Daily Telegraph
in the UK.
Cheers
Malcolm
eric fitch
This from an ICC news item:
"NEWS 658 (05 May 97):
Was Game 2 was really a draw?
Perhaps Kasparov resigned because he thought there was nothing better
than 45...Qxc6, which loses. Perhaps he didn't see any hope to
45...Qe3 because White can easily stop the perpetual after 46 Qxd6
Qc1+ 47 Kf2 Qd2+ 49 Kg1 Qc1+ 50 Kh2 Qf4+ 51 g3 Qxe4 52 Qc7+ Kg8 53
Qxb8+ Kh7 54 Ra2. There are other ways to get to this and similar
positions where Black has no perpetual check.
ICC Admin SJLIM was reporting on the scene for ICC. He told us that
Kasparov was deep in thought and that the commentators had the Champion
buried after 43 Qc6.
However, IM Biblos on ICC soon suggested 45...Qe3 in the game
continuation. But there would no perpetual after 46 Qxd6 and Black's Rb8
would have been under fire, so analysts thought it was a lost cause.
Then, ICC admin Tim gave the key line: 45...Qe3 46 Qxd6 Re8!
Soon after, Kasparov resigned and the puzzled spectators asked, "Why
resign here? He should have taken a shot." Analysts began a post-mortem
centering on Tim's idea. Some of the key analysts during post-mortems
were
BPF, Ferret, Schroer, Securitron and sveshi. (They have not reviewed
this
and are not responsible for any errors.)
Briefly, after 45...Qe3:
a) 46 Qxd6 Re8 47 Bf3 loses the Bf3 to stop the perpetual, and then
after
Black takes the Bf3 he still threatens a perpetual.
b) 46 Qxd6 Re8 47 Qc5 keeps the game alive, but 47... Qxe4 48 d6 Qb1+
looks tenable for black.
c) 46 Qd7+ Kg8 (46...Be7 47 Qe6+ eventually wins for White) 47 Qxd6 Rf8
is
similar to 46 Qxd6 Re8.
There was a rumor that DB had evaluated the final position as +2 (i.e.,
equivalent to a 2-Pawn, or winning, advantage). Ferret said something to
the effect that no one should resign to a computer which thinks it's at
+2
in this kind of position. That's because of the evaluation error
potential
in lines where a quiet move (46...Re8) leads to positions with possible
perpetual checks. Again, the report of DB's evaluation is just a rumor.
We have not performed an analysis of the last moves before Kasparov
resigned. If the final position is equal, either DB made a big error"
And this from IM Pein at the electric minds conference site:
" Topic 40: Could Game 2 have been a draw????
#8 of 31: Malcolm Pein Mon 05 May '97 (01:37 PM)
eric fitch
This is Seirawa's ananlysis at the end:
45.Ra6!
"Now the auditorium's audience was a buzz. For some time the
commentators, IM Maurice Ashley and IM Michael Valvo were
quickly coming to the conclusion that 45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8
48.Bd5+ and Bd5-e6 was lost for Kasparov. The
desperate attempt 45...Qe3 46.Qxd6 Qxe4 (46...Qf4+ 47.Bf3; 46...Qc1+
47.Ke2 Qb2+ 48.Kd1 stops the perpetual)
47.Ra7+ Kg8 48.Qxb8+ picks up Black's Rook with check. What was Garry's
Defense? He had none and extended his hand
in resignation. The applause rocked the auditorium! Folks we have a
match! A superb game by Deep Blue and its team, my
hearty congratulations for a brilliant effort!"
Still he doesn't look into 45...Qe3 46.Qxd6 Re8. This is the only
move for black here and one that, as far as I can see, leads to a draw.
Kjell Tore Sandum
Don C. Aldrich <do...@skypoint.com> wrote in article
<336d6a99....@news.skypoint.com>...
: On Sun, 04 May 1997 23:59:12 -0400, Bill Doyle <bdo...@clark.net>
: wrote:
:
: >Kasparov apparently missed a draw with:
: >
: >Qe3 Qxd6 Re8!
: >
: >After many hours of analysis by carbon and silicon,
: >there is no win for Deep Blue!
:
: Oh, really? Well, it seems to me that a timely Qb6 by white saves
: all, e.g.,
:
: 45. ... Qe3 45. Qxd6 Qxe4 46. Qc7+ Kg8 47. Qxb8+ Kh7 48. Qb6++
:
: Or 45. .... Qc1+ 46. Ke2 Qb1+ 47. Kd1 Qxc3 (what
: else?) 48. Ra7+ Kg8 49. Qxb8+ Kh7 50. Qb6 game over.
Have you looked at "45...Qe3 45.Qxd6 46.Re8!" ?
KT
Bill Doyle
GM Gausel now admits its a draw after:
Qe3 Qxd6 Re8! (not Qc1??)
An unconfirmed report has IBM's web page under the chat
community admitting its a draw!
Cheers,
Bill
Serge Desmarais wrote:
>
> Steven C. Den Beste wrote:
> >
> > Bill Doyle wrote:
> > >
> > > Kasparov apparently missed a draw with:
> > >
> > > Qe3 Qxd6 Re8!
> > >
> > > After many hours of analysis by carbon and silicon,
> > > there is no win for Deep Blue!
> > >
Massimiliano Orsi
Bill Doyle <bdo...@clark.net> wrote:
>Kasparov apparently missed a draw with:
>Qe3 Qxd6 Re8!
>After many hours of analysis by carbon and silicon,
>there is no win for Deep Blue!
As a matter of fact, after 45. ..Qe3 46. Qxd6 Re8 47. Bf3 Qc1 48. Kf2
Qd2 49. Be2 Qf4 50. Ke1 Qc1 51. Bd1 Qxc3 (also 51. ..Qe3 52. Kf1 and
now 52. ..Qc1 53. Ke2 Qb2 54. Kf3 Qxc3 should draw; but not 52. ..Qf4?
53. Kg1 Qe3 54. Kh1! Qe1 55. Kh2 Qxe1 56. Ra7 and White mates) 52.
Kf1 Qc1 White can't improve his position.
White can play 46. Qd7 (instead of 46. Qxd6) and now 46. ..Be7
loses after 47. d6, but 46. ..Kg8 47. Qxd6 Rf8 poses the same problem
of White King's safety.
I think only White's chance is playing something like 48. Qb6 or 48.
Ra2, even after the immediate 46. Qxd6; in both cases White's
resignation would not have been on my agenda.
Did Kasparov miss something or is there something really bad in the
final position?
mclane
massimil...@iol.it (Massimiliano Orsi) wrote:
>Did Kasparov miss something or is there something really bad in the
>final position?
Ah - come one. You are all waisting time. He has lost this game on
purpose and you will find more things he had missed in this game.
The confusion about this game does not come from whites play, it comes
from GK playing without a plan.
Robert Hyatt
: > >
: > > Kasparov apparently missed a draw with:
: > >
: > > Qe3 Qxd6 Re8!
: > >
: > > After many hours of analysis by carbon and silicon,
: > > there is no win for Deep Blue!
: > > Wonder how many times this has happened to Garry?
: >
: > Instead of taking the bishop, what if White uses its queen to check the
: > black King?
: >
: > Since the board had simplified, one must assume that DB had a very long
: > look into the future, since the tree branching was no longer as wicked;
: > surely it could have maintained the pressure on GK to prevent a
: > repetitive check leading to a draw.
: >
: > I know that bishop looks ripe for the plucking, but it'll still be
: > there, and taking it loses incentive. DB proved in this game that it was
: > more subtle than that.
: >
: > ...but then I'm not a very good chess player.
: >
: > --
: > Steven C. Den Beste
: > sden...@san.rr.com "Cheap, good, soon - pick any two."
: > sden...@qualcomm.com
: Here is GM Gausel variation in the post-mortem comments :
: 45...Qe3 46.Qxd6 Qc1+ 47.Ke2 Qb2+ 48.Kf3 Qxc3+ 49.Kg4 h5+
: 50.Kxh5 black's gone
: Serge
I think it's more complicated that that... I don't think h5 is the
right way to go, that pawn keeps white hemmed in on the kingside and
We can't find any way for the white king to escape the checks by the
queen, because the kingside is so open, and white's pieces are out to
lunch on the other side of the board... Ra1 seemed to be the best try
but white has to work to show losing the bishop still leaves a win...
Dave Gomboc
In article <5klcav$i...@proxye1.san.rr.com>,
Steven C. Den Beste <sden...@san.rr.com> wrote:
>
>There's an online chess column associated with the NY Times which has an
>analysis of the game at:
>
>http://www.nytimes.com/library/cyber/week/050597chess.html
>
>Regarding this, he says:
>
>> A desperate attempt at a perpetual check with 45 . . . Qe3 would
>> fail after 46 Qd6 Qc1 47 Ke2 Qb2 48 Kf3 Qc3 49 Kg4.
Unfortunately this totally fails to consider 45...Qe3 46.Qd6 Re8!
Has anyone left their Pentium Pro 200 looking at this position since
the game ended? I am really curious to know if Deep Blue had seen a
win or if it had made a mistake.
Dave Gomboc
drgo...@a.stu.athabascau.ca
Don C. Aldrich
On Mon, 05 May 1997 20:55:28 -0400, Bill Doyle <bdo...@clark.net>
wrote:
>You weren't paying attention:
>
>the line is: Qe3 Qxd6 Re8! not Qxe4?? this is key.
>
>This is still a draw 24 hours and much analysis later.
>
>Cheers
yes, it appears so....
eric fitch
Serge Desmarais wrote:
>
> Steven C. Den Beste wrote:
> >
> > Bill Doyle wrote:
> > >
> > > Kasparov apparently missed a draw with:
> > >
> > > Qe3 Qxd6 Re8!
> > >
> > > After many hours of analysis by carbon and silicon,
> > > there is no win for Deep Blue!
> > >
> > > Wonder how many times this has happened to Garry?
> >
> > Instead of taking the bishop, what if White uses its queen to check the
> > black King?
> >
> > Since the board had simplified, one must assume that DB had a very long
> > look into the future, since the tree branching was no longer as wicked;
> > surely it could have maintained the pressure on GK to prevent a
> > repetitive check leading to a draw.
> >
> > I know that bishop looks ripe for the plucking, but it'll still be
> > there, and taking it loses incentive. DB proved in this game that it was
> > more subtle than that.
> >
> > ...but then I'm not a very good chess player.
> >
> > --
> > sden...@san.rr.com "Cheap, good, soon - pick any two."
> > sden...@qualcomm.com
>
> Here is GM Gausel variation in the post-mortem comments :
>
> 45...Qe3 46.Qxd6 Qc1+ 47.Ke2 Qb2+ 48.Kf3 Qxc3+ 49.Kg4 h5+
> 50.Kxh5 black's gone
>
> Serge
Many strong players are agreeing about this.
Herbert Groot Jebbink
On Tue, 06 May 1997 00:35:02, mcl...@prima.ruhr.de (mclane) wrote:
>Ah - come one. You are all waisting time. He has lost this game on
>purpose and you will find more things he had missed in this game.
On Mon, 05 May 1997 17:39:01 qb...@xs4all.nl (Qbert) wrote:
>Now, if this encounter were just about CHESS, the final score would
>indeed be 6-0 or 5.5-0.5 at its best. But since this match is not so
>much about chess as it is about IBM and Garry Kasparov, I suppose it
>will end up with DB having scored slightly better than it did last time.
On Sun, 04 May 1997 17:23:58, Tim Swetonic <ti...@starwave.com> wrote:
>This is merely a question. Does anyone think there's any possibility
>that Kasparov would throw a match, or draw a match or two on purpose to
>make Deep Blue look good? He's getting a lot of money ...
yes yes !!!, I was waiting for this kind of postings, this is fun !!!
Groeten, Herbert
---
The Trans-Siberian Railroad Page, http://www.xs4all.nl/~hgj/
Francesco Di Tolla
Robert Hyatt wrote:
>
> : 45...Qe3 46.Qxd6 Qc1+ 47.Ke2 Qb2+ 48.Kf3 Qxc3+ 49.Kg4 h5+
> : 50.Kxh5 black's gone
> I think it's more complicated that that... I don't think h5 is the
> right way to go
what about 49....Qe3 again?
Qe3 could have been a draw or not. It was at least worth trying.
bye
Franz
--
Francesco Di Tolla, Center for Atomic-scale Materials Physics
Physics Department, Build. 307, Technical University of Denmark,
DK-2800 Lyngby, Denmark, Tel.: (+45) 4525 3208 Fax: (+45) 4593 2399
mailto:dit...@fysik.dtu.dk http://www.fysik.dtu.dk/persons/ditolla.html
chrisw
--
http://www.demon.co.uk/oxford-soft
mclane <mcl...@prima.ruhr.de> wrote in article
<5kluem$fn0$1...@steve.prima.ruhr.de>...
> massimil...@iol.it (Massimiliano Orsi) wrote:
>
>
> >Did Kasparov miss something or is there something really bad in the
> >final position?
>
> Ah - come one. You are all waisting time. He has lost this game on
> purpose and you will find more things he had missed in this game.
>
OK, so its possible to put forward the conspiracy theory as to win/loss.
Unfortunately we can never know.
Nevertheless, whether Qe3 draws or not is of interest.
CSTal is still failing high after 17 ply on Qe3 - so there's obviously
something there.
Chris Whittington
Howard Exner
Bill Doyle <bdo...@clark.net> wrote in article <336D5B...@clark.net>...
> Kasparov apparently missed a draw with:
>
> Qe3 Qxd6 Re8!
>
> After many hours of analysis by carbon and silicon,
> there is no win for Deep Blue!
>
> Wonder how many times this has happened to Garry?
Here is one of many bizarre lines that support your claim.
43... Qe3 44. Qxd6 Re8 45. h4 Qxe4 46. Ra7+ Kg8 47. Qd7 Qf4+ 48. Kg1 Qe3+ 49.
Kh2 Qf4+ 50. Kh3 Re7! 51. Qc8+ Kh7 52. Rxe7 h5 Still looks like a draw.
Some were suggesting a timely h4 for white as winning but I still don't see it. Henley
and Seirawan are not providing too much analysis on this Qe3
attempt. Perhaps they could show some deeper analysis? Curious to see what Kasparov or
the
Deep Blue team will offer.
chrisw
--
http://www.demon.co.uk/oxford-soft
eric fitch <eric...@sprynet.com> wrote in article
<336E8C...@sprynet.com>...
> The more that strong players look at this line, they seem to think it's
> a draw. This from IM Malcom Pein on the electric minds forum:
>
>
> "Hi everyone, I am writing from the web room at the
> Equitable. The more I look at this position the more I
> think its a draw. It was actually GM Jon Tisdall who
> brough this to my attention. I have checked with Joel
> Benjamin and he said he thought it was a draw but he
> did point out that maybe DB could play Kh1 instead of
> Kf1. Frederic Friedel disagreed at first so I told him
> to switch Fritz on again and lo and behold he reckons
> its a draw as well. I think Garry is going to flip when
> he finds out and Fred was not looking forward to
> telling him !!
> The interesting thing is that everyone was psyched out
> by DB and assumed Qe3 lost because if it did not then
> DB would not allow it !!
> It seems to me we have had a case of mass hypnosis /
> delusion that took in 1000s and 1000s of players.
> Deep Blue did play a fantastic positional game the
> funny thing is its flaw here was that its lookahead
> when it played Kf1 did not extend far enough down Qe3.
> It still played a beautiful Ruy Lopez as far as I am
> concerned.
>
> Anyway, its hold the front page on The Daily Telegraph
> in the UK.
And he did hit the front page ! Well done Malcolm.
Actually there were two stories.
1. was Malcolm Pein's column which said DB won, gave the listing and some
analysis - this must have been posted off as soon as the game was over.
2. was the front page article, saying Garry was asleep, and wasn't going to
be happy when he woke up to the draw story. Pein says it was a definite
draw, and credits Ferret with posing it one hour after the match was over.
Chris Whittington
>
> Cheers
>
> Malcolm
>
chrisw
Dave Gomboc <drgo...@a.stu.athabascau.ca> wrote in article
<5km5b8$g...@aurora.cs.athabascau.ca>...
I left CStal on a PPro 200 for 36 hours and then got fed up and switched
off.
After 17 ply, Qe3 was failing high at -1.87
This implies that there was something deep, quite likely a draw.
At the time of posting there are precious few (if any) posts from
programmers/operators giving a best line and score for this move. I assume
from that that none of the prgs can find a definite draw from this
position, but that none of the prgs find a a definite loss either.
This one is complicated - no doubt there'll be much analysis and comment
about the game on the servers come the next round.
Chris Whittington
Chris Whittington
>
> Dave Gomboc
> drgo...@a.stu.athabascau.ca
>
>
go...@hbar.phys.msu.su
In drgo...@a.stu.athabascau.ca (Dave Gomboc) writes:
>Unfortunately this totally fails to consider 45...Qe3 46.Qd6 Re8!
>
>Has anyone left their Pentium Pro 200 looking at this position since
>the game ended? I am really curious to know if Deep Blue had seen a
>win or if it had made a mistake.
>
>Dave Gomboc
>drgo...@a.stu.athabascau.ca
It looks that 45...Qe3 46.Qd6 Re8! is really a draw but what about 45...Qe3 46.Qd7+
for example 46... Be7 47.Qe6+ Kf8 48. d6 or 46... Kg8 47.Qxd6 Re8 48.Qd7 and Ra7
Ernst A. Heinz
Chris wrote:
> At the time of posting there are precious few (if any) posts from
> programmers/operators giving a best line and score for this move. I assume
> from that that none of the prgs can find a definite draw from this
> position, but that none of the prgs find a a definite loss either.
DarkThought/AEGON'97 locks onto 45.. Qe3! after 8:47 minutes in iteration 14
with a score of -1.57 pawns. Therewith it stayed until interruption after
quite a while in iteration 18.
Because there are enough variations where White can play quiet moves and/or
give back material to avoid/delay the perpetual, I strongly doubt that any
existing chess machine (Cray Blitz, Deep[er] Blue, *Socrates etc. included)
will return an *exact* draw score in this position ...
=Ernst=
Workers World Chicago bureau
chrisw (chr...@cpsoft.demon.co.uk) wrote:
.....
: At the time of posting there are precious few (if any) posts from
: programmers/operators giving a best line and score for this move. I assume
: from that that none of the prgs can find a definite draw from this
: position, but that none of the prgs find a a definite loss either.
: This one is complicated - no doubt there'll be much analysis and comment
: about the game on the servers come the next round.
The latest word on this from the ICC analysis team (Ferret, Schroer,
Securitron, et. al.) has been posted at www.tcc.net/twic/twic.html ..
various lines through move 60, and still a draw
Jouni Uski
Bill Doyle wrote:
>
> Kasparov apparently missed a draw with:
>
> Qe3 Qxd6 Re8!
>
> After many hours of analysis by carbon and silicon,
> there is no win for Deep Blue!
>
> Wonder how many times this has happened to Garry?
All my programs play next h4 !? giving white king one more field in h3.
Then I don't see any forced draw for black, but position is very
difficult...
Jouni Uski
Robert Hyatt
Steven C. Den Beste ("sdenbes1@NOSPAM"@san.rr.com) wrote:
: Bill Doyle wrote:
: >
: > Kasparov apparently missed a draw with:
: >
: > Qe3 Qxd6 Re8!
: >
: > After many hours of analysis by carbon and silicon,
: > there is no win for Deep Blue!
: >
: > Wonder how many times this has happened to Garry?
: Instead of taking the bishop, what if White uses its queen to check the
: black King?
: Since the board had simplified, one must assume that DB had a very long
: look into the future, since the tree branching was no longer as wicked;
: surely it could have maintained the pressure on GK to prevent a
: repetitive check leading to a draw.
: I know that bishop looks ripe for the plucking, but it'll still be
: there, and taking it loses incentive. DB proved in this game that it was
: more subtle than that.
: ...but then I'm not a very good chess player.
doesn't appear to help. If you check, the king retreats to g8, and there are
no good ways to continue the attack. the black rook guards key squares, the
bishop you refused to take guards others, so white has to play a quiet move
to threaten something, by either taking the bishop, or attacking the rook, or
bringing the rook into play. But *any* non-check opens the door for black.
White's only chance is to get the rook back to help defend against the
checks, but then white's bishop is hanging... difficult position...
Amir Ban
It is a draw. I guess by now several analyzers came to that conclusion.
I guess the analysis of the perpetual check is hard for computers
because while the white king marches back and forth, the black queen
must capture f5, so the transposition tables don't discover the
repetition so soon.
The critical variations are:
45 ... Qe3 46. Qxd6 Re8 or
46. Qd7+ Kg8 47. Qxd6 Rf8
The point is that black puts the rook in a safe position and defends
against immediate mate. On the other hand it is threatening to take the
bishop and the same time perpetual check. The white king cannot hide at
h2 by g3 because black just takes the bishop on e4 or f3 and again
threatens perpetual check (this tempo is important. It would not work
if white has immediate mate threat on the other side).
The only try for white is (46. Qxd6 Re8) 47. h4
The plan is for the white king to escape at Kh3 shielded by pawns at h4
and g3. So Qxe4 is forbidden because of mate on the 7th rank. But black
should play
47... h5 !
And there is no escape from perpetual.
This is almost beautiful enough to be a study.
We still have no answer why 36.Qb6 was not played. This is a complete
mystery. I have sources that say that after the game Kasparov himself
was asking this.
Amir
Robert Hyatt
Ernst A. Heinz (hei...@ira.uka.de) wrote:
: Chris wrote:
: > At the time of posting there are precious few (if any) posts from
: > programmers/operators giving a best line and score for this move. I assume
: > from that that none of the prgs can find a definite draw from this
: > position, but that none of the prgs find a a definite loss either.
: DarkThought/AEGON'97 locks onto 45.. Qe3! after 8:47 minutes in iteration 14
: with a score of -1.57 pawns. Therewith it stayed until interruption after
: quite a while in iteration 18.
: Because there are enough variations where White can play quiet moves and/or
: give back material to avoid/delay the perpetual, I strongly doubt that any
: existing chess machine (Cray Blitz, Deep[er] Blue, *Socrates etc. included)
: will return an *exact* draw score in this position ...
: =Ernst=
I'm surprised that DB overlooked this, although I'd bet it can find the
Qe3 move very quickly, because when we studied this in detail, there are
so many positions where there is exactly *one* decent move to avoid a
draw or avoid a loss, or to follow a draw... and singular extensions should
lock into the right line quickly. Were I a betting man, and if this is really
a draw, I'd bet DB's eval was 0.000 at the end. Only a gut feeling, but it
is not that deep, in light of their search...
miK
ww...@wwa.com (Workers World Chicago bureau) wrote:
>chrisw (chr...@cpsoft.demon.co.uk) wrote:
>.....
>: At the time of posting there are precious few (if any) posts from
>: programmers/operators giving a best line and score for this move. I assume
>: from that that none of the prgs can find a definite draw from this
>: position, but that none of the prgs find a a definite loss either.
>: This one is complicated - no doubt there'll be much analysis and comment
>: about the game on the servers come the next round.
>The latest word on this from the ICC analysis team (Ferret, Schroer,
>Securitron, et. al.) has been posted at www.tcc.net/twic/twic.html ..
>various lines through move 60, and still a draw
Even the DBteam seems to agree...
*Murray Campbell of the Deep Blue team just acknowledged on stage in
the auditorium that the second game is a draw after 45...Qe3.* (inside
chess page)
A bit odd, since Garry surely was not under time pressure when he
resigned. I wonder what he was thinking...
Ernst A. Heinz
After hearing of many microcomputer programs that they find 45.. Qe3!
I am interested in what other programs play for White at move 44:
DarkThought/AEGON'97 instantly locks onto 44. Kh1 and later switches to
44. Kh2 in iteration 12 (which hopefully also wins) -- but it never ever
considers Deeper Blue's choice 44. Kf1? to be the best move :-)
What about our fellow micros out there?
=Ernst=
eric fitch
Jouni Uski wrote:
>
> Bill Doyle wrote:
> >
> > Kasparov apparently missed a draw with:
> >
> > Qe3 Qxd6 Re8!
> >
> > After many hours of analysis by carbon and silicon,
> > there is no win for Deep Blue!
> >
> > Wonder how many times this has happened to Garry?
>
> Then I don't see any forced draw for black, but position is very
> difficult...
>
> Jouni Uski
h4 also draws after blacks h5! Kasparov himself has looked at this
continuation.
eric fitch
part, owes Re8! to "Tim" , a strong ICC player who first suggested it.
The ananlysis went on until around 3:00 a.m. CST.
Ilias Kastanas
In article <336E48...@sprynet.com>,
eric fitch <eric...@sprynet.com> wrote:
>brucemo wrote:
>>
>> Kjell Tore Sandum wrote:
>> >
>> > In my opinion, Doyle is correct: 45...Qe3! is a draw.
>> >
>> > After 46.Qd7+ Kg8 47.Qxd6 the simple 47...Rf8 leaves white with the same
>> > problems of avoiding perpetual check. The main difference is that now two
>> > more plies are added, and the draw is probably behind most computers
>> > horizon.
>>
>> If we missed a win for DB, it's not in this variation, I think. 46. Qxd6
>> seems harder.
>>
>> bruce
>
>This is Seirawa's ananlysis at the end:
>
> 45.Ra6!
>
>"Now the auditorium's audience was a buzz. For some time the
>commentators, IM Maurice Ashley and IM Michael Valvo were
>quickly coming to the conclusion that 45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8
>48.Bd5+ and Bd5-e6 was lost for Kasparov. The
>desperate attempt 45...Qe3 46.Qxd6 Qxe4 (46...Qf4+ 47.Bf3; 46...Qc1+
>47.Ke2 Qb2+ 48.Kd1 stops the perpetual)
>47.Ra7+ Kg8 48.Qxb8+ picks up Black's Rook with check. What was Garry's
>Defense? He had none and extended his hand
>in resignation. The applause rocked the auditorium! Folks we have a
>match! A superb game by Deep Blue and its team, my
>hearty congratulations for a brilliant effort!"
>
> Still he doesn't look into 45...Qe3 46.Qxd6 Re8. This is the only
>move for black here and one that, as far as I can see, leads to a draw.
It does seem a draw. W can try to disrupt the perpetual via 47. h4.
Tempting is 47... Qf4+ 48. Bf3, Qc1+ 49. Kf2, Qd2+ 50. Kg3, Qe1+ 51. Kg4,
h5+ drawing, since 52. K:h5?? is mate in two. But W can improve with
50. Be2; the K reaches h3, and winning the B throws Bl off track.
Posted analysis prefers 47... Q:e4 48. Ra7+, Kg8 49. Qd7, Qf4+ ...
and when Kh2, Qf4+. Kh3 not ... Qe3+? but a miracle, ... Re7! Qc8+, Kh7.
Ra8, Re8! and then ... Q:f5+ will draw.
Still, the simplest is 47. ... h5, shutting the door; the WK now
finds no relief at h3.
Ilias
Jouni Uski
Probably -Qe3! is draw, but it doesn't actually
matter, because Gary resigned and lost.
Result 1 - 0.
Jouni
Robert Hyatt
chrisw (chr...@cpsoft.demon.co.uk) wrote:
: Dave Gomboc <drgo...@a.stu.athabascau.ca> wrote in article
: <5km5b8$g...@aurora.cs.athabascau.ca>...
: > In article <5klcav$i...@proxye1.san.rr.com>,
: > Steven C. Den Beste <sden...@san.rr.com> wrote:
: > >
: > >There's an online chess column associated with the NY Times which has an
: > >analysis of the game at:
: > >
: > >http://www.nytimes.com/library/cyber/week/050597chess.html
: > >
: > >Regarding this, he says:
: > >
: > >> A desperate attempt at a perpetual check with 45 . . . Qe3 would
: > >> fail after 46 Qd6 Qc1 47 Ke2 Qb2 48 Kf3 Qc3 49 Kg4.
: >
: > Unfortunately this totally fails to consider 45...Qe3 46.Qd6 Re8!
: >
: > Has anyone left their Pentium Pro 200 looking at this position since
: > the game ended? I am really curious to know if Deep Blue had seen a
: > win or if it had made a mistake.
: I left CStal on a PPro 200 for 36 hours and then got fed up and switched
: off.
: After 17 ply, Qe3 was failing high at -1.87
: This implies that there was something deep, quite likely a draw.
: At the time of posting there are precious few (if any) posts from
: programmers/operators giving a best line and score for this move. I assume
: from that that none of the prgs can find a definite draw from this
: position, but that none of the prgs find a a definite loss either.
: This one is complicated - no doubt there'll be much analysis and comment
: about the game on the servers come the next round.
: Chris Whittington
What we did, on ICC, was to examine the game *after* Qe3, to see if we could
find any win for the computer. From this position, we tried *every* move by
white that looked reasonable, and Crafty (to the best of my knowledge) found a
repetition in every move. I recall trying 5 different moves, which were (I
believe) Bf3, Q check, Qxd6, Ra1, and h4 (don't know why we tried h4 but we
did.) The problem is, that from this point back, it needs two more plies and
Crafty isn't going to get those in any reasonable time control. DB should be
able to see this instantly with their deep search + singular extensions. I'm
*highly* interested in seeing their analysis of course. And it might be that
there is some other "saving" move for white, but the above moves don't seem to
have an answer...
: Chris Whittington
: >
: > Dave Gomboc
: > drgo...@a.stu.athabascau.ca
: >
: >
Keith Reynolds
How about....after 45. ...Qe3 46. Qxd6 Re8 instead of Bf3 White tries
something like 47. h4 to create a hiding place for his King? I can't find a
save for Black....any ideas?
Massimiliano Orsi <massimil...@iol.it> wrote in article
<5kl214$a...@mikasa.iol.it>...
> Bill Doyle <bdo...@clark.net> wrote:
>
> >Kasparov apparently missed a draw with:
> >Qe3 Qxd6 Re8!
> >After many hours of analysis by carbon and silicon,
> >there is no win for Deep Blue!
>
> As a matter of fact, after 45. ..Qe3 46. Qxd6 Re8 47. Bf3 Qc1 48. Kf2
> Qd2 49. Be2 Qf4 50. Ke1 Qc1 51. Bd1 Qxc3 (also 51. ..Qe3 52. Kf1 and
> now 52. ..Qc1 53. Ke2 Qb2 54. Kf3 Qxc3 should draw; but not 52. ..Qf4?
> 53. Kg1 Qe3 54. Kh1! Qe1 55. Kh2 Qxe1 56. Ra7 and White mates) 52.
> Kf1 Qc1 White can't improve his position.
>
> White can play 46. Qd7 (instead of 46. Qxd6) and now 46. ..Be7
> loses after 47. d6, but 46. ..Kg8 47. Qxd6 Rf8 poses the same problem
> of White King's safety.
>
> I think only White's chance is playing something like 48. Qb6 or 48.
> Ra2, even after the immediate 46. Qxd6; in both cases White's
> resignation would not have been on my agenda.
jlu...@hoflink.com
On Mon, 5 May 1997, eric fitch wrote:
[some interesting analysis deleted]
.
> There was a rumor that DB had evaluated the final position as +2 (i.e.,
> equivalent to a 2-Pawn, or winning, advantage). Ferret said something to
> the effect that no one should resign to a computer which thinks it's at
> +2
> in this kind of position. That's because of the evaluation error
> potential
> in lines where a quiet move (46...Re8) leads to positions with possible
> perpetual checks. Again, the report of DB's evaluation is just a rumor.
.
I can confirm the rumor and add more context. After the game, the Deep
Blue team spoke to the spectators in attendance. In response to a
question about Deep Blue's final evaluation, one of the team members said
it was +2.
.
I ran into GM Joel Benjamin outside the playing site yesterday, and he
said that Deep Blue's evaluation at the time Kasparov resigned was +1.50.
.
Today we should get a lot more information about this amazing controversy.
.
[posted & mailed]
.
Jason Luchan
Keith Reynolds
Seems my previous post didn't make it.
Bill Doyle
> : > programmers/operators giving a best line and score for this move. I assume
> : > from that that none of the prgs can find a definite draw from this
> : > position, but that none of the prgs find a a definite loss either.
>
> : with a score of -1.57 pawns. Therewith it stayed until interruption after
> : quite a while in iteration 18.
>
> : Because there are enough variations where White can play quiet moves and/or
> : give back material to avoid/delay the perpetual, I strongly doubt that any
> : existing chess machine (Cray Blitz, Deep[er] Blue, *Socrates etc. included)
> : will return an *exact* draw score in this position ...
>
> : =Ernst=
>
> I'm surprised that DB overlooked this, although I'd bet it can find the
> Qe3 move very quickly, because when we studied this in detail, there are
> so many positions where there is exactly *one* decent move to avoid a
> draw or avoid a loss, or to follow a draw... and singular extensions should
> lock into the right line quickly. Were I a betting man, and if this is really
> a draw, I'd bet DB's eval was 0.000 at the end. Only a gut feeling, but it
> is not that deep, in light of their search...
From the IBM web site Joel Benjamin says DB's eval was +1.5.
Joel also adds that is was a draw.
Apparently DB didn't see the perpertual check because of the
number of squares the King can move to.
Also there are a number of irrelavent pawn captures that can
be made.
Kasparov's and DB's camp both acknowledge it was a draw.
Cheers,
Bill Doyle
Steffen A. Jakob
Robert Hyatt wrote:
[...]
> I'm surprised that DB overlooked this, although I'd bet it can find the
> Qe3 move very quickly, because when we studied this in detail, there are
> so many positions where there is exactly *one* decent move to avoid a
> draw or avoid a loss, or to follow a draw... and singular extensions should
> lock into the right line quickly. Were I a betting man, and if this is really
> a draw, I'd bet DB's eval was 0.000 at the end. Only a gut feeling, but it
> is not that deep, in light of their search...
So if we know in some months fore sure that the position was a draw,
this would be a great test position for computers! :-)
Best wishes,
Yobes.
--
Steffen A. Jakob |"Die Ratten versenken das
s...@gams.net | stinkende Schiff."
+43 1 8176230-18 | (Regina Klotz)
http://www.gams.net/~saj/ |
Ed Schroder
Rebel, Kh2 for hours...
Don't know if this is better than the suggested Kh1
- Ed Schroder -
>
>=Ernst=
Bill Doyle
> > Qd2 49. Be2 Qf4 50. Ke1 Qc1 51. Bd1 Qxc3 (also 51. ..Qe3 52. Kf1 and
> > now 52. ..Qc1 53. Ke2 Qb2 54. Kf3 Qxc3 should draw; but not 52. ..Qf4?
> > 53. Kg1 Qe3 54. Kh1! Qe1 55. Kh2 Qxe1 56. Ra7 and White mates) 52.
> > Kf1 Qc1 White can't improve his position.
> >
> > White can play 46. Qd7 (instead of 46. Qxd6) and now 46. ..Be7
> > loses after 47. d6, but 46. ..Kg8 47. Qxd6 Rf8 poses the same problem
> > of White King's safety.
> >
> > I think only White's chance is playing something like 48. Qb6 or 48.
> > Ra2, even after the immediate 46. Qxd6; in both cases White's
> > resignation would not have been on my agenda.
> >
> > Did Kasparov miss something or is there something really bad in the
> > final position?
> >
> >
> >
45. ... Qe3 46. Qxd6 Re8 47. h4! h5! also draws.
Latest word from GM Kaidanov: game 2 may not have been a draw, but
he gives no analysis.
Cheers,
Bill Doyle
Serge Desmarais
Robert Hyatt wrote:
>
> Serge Desmarais (psy...@total.net) wrote:
> : Steven C. Den Beste wrote:
> : >
> : > Bill Doyle wrote:
> : > >
> : > > Kasparov apparently missed a draw with:
> : > >
> : > > Qe3 Qxd6 Re8!
> : > >
> : > > After many hours of analysis by carbon and silicon,
> : > > there is no win for Deep Blue!
> : > > Wonder how many times this has happened to Garry?
> : >
> : > Instead of taking the bishop, what if White uses its queen to check the
> : > black King?
> : >
> : > Since the board had simplified, one must assume that DB had a very long
> : > look into the future, since the tree branching was no longer as wicked;
> : > surely it could have maintained the pressure on GK to prevent a
> : > repetitive check leading to a draw.
> : >
> : > I know that bishop looks ripe for the plucking, but it'll still be
> : > there, and taking it loses incentive. DB proved in this game that it was
> : > more subtle than that.
> : >
> : > ...but then I'm not a very good chess player.
> : > --
> : > Steven C. Den Beste
> : > sden...@san.rr.com "Cheap, good, soon - pick any two."
> : > sden...@qualcomm.com
>
> : Here is GM Gausel variation in the post-mortem comments :
>
> : 45...Qe3 46.Qxd6 Qc1+ 47.Ke2 Qb2+ 48.Kf3 Qxc3+ 49.Kg4 h5+
> : 50.Kxh5 black's gone
>
> : Serge
>
> I think it's more complicated that that... I don't think h5 is the
> right way to go, that pawn keeps white hemmed in on the kingside and
> We can't find any way for the white king to escape the checks by the
> queen, because the kingside is so open, and white's pieces are out to
> lunch on the other side of the board... Ra1 seemed to be the best try
> but white has to work to show losing the bishop still leaves a win...
I was wrong! I put too much confidence in what the GMs and IMs were
commenting on the game... So, when they all stated that the game was lost
for Black, I took their advice with no critics and a lot of naivety...
Hehehe. I must be cautious about superficial opinions!
Serge
Robert Hyatt
Serge Desmarais (psy...@total.net) wrote:
: Serge
Remember that Kasparov also thought it was lost. Now everyone agrees it
was a dead draw.
Robert Hyatt
Keith Reynolds (kr...@toolcity.net) wrote:
: How about....after 45. ...Qe3 46. Qxd6 Re8 instead of Bf3 White tries
: something like 47. h4 to create a hiding place for his King? I can't find a
: save for Black....any ideas?
h5... then g4 is safe for the black queen and the black king then
can't hide at h3...
: Massimiliano Orsi <massimil...@iol.it> wrote in article
: <5kl214$a...@mikasa.iol.it>...
: > Bill Doyle <bdo...@clark.net> wrote:
: >
: > >Kasparov apparently missed a draw with:
: > >Qe3 Qxd6 Re8!
: > >After many hours of analysis by carbon and silicon,
: > >there is no win for Deep Blue!
: >
: > As a matter of fact, after 45. ..Qe3 46. Qxd6 Re8 47. Bf3 Qc1 48. Kf2
: >
: >
: >
Peter Osterlund
On 6 May 1997, Robert Hyatt wrote:
: I'm surprised that DB overlooked this, although I'd bet it can find the
: Qe3 move very quickly, because when we studied this in detail, there are
: so many positions where there is exactly *one* decent move to avoid a
: draw or avoid a loss, or to follow a draw... and singular extensions should
: lock into the right line quickly. Were I a betting man, and if this is really
: a draw, I'd bet DB's eval was 0.000 at the end. Only a gut feeling, but it
: is not that deep, in light of their search...
Are the DB chess chips able to recognize a draw by repetition? I vaguely
remember reading that the older DT chess chips could not do this. In this
position, it could make a *big* difference, especially if the chess chips
are doing check extensions.
--
Peter Österlund Email: peter.o...@mailbox.swipnet.se
Sköndalsvägen 35 f90...@nada.kth.se
S-128 66 Sköndal Homepage: http://home1.swipnet.se/~w-15919
Sweden Phone: +46 8 942647
Robert Hyatt
Peter Osterlund (peter.o...@mailbox.swipnet.se) wrote:
: On 6 May 1997, Robert Hyatt wrote:
: : I'm surprised that DB overlooked this, although I'd bet it can find the
: : Qe3 move very quickly, because when we studied this in detail, there are
: : so many positions where there is exactly *one* decent move to avoid a
: : draw or avoid a loss, or to follow a draw... and singular extensions should
: : lock into the right line quickly. Were I a betting man, and if this is really
: : a draw, I'd bet DB's eval was 0.000 at the end. Only a gut feeling, but it
: : is not that deep, in light of their search...
: Are the DB chess chips able to recognize a draw by repetition? I vaguely
: remember reading that the older DT chess chips could not do this. In this
: position, it could make a *big* difference, especially if the chess chips
: are doing check extensions.
The DT-II chips did not, although the basic part of the search (first 7 plies
+ extensions as I recall) did. Hsu had reported that this would be fixed in
the DB chips. I believe I've read somewhere where this was fixed. I'd also
bet they knew that game was a draw while it was going on. And that they either
were playing "coy" to avoid giving away clues for just exactly how deep they
are searching, or they didn't want to make a big scene that could be taken as a
"look how dumb Kasparov was, he overlooked a draw that we saw after 30 seconds
of searching..." Perhaps we'll find the real truth after the match...
Al Evans
Hello,
I am looking for a small DOS based chess game that will run on an HP95LX
handheld computer. I have heard there is one that is freeware or
shareware. I think it would help pass the time while waiting in doctors
offices, on airplains etc.
I have not yet been able to find such a chess game. If you know of one,
please send me a note of its where abouts.
Thanks, Jim
Robert Hyatt
Alan Reider (a...@reider.net) wrote:
: >Remember that Kasparov also thought it was lost. Now everyone agrees it
: >was a dead draw.
: Ahem.. not everyone. GM Dzindzhi was saying that it is not dead at
: all, just not a forced win. I dont have a set in front of me but i
: think it was something like, after re8, to play qc5, let black take
: back on e4, then kf1 and black is in for a miserable time, maybe even
: lost. I dont know if theres any analysis on it, this was from
: comments when he was analysing game 3 with yasser and ashley.
: it seems everyone is so transfixed by the tricky tactics by black to
: avoid a forced loss, that it is just assumed to be 'dead' drawn, and
: this positional attempt to continue the struggle ignored. i'd like to
: see this analysed.
Hard to say. I tried this yesterday, playing the Qc5 move and Crafty still
finds a forced draw for black... I posted the analysis for whomever asked
about the Qc5 move...