Silicon Graphics World Checkers Championship
Robert Lake
The Silicon Graphics World Checkers Championship is being held in
Boston at the Computer Museum, August 15-25. The World Champion,
Dr. Marion Tinsley, is putting his title on the line against the
computer program Chinook (Martin Bryant, Robert Lake, Paul Lu,
Jonathan Schaeffer, Norman Treloar). It is a best-of-30-game
match.
On Monday, the opening ceremony was held, with speeches from
Raymond Keene and David Levy (match organizers), Marvin Minsky
(professor in artificial intelligence), Bob Bishop (sponsor,
president of Silicon Graphics International, and Tony Buzan.
Game 1 was a quiet draw, while in game 2 Chinook pressed hard
for victory, forcing Tinsley to find the only saving draw.
Games 3-6 will be played on Tuesday.
BLACK : Chinook
WHITE : Tinsley
EVENT : Match01
1. 09-14 22-18 2. 11-15 18-09 3. 05-14 25-22 4. 15-19 24-15
5. 10-19 23-16 6. 12-19 22-17 7. 06-10 27-24 8. 02-06 24-15
9. 10-19 17-10 10. 07-14 29-25 11. 08-11 25-22 12. 04-08 26-23
13. 19-26 30-23 14. 11-15 31-27 15. 08-11 28-24 16. 06-10 24-20
17. 01-06 22-17 18. 15-18 23-19 19. 18-22 19-16 20. 11-15 16-11
21. 15-19 11-07 22. 14-18 32-28 23. 22-26 17-13 24. 10-15 07-02
25. 19-23 02-09 26. 23-32 09-14 27. 18-23 14-10 28. 15-18 20-16
29. 26-31 16-11 30. 32-27
RESULT: Drawn
BLACK : Tinsley
WHITE : Chinook
EVENT : Match02
DATE : Mon Aug 15 11:36:44 1994
1. 09-14 22-18 2. 11-15 18-11 3. 08-15 25-22 4. 06-09 29-25
5. 01-06 23-18 6. 14-23 27-11 7. 07-16 22-18 8. 03-07 25-22
9. 04-08 24-20 10. 08-11 28-24 11. 09-14 18-09 12. 05-14 26-23
13. 10-15 32-28 14. 07-10 31-26 15. 06-09 30-25 16. 02-06 24-19
17. 15-24 28-19 18. 11-15 20-11 19. 15-24 23-19 20. 24-27 11-07
21. 27-32 26-23 22. 32-28 22-18 23. 28-32 07-03 24. 32-27 03-07
25. 27-31 25-22 26. 31-27 07-11 27. 27-31 11-16 28. 31-27 16-20
29. 09-13 18-02 30. 27-25 02-06 31. 10-14 19-15 32. 14-18 15-11
33. 18-22 06-10 34. 22-26 10-15 35. 25-22 11-07 36. 26-30 07-02
37. 30-26 02-06 38. 26-30 06-10 39. 30-26 10-14 40. 26-30 14-18
41. 30-25 18-23 42. 25-30 20-24 43. 30-25 24-19 44. 25-30 23-18
45. 22-26 15-11 46. 30-25 19-15 47. 26-22 18-23 48. 25-30 11-07
RESULT: Drawn
SCORE: Tinsley: 1.0
Chinook: 1.0
--
Rob Lake la...@cs.ualberta.ca
Department of Computing Science http://web.cs.ualberta.ca/~lake
University of Alberta Office: (403) 492-5817
Edmonton, Alberta, Canada Fax: (403) 492-1071
Jeffrey Ely
>
>
>The Silicon Graphics World Checkers Championship is being held in
>Boston at the Computer Museum, August 15-25. The World Champion,
>Dr. Marion Tinsley, is putting his title on the line against the
>computer program Chinook (Martin Bryant, Robert Lake, Paul Lu,
>Jonathan Schaeffer, Norman Treloar). It is a best-of-30-game
>match.
Can anyone give me directions to the Computer Museum? I will be in
Boston around that time, and I would like to stop by to check this
out. Thanks.
Jeff
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Jeff Ely "The Sun's not yellow...
Department of Economics, UC Berkeley It's Chicken!"
je...@econ.berkeley.edu
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
"...a man of great talent will normally pay less attention to other
people's foolishness than would a fool." -Proust
David Bell
> says:
> >
> >
> >The Silicon Graphics World Checkers Championship is being held in
> >Boston at the Computer Museum, August 15-25. The World Champion,
> >Dr. Marion Tinsley, is putting his title on the line against the
> >computer program Chinook (Martin Bryant, Robert Lake, Paul Lu,
> >Jonathan Schaeffer, Norman Treloar). It is a best-of-30-game
> >match.
>
Why is there a computer playing in the checkers world championship???
surely this should be for PEOPLE only - especially seeing that checkers
has almost been 'solved' by computers.
You won't see a chess world championship with a computer in it! ( I hope :-) )
Robert Lake
>Why is there a computer playing in the checkers world championship???
>surely this should be for PEOPLE only - especially seeing that checkers
>has almost been 'solved' by computers.
>
>You won't see a chess world championship with a computer in it! ( I hope :-) )
There are 500 billion billion (i.e. 5 x 10**20) positions in checkers.
So far I have solved only about 250 billion. Where have you heard that
checkers has almost been 'solved ' by computers?
Kelly Martin
David> Why is there a computer playing in the checkers world
David> championship??? surely this should be for PEOPLE only -
David> especially seeing that checkers has almost been 'solved' by
David> computers.
i once read that checkers had been essentially mastered by humans as
well and that the only way of winning a tournament was to come up with
more 'cooks' than the other guy, which essentially consists of baiting
your opponent into making a mistake. samuel's algorithm is nearly
unbeatable, as i understand it.
David> You won't see a chess world championship with a computer in it!
David> ( I hope :-) )
maybe not yet, but i wouldn't bet on it not happening soon. the best
chess programs are master-rank now.
k.
--
kelly martin GCS/M/O -d+(x) H+ s+:- g+? p? au-- !a w++ v+(*)
<ke...@poverty.bloomington.in.us> C+(++)>$ U>++++$ P+++>$ L 3- E++ N++ K--
!W--- M-- V -po+(--) Y+ t+(+) !5 j R@ G?
interested in astrology or tarot and tv b++>+++ !D B-- e+ u+ h---- f- r+++
live in the bloomington, indiana area? n---- z++++*
send me email for more info!
David Wolfe
Permit me to contribute some information I received from a talk given
by Jon Schaeffer (one of the authors of the checkers program Chinook)
at the recent workshop in combinatorial games.
1) It's clear that the top two players (human or otherwise) are
Tisdale and Chinook. They are both far above the ability of all other
players.
2) Tisdale is happy to compete against a computer for the world
championship. Not only does he find no philosophical problem with it,
but he has the added advantage of finally getting some "decent"
competition (by his standards).
3) The world checkers federation (I don't know its name) does not
agree. Therefore, Tisdale is playing in a Silicon Graphics sponsored
event and is disassociating himself from the largest checkers
organization.
4) Concerning the assertion that checkers has essentially been
mastered by humans: Yes, it is true that in practice strong players
will draw virtually all there games. Consequently, current high level
tournament practice is start from a randomly selected position, P,
which is obtained by making a few random moves from the usual starting
configuration. Games are played in pairs, with the two players
getting the opportunity to play both colors from P.
5) There are many different checkers variants, and the standard
variant played in these tournaments is American. When a checker
reaches the last rank, the check is promoted to a "king", which can
then move backward or forwards one square (or jump to capture). It
may NOT move like a bishop, as in some variants played especially in
Spain, Eastern Europe and Asia.
6) I'm not an authority. I hope I've been vague enought to be
accurate.
David
> From: ke...@poverty.bloomington.in.us (Kelly Martin)
> "David" == David Bell <Da...@sparky2.demon.co.uk> writes:
>
> David> Why is there a computer playing in the checkers world
> David> championship??? surely this should be for PEOPLE only -
> David> especially seeing that checkers has almost been 'solved' by
> David> computers.
>
> i once read that checkers had been essentially mastered by humans as
> well and that the only way of winning a tournament was to come up with
> more 'cooks' than the other guy, which essentially consists of baiting
> your opponent into making a mistake. samuel's algorithm is nearly
> unbeatable, as i understand it.
>
> David> You won't see a chess world championship with a computer in it!
> David> ( I hope :-) )
>
> maybe not yet, but i wouldn't bet on it not happening soon. the best
> chess programs are master-rank now.
> From: la...@cs.ualberta.ca (Robert Lake)
> The Silicon Graphics World Checkers Championship is being held in
> Boston at the Computer Museum, August 15-25. The World Champion,
> Dr. Marion Tinsley, is putting his title on the line against the
> computer program Chinook (Martin Bryant, Robert Lake, Paul Lu,
> Jonathan Schaeffer, Norman Treloar). It is a best-of-30-game
> match.
>
Jeffrey Ely
>i once read that checkers had been essentially mastered by humans as
>well and that the only way of winning a tournament was to come up with
>more 'cooks' than the other guy, which essentially consists of baiting
>your opponent into making a mistake. samuel's algorithm is nearly
>unbeatable, as i understand it.
Let me ask you: if as you say checkers has been mastered by humans,
how would such a human ever make the type of mistake you mention??
>maybe not yet, but i wouldn't bet on it not happening soon. the best
>chess programs are master-rank now.
So like, you don't know much about computers and chess do you?
>interested in astrology or tarot and tv b++>+++ !D B-- e+ u+ h---- f- r+++
>live in the bloomington, indiana area? n---- z++++*
>send me email for more info!
hmmmmm.....
Toby J. Sommer
>Permit me to contribute some information I received from a talk given
>by Jon Schaeffer (one of the authors of the checkers program Chinook)
>at the recent workshop in combinatorial games.
[...]
>4) Concerning the assertion that checkers has essentially been
>mastered by humans: Yes, it is true that in practice strong players
>will draw virtually all there games. Consequently, current high level
>tournament practice is start from a randomly selected position, P,
>which is obtained by making a few random moves from the usual starting
>configuration. Games are played in pairs, with the two players
>getting the opportunity to play both colors from P.
I read about Tinsley in the (printed) newspaper: He has lost but a
single match since 1955 (repeat: 19_55_), and that on a technicality
(tie breaker ruling). And so on ... An interesting personality.
Can someone please post a quick and simple explanation the notation used
in the posted checkers games? E.g., 1. 10-15 22-18 or 1. 09-14 22-18
etc.. And how do you determine what the "randomly selected position,
P," is?
(I guess I should go back and look to see if pairs of games begin
identically to position P and then diverge from move x onward.)
Thank you.
Dennis Breuker
>Permit me to contribute some information I received from a talk given
>by Jon Schaeffer (one of the authors of the checkers program Chinook)
>at the recent workshop in combinatorial games.
>
>1) It's clear that the top two players (human or otherwise) are
>Tisdale and Chinook. They are both far above the ability of all other
^^^^^^^
Tinsley, you mean.
>players.
[snip snip]
Dennis.
Johannes Fuernkranz
David Bell <da...@sparky2.demon.co.uk> wrote:
>
>Why is there a computer playing in the checkers world championship???
>surely this should be for PEOPLE only - especially seeing that checkers
>has almost been 'solved' by computers.
Is that so?
>
>You won't see a chess world championship with a computer in it! ( I hope :-) )
Why not?
Juffi
engelkes
>In article <32slo3$g...@infosrv.edvz.univie.ac.at>,
>Johannes Fuernkranz <ju...@ai.univie.ac.at> wrote:
>>In article <777066...@sparky2.demon.co.uk>,
>>David Bell <da...@sparky2.demon.co.uk> wrote:
>>>
>>>Why is there a computer playing in the checkers world championship???
>>Why not?
>>>You won't see a chess world championship with a computer in it! ( I hope :-)
>>Why not?
> While this isn't exactly in the spirit of this post...
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
>-Regards, Mark
There is one big difference: all the machines you name are BETTER than
humans, while every chess program can be beaten by grandmasters. I hope
Old Tinsley will prove that for checkers too.
--
W.E. Engelkes
Spaarnrijkstraat 26
2024 EK Haarlem (NL)
Tel. 023-259279
Hal Bogner
>In article <32slo3$g...@infosrv.edvz.univie.ac.at>,
>Johannes Fuernkranz <ju...@ai.univie.ac.at> wrote:
>>In article <777066...@sparky2.demon.co.uk>,
>>David Bell <da...@sparky2.demon.co.uk> wrote:
>>>
>>>Why is there a computer playing in the checkers world championship???
>>Why not?
>>>You won't see a chess world championship with a computer in it! ( I hope :-)
>>Why not?
>
> While this isn't exactly in the spirit of this post...
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
>
>-Regards, Mark
Perhaps this will be obvious to you upon reading your own words again, but it
I suppose that if the world's fastest dragster was very nearly as fast in the
100 yard dash as the world's fastest runners, then it *would* be interesting,
too.
-hal
an obsolete chess master,
who also lost three of his first four games to a hand-held checkers computer
at five seconds per move ;-(
Johannes Fuernkranz
Mark A Zabel <mza...@magnus.acs.ohio-state.edu> wrote:
>In article <32slo3$g...@infosrv.edvz.univie.ac.at>,
>Johannes Fuernkranz <ju...@ai.univie.ac.at> wrote:
>>In article <777066...@sparky2.demon.co.uk>,
>>David Bell <da...@sparky2.demon.co.uk> wrote:
>>>
>>>Why is there a computer playing in the checkers world championship???
>>Why not?
>>>You won't see a chess world championship with a computer in it! ( I hope :-)
>>Why not?
>
> While this isn't exactly in the spirit of this post...
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
>
>-Regards, Mark
Agreed.
But until it is clearly established that the human World Chess Champion
doesn't have a chance against the best computer(s) I'll look forward to a
match between the two. It might still be very interesting thereafter.
You don't get much out of watching a hydraulic lift in a weightlifting.
It can be very exciting though to watch somebody/something beat Kasparov
6 games in a row. Maybe we could even learn something about chess.
Juffi
Marc ROGER
in rec.games.chess:
> Can someone please post a quick and simple explanation the notation used
> in the posted checkers games? E.g., 1. 10-15 22-18 or 1. 09-14 22-18
White squares are numbered from bottom to top and left to right, notation
indicates start-final destination.
-- Marc.
Robert Williams
>>>Why is there a computer playing in the checkers world championship???
>>Why not?
>>>You won't see a chess world championship with a computer in it! ( I hope :-)
>>Why not?
>
> While this isn't exactly in the spirit of this post...
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
>
> -Regards, Mark
I think that if you look back at the beginning of the introduction of
machines, you'll see many contests between man and machine... or between
horse and machine. (e.g. John Henry)... it was only after machines far
outstripped man that the competition ended... I think when you see a
computer beat the world champion 13 - 0, it will go back to a human only
sport... Right now, it's still interesting...
skip
Jonathan Schaeffer
>"David" == David Bell <Da...@sparky2.demon.co.uk> writes:
>David> Why is there a computer playing in the checkers world
>David> championship??? surely this should be for PEOPLE only -
>David> especially seeing that checkers has almost been 'solved' by
>David> computers.
>i once read that checkers had been essentially mastered by humans as
>well and that the only way of winning a tournament was to come up with
>more 'cooks' than the other guy, which essentially consists of baiting
>your opponent into making a mistake. samuel's algorithm is nearly
>unbeatable, as i understand it.
Samuels algorithm was "good" for 1962, but is insufficient for 1992.
The program did The program had limited search depth (5 ply?) and a
simpole evaluation function. Its reputation rests on one game
that it won against a "master" player. This single game has distorted
a correct perspective on the program.
Jonathan Schaeffer
>I read about Tinsley in the (printed) newspaper: He has lost but a
>single match since 1955 (repeat: 19_55_), and that on a technicality
>(tie breaker ruling). And so on ... An interesting personality.
Since 1950, Tinsley has won every match played for the World Championship,
usually by a large margin. Until this year he had placed first
in every tournament he competed in (in the US Oopen earlier this
year, there was a 3-way tie for first, with Tinsley placing 3rd on
tiebreak).
In the last 44 years, Tinsley has lost only NINE games (2 against Chinook
in 1992) - an amazing accomplishment.
Peter Rice
> Perhaps this will be obvious to you upon reading your own words again, but it
> I suppose that if the world's fastest dragster was very nearly as fast in the
> 100 yard dash as the world's fastest runners, then it *would* be interesting,
> too.
Let's just make sure it abides by the rules:
No starting the engine before the race
No straying over the lines marking the lane
No use of artificial aids (like a parachute) to stop
Testing for illegal substances, especially in the fuel tank
... and only purple yoghurts during the race :-)
--
------------------------------------------------------------------------
Peter Rice | Informatics Division
E-mail: p...@sanger.ac.uk | The Sanger Centre
Tel: (44) 1223 494967 | Hinxton Hall, Hinxton,
Fax: (44) 1223 494919 | Cambs, CB10 1RQ, England.
T. M. Cuffel
Mark A Zabel <mza...@magnus.acs.ohio-state.edu> wrote:
>
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
There is a qualitive difference between physical exploits like racing and
mental ones like chess playing.
Man is used to getting beat in the physical realm. Horses and dogs have
outrun the fastest humans since time immemorial, so watching a dragster
do it is no big deal. But chess involves brainpower, and in matters
of mind, humans have been largely unrivalled until recently. So
human-computer chess matches will maintain a John Henry-esque type
interest for some time to come.
David Bell
> So far I have solved only about 250 billion. Where have you heard that
> checkers has almost been 'solved ' by computers?
> --
> Rob Lake la...@cs.ualberta.ca
The computer doesn't have to know every possible position in it's
database!!
What's happened is that the opening library has expanded forwards into the
game, and the endgame library has expanded into endgames with more and more
pieces on the board.
What this means is that the opening and endgame books will meet up, providing
perfect play by the computer. I read in a chess column that this has almost
happened. (I think it *was* Chinook).
--
/---------------------------\ /---------------------------\
| 'Kryten, unpack Rachel |---------------------| David Bell |
| and get out the puncture |---------------------| |
| repair kit!' - Rimmer |---------------------| da...@sparky2.demon.co.uk |
\---------------------------/ \---------------------------/
David Bell
> Johannes Fuernkranz <ju...@ai.univie.ac.at> wrote:
> >In article <777066...@sparky2.demon.co.uk>,
> >David Bell <da...@sparky2.demon.co.uk> wrote:
> >>
> >Why not?
> >>You won't see a chess world championship with a computer in it! ( I hope :-)
> >Why not?
>
> Why not have the world's fastest dragster (or rocket sled or whatever)
> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
> lift compete in weightlifting competition? Chess is a sport - not
> a (pure) measure of intelligence (whatever "intelligence" is).
>
Thanks for agreeing with me. This is why computers should NOT play in human
events.
David Bell
> David> ( I hope :-) )
>
> maybe not yet, but i wouldn't bet on it not happening soon. the best
> chess programs are master-rank now.
>
The POINT is, the chess world championship is for HUMANS ONLY. It doesn't
matter how good the computers get, they will not be included.
Jan Eric Larsson
>Why is there a computer playing in the checkers world
IMHO, you can arrange three kinds of world championships, human only,
computer only, and mixed. This is certainly true for both chess and
checkers, and I think it woul'd be quite silly to try and forbid any
of them.
In practise, it is of course an open question what kind of championship
that different organizations will want to arrange, and which kind that
will be considered most interesting.
David Bell writes:
>especially seeing that checkers has almost been 'solved' by
>computers.
Not true. Jonathan Schaeffer's team (Edmonton, Canada) is working on
solving more and more positions for their Checkers position library,
(used by the Chinook program). This is a huge task, but according to
the posting by Robert Lake, they seem to be half-way.
Kelly Martin writes:
>samuel's algorithm is nearly unbeatable, as i understand it.
No true either. Samuel's programs were a great step forward for the
early sixties, but their play is mediocre by current standards. Any
simple (and non-learning) program written according to current
knowledge will play far better.
The reason that Samuel is so praised is that he pulled of an early
successful learning result. But what was written in those days must
be reevaluated before you compare with today.
Jan Eric Larsson Phone: +1 (415) 723-0948
Knowledge Systems Laboratory E-mail: Lar...@KSL.Stanford.Edu
Stanford University
701 Welch Road, Building C "We watched the thermocouples dance to the
Palo Alto, CA 94304, USA spirited tunes of a high frequency band."
Joseph Albert
Robert Williams <rlwil...@gallua.gallaudet.edu> wrote:
>... it was only after machines far
>outstripped man that the competition ended... I think when you see a
>computer beat the world champion 13 - 0, it will go back to a human only
>sport... Right now, it's still interesting...
well, apparently the chinook software can beat every human except Tinsley,
and it is unclear if there will ever again be a (human) draughts player of
Tinsley's calibre. i think this is a pretty unique situation. in
athletics and in chess, it is a given that when the best start to wane,
new, young competitors, the best of the next generation takeover.
if software doesn't get good enough to defeat Tinsley during Tinsley's
lifetime, we'll never know whether some future, better software would
have defeated Tinsley. in fact, it seems to be unclear whether a
machine could do better than draw a match with Tinsley through any
means other than sheer stamina. from a scientific point of view, this might
be as interesting as showing that a slow, electric golf cart that runs 10mph,
and can can go for 25hrs on a battery charge can defeat a human runner in a
250 mile race.
Joseph Albert
alb...@cs.wisc.edu
T. M. Cuffel
Andrew Pierce <ajpi...@med.unc.edu> wrote:
>In article <330914$a...@lace.Colorado.EDU> cuf...@cs.colorado.edu (T. M. Cuffel) writes:
>>Man is used to getting beat in the physical realm. Horses and dogs have
>>outrun the fastest humans since time immemorial, so watching a dragster
>>do it is no big deal. But chess involves brainpower, and in matters
>>of mind, humans have been largely unrivalled until recently. So
>>human-computer chess matches will maintain a John Henry-esque type
>>interest for some time to come.
>
>mental realm, they have come to rely on being beaten. Your calculator and
>spell-checker are two examples of this. Of course, no one would think of
>entering a computer in a spelling-bee.
No one would think of entering a dictionary in a spelling bee either.
Like horses and dogs, spell checking devices and calculating devices
have been with mankind for a very long time. People have grown
accustomed to them, and consider the feats they accomplish to be
uninteresting. That's why I qualified "unrivalled" with the word
"largely" when referring to traditional mental accomplishments.
Johannes Fuernkranz
David Bell <da...@sparky2.demon.co.uk> wrote:
>> David> You won't see a chess world championship with a computer in it!
>> David> ( I hope :-) )
>>
>> maybe not yet, but i wouldn't bet on it not happening soon. the best
>> chess programs are master-rank now.
>>
>
>The POINT is, the chess world championship is for HUMANS ONLY. It doesn't
>matter how good the computers get, they will not be included.
>
Same with checkers.
Hope we'll have a champ like Tinsley when the going gets tough.
He just cares about good games, not about Man or Machine.
Therefore he resigned his Checkers Worldchampionship in 1991 to play
the computer.
Juffi
Jonathan Schaeffer
>> There are 500 billion billion (i.e. 5 x 10**20) positions in checkers.
>> So far I have solved only about 250 billion. Where have you heard that
>> checkers has almost been 'solved ' by computers?
>> --
>> Rob Lake la...@cs.ualberta.ca
>The computer doesn't have to know every possible position in it's
>database!!
>What's happened is that the opening library has expanded forwards into the
>game, and the endgame library has expanded into endgames with more and more
>pieces on the board.
>What this means is that the opening and endgame books will meet up, providing
>perfect play by the computer. I read in a chess column that this has almost
>happened. (I think it *was* Chinook).
You are correct; this is what we are doing. We are close to solving
some openings, in the sense that they are draws and we will never lose
a game in the opening (barring a machine failure). It is technologically
possible to solve the entire game (5*10**20 positions) but we do not
have the resources right now to do it. Anyone have some idle cycles and
(more importantly) disk space available to us?
Mark A Zabel
T. M. Cuffel <cuf...@cs.colorado.edu> wrote:
>In article <32t3ln$d...@charm.magnus.acs.ohio-state.edu>,
>Mark A Zabel <mza...@magnus.acs.ohio-state.edu> wrote:
>>
>> Why not have the world's fastest dragster (or rocket sled or whatever)
>> compete in the 100 meter dash in the Olympics? Why not have an hydraulic
>> lift compete in weightlifting competition? Chess is a sport - not
>> a (pure) measure of intelligence (whatever "intelligence" is).
>
>mental ones like chess playing.
I don't think racing is a purely physical exploit...neither do I think
chess is purely mental. Is calculation one of the higher things we do
mentally? I don't know.
>
>Man is used to getting beat in the physical realm. Horses and dogs have
>outrun the fastest humans since time immemorial, so watching a dragster
>do it is no big deal. But chess involves brainpower, and in matters
>of mind, humans have been largely unrivalled until recently.
It will still be the case when computers beat all humans at chess,
won't it? Computers out-calculate humans. They have out-calculated
humans for many years now in areas of mathematics. If there is
a single mathematician who thinks a computer is more intelligent than
he because it can calculate an integral faster, I've yet to meet him.
Mathematics is not generally considered a physical exploit.
>So human-computer chess matches will maintain a John Henry-esque type
>interest for some time to come.
I agree, they will. But not because we haven't been beat in matters
of the brain.
Anyway, I can see I confused some people with my post...I know some
people are interested in it - that alone justifies doing it. The
question is: Does it justify calling it a "World Championship?" I happen
to think not.
-Regards, Mark
John M. Sullivan
Joseph Albert <alb...@trigger.cs.wisc.edu> wrote:
>if software doesn't get good enough to defeat Tinsley during Tinsley's
>lifetime, we'll never know whether some future, better software would
>have defeated Tinsley.
Surely as computer speeds increase we will eventually have a computer
program that plays a _perfect_ game of checkers. (It's only a finite
game, of course.) Such a program would defeat Tinsley (whenever
possible).
-John Sullivan
David desJardins
> Surely as computer speeds increase we will eventually have a computer
> program that plays a _perfect_ game of checkers. (It's only a finite
> game, of course.)
Now there's a deep comment. "It's only a finite game, of course." So
is chess. So is go. So is go on a 1000x1000 board. Should we
therefore conclude that it's only a matter of time before we have a
computer that plays go perfectly on any particular finite-size board?
If solving go positions is PSPACE-hard, does that change your answer?
David desJardins
--
Copyright 1994 David desJardins. Unlimited permission is granted to quote
from this posting for non-commercial use as long as attribution is given.
Ed Hanway
: In article <3310jl$6...@daffy.cs.wisc.edu>,
That's correct, unless Tinsley can already play perfectly. Actually he
wouldn't even have to be perfect, he'd just have to be good enough to
avoid making a game-losing mistake during the course of the match.
--
Ed Hanway --- j...@raster.kodak.com
Any opinions are mine, not Kodak's.
Joseph Albert
John M. Sullivan <sull...@geom.umn.edu> wrote:
>Surely as computer speeds increase we will eventually have a computer
>program that plays a _perfect_ game of checkers. (It's only a finite
>game, of course.) Such a program would defeat Tinsley (whenever
>possible).
yes, I believe we will see the day thata computer plays a perfect game
of checkers. It isn't clear that this will defeat Tinsley through any
method other than stamina. In particular, if the match conditions were
set up so that Tinsley was fully rested, in perfect health, etc. for
every game, it might be that the computer would only draw a match with
Tinsley.
joseph albert
alb...@cs.wisc.edu
T. M. Cuffel
Assuming checkers is drawn with best play, is there a meaningful difference
between a perfect strategy and one that never loses?
Dietmar Budelsky
Under two assumptions:
1) A program solves the chess game (This does not mean, it generates all
possible moves in the momentary game tree!)
2) The game is a draw with perfect play
we found, that the perfect program, which only is interested in the best
momentary result gets a lower result against relatively bad players
(mostly remis, some wins), than todays imperfect programs. As there
is in real no best move (only maximal three classes of moves), a really
good program has to find the move in the best class, which makes it most
likely for the opponent to make a result-changing error.
For checkers this is even true. chinook 2000 has to choose the move, which
most likely gives Tinsley the opportunity to fail.
Dietmar
Jonathan Schaeffer
Correct. Chinook is learning more about the game every day. With ennough
disk space and compute cycles, checkers could be solved in a few
years.
Jonathan Schaeffer
The "stamina" comment is not correct. We have done a thorough analysis of
all of Tinsley's published games. He has been in several losing positions
that have never been mentioned in the literature. Tinsley is almost perfect
but not quite as perfect as portrayed in the press.
Johannes Fuernkranz
T. M. Cuffel <cuf...@alumni.cs.colorado.edu> wrote:
>
>Assuming checkers is drawn with best play, is there a meaningful difference
>between a perfect strategy and one that never loses?
>
If one of them is fallible, however, you might miss wins with the latter
strategy. It's easy for white to follow a never-lose strategy in, say, a
drawn KPK endgame. (Just sac the pawn! :-)
It's a little harder to play perfectly which would include knowing when
your opponent makes a mistake, knowing how to win those positions and maybe
even knowing in which positions your opponent is most likely to make a
mistake.
Juffi
Jan Eric Larsson
>The POINT is, the chess world championship is for HUMANS ONLY. It doesn't
>matter how good the computers get, they will not be included.
Not a good point. You forgot to tell which chess world championship
you are talking about...
Humans only, machines only, women only, juniors only, etc.
Jan Eric Larsson
>What this means is that the opening and endgame books will meet up,
>providing perfect play by the computer. I read in a chess column that
>this has almost happened. (I think it *was* Chinook).
Some facts you might want to know:
1. You are talking directly over e-mail to the guy that is doing
the "solving checkers" stuff, and the world is listening.
2. The Chinook team has made a great effort in coming closer to
solving checkers. The search (seems to be a minimum of 19 full
plies now), and the size of the endgame databases means that they
were one the way of solving checkers a while ago. When Robert
Lake ha finished his work, there you have it. If he says he is
half way, then let's trust him: he is half way.
3. The older Chinook results were so encouraging that many described
checkers as "solved", "almost solved", etc. It isn't yet, and
Robert Lake is probably the one in the world that knows best about
the progress...
Sorry to respond/attack two of your postings on one day, but this
latter one was really silly. Quoting memories of some chess column
you read in the face of the guy doing the work that chess column was
about make you seem like a blithering idiot at best. I don't think
you are, but you may want to read and think, and check your facts a
bit more thoroughly before posting.
Siva Chelliah
>> David> ( I hope :-) )
>>
>> maybe not yet, but i wouldn't bet on it not happening soon. the best
>> chess programs are master-rank now.
>>
>
>The POINT is, the chess world championship is for HUMANS ONLY. It doesn't
>matter how good the computers get, they will not be included.
>
>
Why? who told you so?
Please note that it has been proven over and over again that just
having a faster computer is not enough to win at chess. It might be so
in the future, but as of now speed is not alone enough. This is why
you can't just compare chess with 100 meeters run in olympics (as a
previous poster compared.)
Because of this reason, IMHO, it is ok to allow computers in the world
championship.
My only complain is that a computer can remember 1000s of opening moves
and I can't.
--
Siva Chelliah BNR,Richardson.TX
------------------------------------------------------------------------------
Life is a game you play once. In order to win you must make a difference.
Remember: This is not a practice session.
Andrew Pierce
>My only complain is that a computer can remember 1000s of opening moves
>and I can't.
My only complaint is that a computer is allowed to get out additional
boards and move the pieces around to explore various lines while I am
restricted to staring at one game board with the current position on it.
-Andy
ajpi...@med.unc.edu
Lance A. Brown
> My only complaint is that a computer is allowed to get out additional
> boards and move the pieces around to explore various lines while I am
> restricted to staring at one game board with the current position on it.
You mean you don't look at the board and play out sequences in your
head? How do you decide on what move to make?
Lance
Andrew Pierce
Yes... I do, however, the computer is allowed an unlimited number of
perfect boards and I am not allowed that (since my memory and
visualization are not perfect). Furthermore, computers can increase the
number of these boards by simply adding in more chips and I have no such
option. I don't know how much more strength a human player would gain if
he/she was allowed to use additional boards to explore variations but I
would guess that it would be a considerable amount. I don't dispute that
computers have better memories than humans (or at least the potential for
better memories) but games like chess, go and checkers don't interest me
if they are simple memorization/visualization contests. I say give the
human an extra board and let him/her play out the variations they want (in
the time alloted of course).
-Andy
ajpi...@med.unc.edu
David desJardins
> I don't know how much more strength a human player would gain if
> he/she was allowed to use additional boards to explore variations but I
> would guess that it would be a considerable amount.
For competent high-level players, almost exactly zero.
> I say give the human an extra board and let him/her play out the
> variations they want (in the time alloted of course).
The reason the game isn't played this way is that it isn't particularly
useful, not because it would confer some tremendous advantage.
Robert Giffords
unfortunately. However, sometime in the late 1800's or early 1900's there
was a match played between two of the strongest players of the day, with the
proviso the players could, if they wished, move the pieces on the board
to assist in thier analysis. (Presumably, they could only do this on
their move).
Anyway, one player chose to take advantage of this option, the
other did not. The player who did not move the pieces during analysis won
the match.
I am sure I could dig up the exact details on the contest if
someone were interested.
Joseph Albert
Andrew Pierce <ajpi...@med.unc.edu> wrote:
>
> My only complaint is that a computer is allowed to get out additional
>boards and move the pieces around to explore various lines while I am
>restricted to staring at one game board with the current position on it.
yeah, but the computer has to play blindfolded :-)
Joseph Albert
alb...@cs.wisc.edu
ed.knowles
thousands of opening moves, too. You could too, if you trained your brain.
Now you can argue that the computer doesn't "remember" in the same way you
do, but then probably no two people remember in the same way, since human
memory is to a large extent associative, and we all have different experi-
ences. Also, through visualization techniques, you too can "get out" additional
boards and move the pieces around... In fact, the human brain has a huge
advantage over the computer because of far superior pattern recognition
capabilities. I think the best humans will always be "better" than the best
computers at chess, but the humans will lose matches to the computers be-
cause biological hardware is less reliable than electronic hardware. Now
I think that computers will very soon be unbeatable at checkers, but will
never even come close to being really good at GO. But this is rec.games.chess
isn't it? What am I mentioning those other games for ? 8^)
-Ed Knowles
The novice said to the GM, "Everytime I move a piece it's a tactical
mistake, and every time I move a pawn it's a positional mistake. What
should I do?". The GM replied," Move something else."
Steven L Harrington
David desJardins <de...@ccr-p.ida.org> wrote:
>Andrew Pierce <ajpi...@med.unc.edu> writes:
>> I don't know how much more strength a human player would gain if
>> he/she was allowed to use additional boards to explore variations but I
>> would guess that it would be a considerable amount.
>
>For competent high-level players, almost exactly zero.
Wrong! Already such a forum exists - postal chess. I think you
will find that the games in postal chess are at a *very* high level
(the top players that is...). In fact, if you doubt this then take
your best computer (Deep Thought, or whatever) and see how it fares
in a top flight postal tourney. I am willing to bet it'd come in
a solid last place. However, take those same postal players and
sit them down at a board under normal tourney conditions against DThought
and DT would probably score its fair share of pts.
>The reason the game isn't played this way is that it isn't particularly
>useful, not because it would confer some tremendous advantage.
Again, I refer you to postal chess. Humans are allowed opening manuals
and to use the board. This is the closest setting to truly fair
playing conditions between human and machine. In fact, were a computer
to win the world postal championsip I'd have to say that it had
done so completely fairly. I might not say the same thing if
a computer won a normal OTB world championsip for reasons such as
opening book and ending book stored on disk etc. Of course, this is
all a moot point until computers get significantly better at chess
since they presently offer no real threat to win the World Championship
at any time control other than blitz.
-s.l.h.
-ics:cowboy
David desJardins
> Wrong! Already such a forum exists - postal chess. I think you
> will find that the games in postal chess are at a *very* high level
> (the top players that is...).
This is getting completely ridiculous. The quality of play in postal
chess is better than over the board because players have days rather
than minutes to make each move. Not because they are allowed to push
the pieces around on the board.
Charles Blair
>Andrew Pierce <ajpi...@med.unc.edu> writes:
>> I don't know how much more strength a human player would gain if
>> he/she was allowed to use additional boards to explore variations but I
>> would guess that it would be a considerable amount.
>For competent high-level players, almost exactly zero.
The first edition of OXFORD COMPANION TO CHESS reports an incident
in which the grandmaster Matulovic took back a move after taking his
hand from a piece. Seems unlikely this would happen if he had an extra
board to play with. (I have seen claims of similar activities by other
grandmasters, but don't have references handy)
Steven L Harrington
David desJardins <de...@ccr-p.ida.org> wrote:
>Steven L Harrington <shar...@cs.tamu.edu> writes:
>> Wrong! Already such a forum exists - postal chess. I think you
>> will find that the games in postal chess are at a *very* high level
>> (the top players that is...).
>
>This is getting completely ridiculous. The quality of play in postal
>chess is better than over the board because players have days rather
>than minutes to make each move. Not because they are allowed to push
>the pieces around on the board.
Yes, it's getting ridiculous but for reasons other than those you may
expect... Ask any postal player (hell, ask me - i've played postal
before) whether they ever move the pieces around in examining variations
in their games. You will find that the answer is a resounding "yes".
That isn't to say that they do so always or exclusively, nothing can
remove the necessity of looking in one's head, but they nevertheless
do so. Hell, ask most any OTB player of merit whether they move the pieces
when studying at home or preparing/examining a new opening line. Same
answer... If postal players don't do it, then I suggest that we
poll them and see if this is so. Otherwise, if they do move pieces
then I ask the (perhaps rhetorical) question: why do they do it,
if it doesn't help them or offer them any advantage? In addition,
there is the issue of using opening books and ending books
which is the closest analog to a computer's disk storage that you
will find in the human realm - namely, chess books are non-volatile
much like the computer's hard drives. (And, yes, I know that hard
drives are more volatile than books for the most part but for all
practical purposes they are "static" in this context). As for the issue
of days rather than minutes, this does play a role but not as much as
you might think. Namely, most players don't actually spend "days"
contemplating their moves - usually minutes to hours. So,
this does have an effect on the play (removing the problems associated
with time pressure) but it is not the sole factor and it still doesn't
address the issue of why they move the pieces if it doesn't help. Maybe
you can tell me an anecdote of a great postal player who doesn't move
the pieces but I'm willing to bet that he/she would be an exception rather
than the rule.
-s.l.h.
-ics:cowboy