Main
Date: 03 Oct 2007 07:15:26
From: raylopez99
Subject: Computer Science Experts: Chess 'harder' than Go
In an article in Science, 14 September 2007 p. 1519, written by the
Chinook team that weakly solved the game of checkers (draughts), it's
asserted essentially that chess is harder than Go.

What is meant is that the "decision complexity" of chess (and
checkers) is higher than for Go (actually Go-Moku was mentioned, which
is not exactly the same as Go, but I take it the analogy holds for
Go), while the game of Go has a higher "space complexity" (essentially
non-repeating patterns--more such patterns in Go than Chess/Checkers).

What is the consequence of this? It's harder for a computer to
"solve" the game of Go only because there is more "stuff" the computer
has to consider--even though, for each move, less thought is
involved. From the author's paper, page 1519: "Checkers [and by
extention, chess, mentioned in the article--RL] ...requires extensive
skill [<--note that phrase--RL] to make strong move choices...and
[has] moderate space complexity... All the games solved thus far have
either low decision complexity (Qubic and Go-Moku), low space
complexity (Nine Men's Morris...Awari).. or both (Connect Four)."

So, let me translate for you Go players: Go is a game that even a
monkey can play, the moves require little or no thought, you can't
make a "one-move blunder" easily, like in chess, and, consequently,
because so many possible patterns exist, computers with present day
hardware are "fooled" into playing weakly. However, once hardware
improves, Go will be weakly solved (meaning with 99.9999999...%
probability, like checkers) and it will be seen that Go players are
nothing but egotistical naked apes. Watching and playing pachinko is
more interesting than Go. The true cerebral game of skill is, will be
and remains, international chess.

That's what SCIENCE sez, sucker. Don't blame me for reporting the
FACTS.

RL





 
Date: 01 Nov 2007 09:39:37
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On 3 okt, 18:35, [email protected] (-) wrote:
> > raylopez99 wrote:
> >> ... --even though, for each move, less thought is involved.
> Michael Alford <[email protected]> wrote:
> > utter nonsense
>
> Agree. We don't need to discuss "what is thought" in order
> to build a game-playing device.
>
> >> From the author's paper, page 1519: "Checkers [and by
> >> extention, chess, mentioned in the article--RL] ...requires extensive
> >> skill [<--note that phrase--RL] to make strong move choices...and
> >> [has] moderate space complexity... All the games solved thus far have
> >> either low decision complexity (Qubic and Go-Moku), low space
> >> complexity (Nine Men's Morris...Awari).. or both (Connect Four)."
>
> >> [ ... ]
> > Total bullshit.
>
> First of all, in order to discuss the Science Magazine article we
> should have the author names cited. Next, at risk of treading upon
> "sacred temple territory" of these Holy Go Spaces, the considerations
> expressed are sound. Frank de Groot has also been here, a prophetic
> champion of "gate arrays." There exists a hardware solution to the
> problem of exponential fan-out, albeit that the hardware solution is
> still elusive. Disinterest by much of a computer artificial intelligence
> community should not be misinterpreted to indicate severe impossibility.
> Instead we could examine the sociological aspect: what led computer
> scientists first to attack Chess instead of Go? Was there a perception
> that Chess is more highly respected as a decision-rendering tool for
> purposes of conducting our state-of-affairs in World Diplomacy?
>
> As the Science Magazine authors note, a solved game may fall due
> to low decision -OR- low space complexity. The decision side is more
> difficult to visualize, and to devise in hardware, however it remains
> true that the impenetrability of a game to machine analysis is the
> result of a combination of both and not simply one aspect or the other.
> The fact that many millenia of human history had elapsed prior to atomic
> weapons does not invalidate the reality that atomic weapons exist today.
> When the hardware devices are invented people may come to regard
> that scientific team with awe and respect, as was accorded to Einstein.
>
> Go Players unprepared for the prospect that their favorite sanctuary
> will someday be invaded, and overrun by electronic hoardes, share an
> immaturity aspect which the Chess Community has long ago shed. It
> will be interesting to observe how many Go Players continue once those
> hardware devices gain an advantage: this would surely be a test of faith
> once the rubber meets the road. Who prefers not to live to see that day?
>
> - regards
> - jb
>
> -------------------------------------------------------------------
> Tamiflu mutates in sewage systems ...http://afp.google.com/article/ALeqM5johTPaZnrC32MSU6WyDgsQL4KcGA
> -------------------------------------------------------------------

i`d love to see that day since i want to become very very very old



 
Date: 10 Oct 2007 03:27:57
From: shapenaji
Subject: Re: Computer Science Experts: Chess 'harder' than Go
Harry Fearnley wrote:

> Using an analogy, I am pleased to report that I have found a
> conclusive proof of your assertion:

> A clock is an object.
> A brick is an object.
> All objects are of equal value.
> A clock has many components, and a brick only one.
> To have more components is to be simpler.
> Therefore, a clock is simpler than a brick.
> Q.E.D. (just to prove I know some Latin)

> From this I can further safely conclude that a brick will offer
> _you_ more intellectual stimulation than a clock.

I'm not an rgg'er and this thread is quite obviously a troll, but I
found this hilarious.

I'd like to go one further though,

All objects are of equal value
Some objects contain components
Components are also objects
If all objects have equal value, then in order for two objects with
different numbers of components to have the same value, all objects
must have no value.

Wow, makes go and chess seem rather pointless doesn't it?

-Nicholas Jhirad




 
Date: 08 Oct 2007 07:20:36
From: --CELKO--
Subject: Re: Computer Science Experts: Chess 'harder' than Go
>> I have often wondered why nobody seems to make a Go board with dimples like a Chinese checkers board; I suspect that this is because it would interfere with that satisfying click that you get with a good Go set. <<

I saw such a board decades ago at a folk craft shop that had Morris,
Chinese Checkers, and a host of other games with identical pieces done
with "bles and cups". The boards were wood and the Go board was
2-3 inches thick. I have no idea how to find them.

I have heard of but not seen a Korean board with wire inset into the
lines so that you get a slight musical tone when you set a stone.



 
Date: 06 Oct 2007 19:30:45
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 6, 7:56 pm, "Roy Schmidt" <[email protected] > wrote:

> HAHAHAHAHAHAHAHA OOH HAHAHAHAHAHA
> Mr. Lopez, you have been trolled by the master troller (i.e., jb


In another thread, I read a small portion of a post
which attempted to describe how a master "troller"
operates. Somewhere in that very long posting, it
was said that the true master does not return to
the thread once he posts his "troll"; yet this guy
seems to be a relative newbie, having both returned
and gloated. (Yawn.) Go back to school.


> Although _Science_ is a famous publication, sometimes a lemon gets into
> their pages.

You posit a correlation between fame and accuracy?
Naivete: look it up (and correct my spelling, if you like).


> Remember the "cold fusion" debacle?

No. Please elaborate: what was the cold fusion
debacle? I do recall the mate-in-one debacle in
which world chess champ Kramnik embarrassed
humans the world over in front of every two-bit
computer chess program -- even Sanny's GetClub
program does not often overlook one of those.


-- help bot





 
Date: 06 Oct 2007 00:50:12
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go

raylopez99 wrote:

> That's what SCIENCE sez, sucker. Don't blame me for reporting the
> FACTS.

After reading for years about how go is supposedly superior
to chess, I went to Wikipedia.com and read up on that "other
game" a bit. Although it is clear that there are more turns and
less reliance upon a single "one-move-blunder" to, perhaps,
decide a game of go, I was a bit disappointed by the dullness
of the playing field. Just a flat board, with lines drawn across
it, both the field of play and the pieces/men are dull, boring.

In sum, I think that both chess and go could easily be
improved upon, a much more interesting and varied game
invented. I'm not saying that *I* could do it, but someone
who is very creative should have little trouble since these
games are so one-dimensional, have such tiny playing fields
and limited scope. Most likely, the invention and spread of
the computer will -- eventually -- render old-style board games
passe.

Here are a few things I learned about go from the Wiki
article:
It is a game of territory, of surrounding men to capture
them and remove them from the board. It is probably much
easier to give odds in go than in chess, without mucking up
the original character of the game. One powerful sneeze and
countless hours of work disappear in the blink of an eye, for
the men are small plastic "stones" that just sit loose upon a
wooden board. Strategy consists largely in surrounding
"space", while not allowing one's own men to get surrounded.
(Dull as dirt; in online Star Trek style games, you *explore*
new territories, fight unknown enemies, etc.) Like chess,
you play one against one, taking turns. (Again, dull as dirt;
in reality, one can be overtaken by faster opponents, like on
approaching a freeway exit and five people cut in front at the
last minute.)

As I have said before, I should give up chess. The only thing
stopping me is my amazing results; I am by far the greatest
player ever on GetClub; undefeated on ChessWorld; able to
save many lost games with Superman-like powers OTB. In
sum, both chess and go bore me, yet I am seemingly,
hopelessly addicted to chess.


-- help bot



  
Date: 07 Oct 2007 14:02:13
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go



help bot wrote:

>the original character of the game. One powerful sneeze and
>countless hours of work disappear in the blink of an eye, for
>the men are small plastic "stones" that just sit loose upon a
>wooden board.

Those boards are like the chess sets with the super-lighweight
hollow plastic men and the red/black board.

A good Go set has ceramic, porcelain, clamshell, slate or
glass stones, and a wood surface that, with years of play,
develops indentations on the playing surface, which adds to
the value of the board. I have ofen wondered why nobody seems
to make a Go board with dimples like a Chinese checkers board;
I suspect that this is because it would interfere with that
satisfying click that you get with a good Go set.

I am told that Yunzi stones from the Yunnan province of China are
really nice - sintered jade and sintered quartz.

BTW, I consider Renju to be a far superior version of Gomoku.
It isn't up there with chess or Go in my opinion, but it is
a serious game, at least as deep as Backgammon.
More info: [ http://www.renju.net/study/rules.php ]

Also, [ http://senseis.xmp.net/?StartingPoints ] is a really
nice Go site.

--
Guy Macon
<http://www.guymacon.com/ >



  
Date: 06 Oct 2007 17:48:55
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Sat, 06 Oct 2007 00:50:12 -0700, help bot <[email protected] >
wrote:

> After reading for years about how go is supposedly superior
>to chess, I went to Wikipedia.com and read up on that "other
>game" a bit. Although it is clear that there are more turns and
>less reliance upon a single "one-move-blunder" to, perhaps,
>decide a game of go, I was a bit disappointed by the dullness
>of the playing field. Just a flat board, with lines drawn across
>it, both the field of play and the pieces/men are dull, boring.

Some find a minimalist esthetic more appealing.

>One powerful sneeze and
>countless hours of work disappear in the blink of an eye, for
>the men are small plastic "stones" that just sit loose upon a
>wooden board.

Very few players use plastic stones for a more immediate reason: they
are too light, and stick to your finger when you try to play one on
the board. Glass is by far the most common go stone material.

-- Roy L


 
Date: 05 Oct 2007 18:42:10
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 5, 8:38 am, raylopez99 <[email protected] > wrote:

> Dam bot, I thought you were a chess professional, the way you posted
> inside stuff?! Maybe you just hang out at the Manhatten chess club
> so much that you've become a regular and equal in playing strength to
> a grandmaster.


I visited the MCC once. It was late, and the only
person I saw was someone who wanted a visitor's
fee, so I left. Bad timing, I guess.

The New York crowd seems to have gotten in a
bit of a huff over the USCF moving its headquarters
to a backwoods state (Davy Crockett!).

Look, if I had the playing strength of a grandmaster,
would I have lost two games as help bot and a few
more as nomorechess against the GetClub program?
I don't think so! Just recently, I played and drew a
game which was "unrecorded" there, on account of
having run out of credits to play rated games for free.
I was basically forced to repeat the position, because
I had a Queen against an army of invaders, as lost
as a cabbage, I was.

And what on earth do you mean by "inside stuff"?
Did I write about the time I overheard GM "Karmink"
whisper to his second: "what was I supposed to do
after 31.K-h8 in this line again"? Or the time that
GM Larry Evans demanded $5,000 "up front" when
he learned that BF had not yet signed a contract?
No. I have never to this very day revealed the real
bust to the King's Gambit, how BF will likely never
have to admit the French Defense is sound, or any
other such things I have learned over the years.
Loose ships sink lips.


-- help bot



 
Date: 05 Oct 2007 06:38:51
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 5, 12:20 am, help bot <[email protected] > wrote:

> For some reason, when a player is a math teacher
> everyone seems to know it, and over the years I
> have played quite a few, many of them excellent
> chess players, but certainly not all. My toughest
> opponents were of course the professionals, the
> players who travel around playing chess for a
> living. But after them, I would put the people who
> have a burning desire to win, along with enough
> talent to keep that from being rendered irrelevant.

Dam bot, I thought you were a chess professional, the way you posted
inside stuff?! Maybe you just hang out at the Manhatten chess club
so much that you've become a regular and equal in playing strength to
a grandmaster.

RL



 
Date: 05 Oct 2007 06:35:52
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 11:45 pm, [email protected] (-) wrote:
> raylopez99 <[email protected]> wrote:
> > Reinfororcement Learning of Local Shape in the Game of Go
> > David Silver, Richard Sutton, and tin Muller

> > Essentially, there is a crack in the GO facade of impenetrability--if
> > trees exist in GO, GO can be solved like chess, since computers are
> > good at solving trees (much better than pattern recognition it seems).
>
> You're walking up on a discussion which periodically washes
> over this newsgroup. A lot of people have "been there, done that."
> Gaming analyzes strong and weak, and there are inroads by computers.
> Yet, among the three scales for Go Ranks -- kyu, amateur dan, pro da=
n --
> known computer programs have not yet graduated beyond the kyu scale.
> Kyu level is through high-school, amateur dan is college level, and
> pro scale is competitve play at a very high level, to teach college.
> Despite some very intense computer science efforts by the best among
> the best, the Go Programs do not even have a high school diploma.

You're so harsh...perhaps the next generation of Moore's law will
change your thinking (though supposedly we're getting close to the
limit for Moore's law).


>
> > "CHESS IS SHARPER"--this is what I was driving at. I've not played
> > GO except on a Palm with a small 9x9 board, only a few times, but I
> > somewhat got the feel each move is not absolutely critical, but it's
> > the accumulation of moves (of course the same is true in chess, but to
> > a lesser degree--think how often have you 'blown' a winning position
> > by making a one-move blunder, meaning every move 'packs more punch'
> > than in Go). But chess had a more tactical feel than (and chinese
> > chess has an even more tactical feel than chess).
>
> > ... thus I prefer chess over more "complex" games like GO or
> > imitating schoolchildren. I do have a 18x18 board and stones that I
> > intend to at some point learn, but unfortunately there's not much
> > computer help in GO as there is in chess, which hinders learning the
> > game.
>
> The Japan Go Academy (Nihon Ki-In) is sometimes rendered by
> machine translators as "the Chessman Institute." In practice most
> Asians are not quite so much interested in competitive distinctions
> among their many varieties of cognitive-skill boardgames. Games
> each exist in their own right -- forms of mental sports tial arts=
--
> and obtain intrinsic respect in that regard. In abstract, computers
> don't care much what sort of gaming problem they're really solving.
> If one were to examine code the initial appearance of a game gets
> lost pretty quickly among those artificial intelligence entanglement=
s=2E

Interesting. I did not know that. I thought each game had it's own
distinct "signature", with chess being distinquished by the "alpha-
beta"/ min-max "signature".

>
> "raylopez99" <[email protected]> wrote in message
> > jb is one of the pioneers of chess programming and AI--and he
> > agrees with part of what I said. Who are YOU? (a nobody, that's who).
>
> I have been internet-cited on a dispute regarding copyrights and
> game programming, and am published for some math conjectures,
> but I am not even a minor contributor for game programming A.I.

Well you're almost famous. I'll add you to the list of people I've
flame-baited and trolled with over the years, including vin Minsky,
back over 10 years ago when the internet was young and people actually
thought monkeys were not typing at the other end! I think I was
discussing randomness with Dr. M, if I recall.

>
> Luckily I obtained a copy of the _Science_ paper today. It is =
not
> particularly newsworthy but reflects much of the same information as
> was on record 3-to-5 years ago. Authors delayed publication while
> they conducted lint-checking and sanity-checking, and wordsmithed
> it for _Science_. =20

Yes, pretty obvious, I saw that myself and don't even work in the
field. But I learned a new definition from that paper: "weakly
solved" means you can get to the end-game using an algorithm (alpha-
beta/ min-max I think), without of course exhaustively searching each
ply, 'breath-wise'.

> The game of Go is mentioned nowhere in the article:
> a sign to me which suggests that the authors intended to distance th=
is
> "solvability" concept from the Go game as much as they could. So
> "raylopez99" breached the very wall that esteemed _Science_ authors
> tried to carefully build, and paraded before our newsgroup luminaries
> his unwarranted example of intellectual dishonesty and arrogance.

There are no luminaries here but you jb. There rest of us are Go and
chess bums with at best 'high school' (gymnastium?) degrees or less.

>
> The _Science_ paper is surely a vellous coordination of geeks
> arriving at "one of the longest running computations completed" in t=
he
> history of computer proofs. One wonders why it might be so difficult
> to accept that best play on both sides produces a draw in Checkers.
> Powerful forces sought to instill the notion that either one side or=
the
> other "must have" an irrefutable advantage. Against those powerful
> forces their comptuers labored day-and-night for more than 15 years
> to establish these ceaseless facts of dynamical balance in Checkers.
> According to the publication rules at _Science_ magazine us ordinary
> folks must wait one year before electronic document text is released.
> Only then may we ingest their stale report without a trip to the lib=
rary.

Yeah, Science magazine is kind of expensive--I'm a businessman and I
write it off my expenses, but even so I might cancel my subscription.
The abstracts are good but who can read the detailed reports but a
specialist in that field? And better if they release the detailed
reports in electronic form, like you say, but there's money to be made
in pulp wood sales. BTW, the same issue of science has a very cool
paper on ethnic violence and boundary lines (Global Pattern Formation
and Ethnic/Cultural Violence May Lim, Richard Metzler, Yaneer Bar-
Yam* *To whom correspondence should be addressed. E-mail:
yaneerATnecsi.edu Published 14 September 2007, Science 317, 1540
(2007) DOI: 10.1126/science.1142734)

Basically it says that where ethnicities are "pure" or "completely
mixed" there is no violence, but where there's a critical ratio of
majority/minority, typically at a boundary between two groups, the
majority beats up on the minority.

I've tried to contact the author of the above but no reply--this looks
very much in your field of expertise--kind of like a super
sophisticated version of the famous LIFE program.

>
> More author-names are listed, than were supplied by the website
> record of abstract: Jonathan Schaeffer, Neil Burch, Yngvi Bj=F6rnss=
on,
> Akihiro Kishimoto, tin M=FCller, Robert Lake, Paul Lu, Steve Sutp=
hen.
> A "google" search on everything these guys have done might save
> all of us needless muck-churning and then we can carry-on with this
> ionospheric discourse already so familiar to the researchers spotlig=
hted.
>
> - regards
> - jb
>

Don't tell me jb that you're looking for intelligent conversation
online? LOL, you won't find it here that's for sure! I use this name
for flame baiting. My posts are somewhat factual only because those
kind of posts draw the most replies I've found, from heuristic tests
over the years. Not unlike the US (and maybe French?) evening news on
the television, which are designed to keep your attention in-between
commercial breaks (adverts). They are not really designed to inform
you, since the average 30 minute newscast only has about two pages of
hard information in it, which you could read in about 5 minutes. TV,
like the internet, is usually only for entertainment.

>From your website: "I mainly work on global optimization techniques
(Genetic algorithms, simulated annealing, interval programming, branch
and bound, neural networks, etc) applied to Air Traffic Management and
Air Traffic Control. I co-organized a few conferences on GA EA95,
EA94. We work on trajectory planning, conflict probe and conflict
resolution, Free Flight, slot allocation..."

I remember a standard "C" program assignment at university was to
design an air traffic control simulator. Mine crashed, literally and
figuratively, but I still got a passing grade. Then I decided there's
more money outside of science. BTW the US FAA is very primitive,
nothing like France or Canada's equivalent, which is much more
advanced. THey don't even use GPS officially in the US for air
traffic control, last I checked.

Nice talking to you professor director jb... I bet you have a nice
cushy job, and your pick of attractive students to tutor once in a
while (French girls are very cute, if you swing that way)--good for
you, viva la France. :-)

Au voir, or goodbye,

RL



  
Date: 05 Oct 2007 14:54:20
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

>> raylopez99 <[email protected]> wrote:
>>> Reinfororcement Learning of Local Shape in the Game of Go
>>> David Silver, Richard Sutton, and tin Muller
>>>
>>> [ ... ]
>>>
>>> Essentially, there is a crack in the GO facade of impenetrability--if
>>> trees exist in GO, GO can be solved like chess, since computers are
>>> good at solving trees (much better than pattern recognition it seems).

> (-) wrote:
>> Despite some very intense computer science efforts by the best among
>> the best, the Go Programs do not even have a high school diploma.

raylopez99 <[email protected] > wrote:
> You're so harsh...perhaps the next generation of Moore's law will
> change your thinking (though supposedly we're getting close to the
> limit for Moore's law).


I began, within this thread, speaking of futuristic hardware devices
(extensions upon "gate arrays") which would crack a Go Problem that
proves to be unmanageable by software (as presently understood). In
the past I have also spoken of the "distributed computing" approach for
addressing an online Go Program. Recently IBM is working on a new
form of memory, speedier, more compact, and faster than disk drives.



> ... I thought each game had it's own
> distinct "signature", with chess being distinquished by the "alpha-
> beta"/ min-max "signature".


Alpha-beta minimax is a common component of most game engines.
Other novel techniques are also included, with their own nomenclatures.



> ... I'll add you to the list of people I've
> flame-baited and trolled with over the years, including vin Minsky,
> back over 10 years ago when the internet was young and people actually
> thought monkeys were not typing at the other end! I think I was
> discussing randomness with Dr. M, if I recall.


I exchanged some email with Dr. Minsky when he stated that elderly
people past their usefulness should be sumily terminated. I assume
he was the real Dr. Minsky ... appearance of a legitimate stanford addr.



> I learned a new definition from that paper: "weakly solved" means you
> can get to the end-game using an algorithm (alpha-beta/ min-max I think),
> without of course exhaustively searching each ply, 'breath-wise'.


Doesn't visit all leafs of the game tree but, through mathematical
argument, establishes that entire classes and branches are uninteresting.



> There are no luminaries here but you jb. There rest of us are Go and
> chess bums with at best 'high school' (gymnastium?) degrees or less.


If I were the only luminary then what why would I read what others write?



> BTW, the same issue of science has a very cool paper on ethnic violence
> and boundary lines (Global Pattern Formation and Ethnic/Cultural Violence
> May Lim, Richard Metzler, Yaneer Bar-Yam* *To whom correspondence
> should be addressed. E-mail: yaneerATnecsi.edu Published
> 14 September 2007, Science 317, 1540 (2007) DOI: 10.1126/science.1142734)
>
> Basically it says that where ethnicities are "pure" or "completely
> mixed" there is no violence, but where there's a critical ratio of
> majority/minority, typically at a boundary between two groups, the
> majority beats up on the minority.


We don't need a _Science_ article to read the police blotter.



> I've tried to contact the author of the above but no reply--this looks
> very much in your field of expertise--kind of like a super
> sophisticated version of the famous LIFE program.


Maybe the other way around: John H. Conway is a flesh-and-blood
person who -imagined- the LIFE program, and then there are variants.



> Don't tell me jb that you're looking for intelligent conversation
> online? LOL, you won't find it here that's for sure! I use this name
> for flame baiting. My posts are somewhat factual only because those
> kind of posts draw the most replies I've found, from heuristic tests
> over the years. Not unlike the US (and maybe French?) evening news on
> the television, which are designed to keep your attention in-between
> commercial breaks (adverts). They are not really designed to inform
> you, since the average 30 minute newscast only has about two pages of
> hard information in it, which you could read in about 5 minutes. TV,
> like the internet, is usually only for entertainment.


You've been sheepled by the controlled mainstream-media TV culture.
Internet can offer for us alternatives to mind-numbing tune-out droning.



> From your website: "I mainly work on global optimization techniques
> (Genetic algorithms, simulated annealing, interval programming, branch
> and bound, neural networks, etc) applied to Air Traffic Management and
> Air Traffic Control. I co-organized a few conferences on GA EA95,
> EA94. We work on trajectory planning, conflict probe and conflict
> resolution, Free Flight, slot allocation..."


No, this was somebody else ("james") claiming to be that guy.



> Nice talking to you professor director jb...


I was called "professor" in the Army, but for different reasons.




- regards
- jb

-----------------------------------------------------------------
Pressure Is Finally On US Germ Labs
http://www.rense.com/general78/pressure.htm
-----------------------------------------------------------------



   
Date: 05 Oct 2007 15:15:53
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

[email protected] (-) wrote:
> I exchanged some email with Dr. Minsky when he stated that elderly
> people past their usefulness should be sumily terminated. I assume
> he was the real Dr. Minsky ... appearance of a legitimate stanford addr.


Oooops! Strike that. Was John McCarthy at Stanford, not Dr. Minsky.





    
Date: 05 Oct 2007 18:36:41
From: Kenneth Sloan
Subject: Re: Computer Science Experts: Chess 'harder' than Go
- wrote:
> [email protected] (-) wrote:
>> I exchanged some email with Dr. Minsky when he stated that elderly
>> people past their usefulness should be sumily terminated. I assume
>> he was the real Dr. Minsky ... appearance of a legitimate stanford addr.
>
>
> Oooops! Strike that. Was John McCarthy at Stanford, not Dr. Minsky.
>
>
>

So, what are we to make of the strange circumstance that vin had a
Stanford e-mail address?


--
Kenneth Sloan [email protected]
Computer and Information Sciences +1-205-932-2213
University of Alabama at Birmingham FAX +1-205-934-5473
Birmingham, AL 35294-1170 http://www.cis.uab.edu/sloan/


 
Date: 05 Oct 2007 00:20:52
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 8:21 am, Harry Sigerson <[email protected] > wrote:

> There's that divide and rule thing in action.
> Take the middle one first, mathematics. [I'll infer possibly wrongly that
> your 'real-world value' relates to financial gain]

Ah, that had slipped my mind. I was rather thinking
of science, and all the real good that can come from
mathematics as it is used in science. Now, there are
some who talk about the "science" of economics as
something really great, but of course they just happen
to have lots of money, so they /would/ think that.


> The music one, is almost simpler in that you can't turn your ears off; you
> can look away from a tv set but the ears keep on hearing the crap ads. If you
> live alone you can dump the goggle box but few do.

I keep mine around "in case of emergency"; a real
dust collector. That reminds me: in a year or two, all
air signals will go digital, so I need to unload it before
it's too late.



> The present day stories are told by moving pictures, to millions at a time
> with minimal recourse to language; it's all there in the coloured pictures. Even
> to the bullets battering through the baddy's sternum and, nowadays, you can even
> watch them squirting out through his back.
> Who needs literature in such an society; as for those 'illiterates', now
> there's a very flexible decation line all ready and renewed for addressing.
> /Flashing digression alarm noticed/


Let me just say that I have never read any poem in
Chinese; never read the greats in French, nor in
German. Even when I go to read titles in English
which I have read or heard about for years, I am
often let down and come away thinking I would not
have missed out on anything, had I never done so.
Much is empty hype.


> The old-fashioned way is still the best,

That's almost a dead-on quote from "Barbarella"!


> The cartoonist has you looking at the backend of a large car of the Rolls or
> Bently variety, being driven sedately along by an old buffer, his wife sitting
> upfront with him.
> They are being overtaken by a Mini going hell-for-leather, up in the air a
> couple of feet above the tarmac (asphaltic concrete) as both cars breast a rise.
> The old gent says...
> "If the good God had intended little cars to go that fast he'd have given
> them bigger wheels."

The more things change, the more they stay the same.
Bentley is still (nearly) tops in horsepower and heft.


> > ...But there are always a few who strive but
> > fail, in spite of any accomplishments in other fields.
> > Look at Dr. Einstein and Dr. Blair, for instance. ;>D
>
> I can't recall the names in this possibly apocryphal tale about Albert. It
> seems that he had a friend who was a good or even great violinist and Einstein
> had some pretensions as a player. They played together.
> The friend, irritated at some point in their duet admonishes,
> "No, no, no, Albert! It's 'One two three; One two three,...'
>
> To make comparisons across such a boundary as that between Einsteinian
> mathemtics and anything else he did is normal; celebrity has to pay its dues but
> it is to some extent pointless.

I used this particular example because it is well-
known that AE was no great chess player, though
he was friends with someone who /was/. Chess, it
seems, goes well beyond the realms of theory or
calculation. "Dr. Blair" was a reference to the
fastest, meanest, shootingest, hard-ridin'est quote
bot that ever rode a horse around these here parts.
Supposedly a Class C player, despite sporting a
PhD. in math.

For some reason, when a player is a math teacher
everyone seems to know it, and over the years I
have played quite a few, many of them excellent
chess players, but certainly not all. My toughest
opponents were of course the professionals, the
players who travel around playing chess for a
living. But after them, I would put the people who
have a burning desire to win, along with enough
talent to keep that from being rendered irrelevant.


> There is a play in which ilyn Monroe really wants to meet Albert and does
> eventually meet him in his hotel room, in New York I think

She needs to be more careful about the men in
her life. Better still, avoid men altogether since,
after all, /diamonds/ are a girl's best friend.


> > Um, no. That is a well-worn path into a deep, dark
> > forest, leading nowhere.
>
> That's good to know; though I get the impression there are a lot of players
> well-treed in that deep dark.

Advanced players would do well to learn the layout
of the forest, but as for beginners, they would do well
to learn not to travel at night, to carry a compass, and
to keep their eyes pealed.


-- help bot






  
Date:
From:
Subject:


 
Date: 04 Oct 2007 23:42:39
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 9:16 am, gaga <[email protected] > wrote:

> > A battleship is infinitely more complex than a car.
> > Therefore, we should all drive... battleships?
>
> I live in Germany. majority if not all of battleships that were
> registered here were sank very fast during traffic accidents. This makes
> me believe that it is dangerous to drive (German) battleships. Besides
> this being Germany i.e. a country of zillion laws would make a life of
> battleship driver a hell.


Look, what police officer is going to try and pull you
over if you run a red light, "roll through" a stop sign, or
pass on the right? The real problem is the taxes and
of course, refueling (do you have any idea what it costs
to buy three metric tonnes of Uranium238 pellets?). Oh,
and you can just forget about finding a place to park in
the city.


-- help bot




 
Date: 04 Oct 2007 22:36:46
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 2:11 pm, james <[email protected] > wrote:
> David Richerby a =E9crit :> C'mon, guys. Anything cross-posted between r=
ec.games.chess.* and
> > rec.games.go is almost certainly a troll. Just leave it alone
> > already.
>
> > Once one person's pointed out that Go and Go-Moku are totally separate
> > games, there really isn't anything else to say.
>
> > Dave.
>
> Right.

You missed jb's reply, which essentially agreed with me (see below).
jb is one of the pioneers of chess programming and AI--and he agrees
with part of what I said. Who are YOU? (a nobody, that's who).

RL

On your point concerning "decision-complexity per move" my jury
is out.
There are several methods for obtaining a measure of this: the
phrase
refers to "state-of-the-art" heuristics. On balance, if
tournament prize
money were equal, number of rounds equal, number of players
equal,
then the fewer moves in Chess appear to argue for "higher
decision
complexity per move" because we acknowledge that Chess is
sharper.
We may also state that, beyond the opening, few Chess
continuations
appear obvious while, at times, Go presents rather obvious
urgencies.



  
Date: 06 Oct 2007 19:56:23
From: Roy Schmidt
Subject: Re: Computer Science Experts: Chess 'harder' than Go
"raylopez99" <[email protected] > wrote:

> jb is one of the pioneers of chess programming and AI--and he agrees
> with part of what I said.

HAHAHAHAHAHAHAHA OOH HAHAHAHAHAHA
Mr. Lopez, you have been trolled by the master troller (i.e., jb).

And on top of that, you thumbed your nose at one of the top researchers in
AI: http://www.recherche.enac.fr/~alliot

Although _Science_ is a famous publication, sometimes a lemon gets into
their pages. Remember the "cold fusion" debacle? I am sure everyone would
grant that chess is deeper than go-moku. Just because go-moku uses the same
equipment as igo/weiqi/baduk does not make it the same game. The authors of
that article just revealed their ignorance by assuming the games were
identical.

Ta!

Roy

--
my reply-to address is gostoned at insightbb dot com
------
The Bradley Go Association meets every Wednesday evening at Kade's Coffee on
War Memorial Drive (opposite the Target/Cub Food/Lowe's/Best Buy center).




   
Date: 07 Oct 2007 18:50:44
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Sat, 6 Oct 2007 19:56:23 -0500, "Roy Schmidt"
<[email protected] > wrote:

>Although _Science_ is a famous publication, sometimes a lemon gets into
>their pages. Remember the "cold fusion" debacle? I am sure everyone would
>grant that chess is deeper than go-moku. Just because go-moku uses the same
>equipment as igo/weiqi/baduk does not make it the same game. The authors of
>that article just revealed their ignorance by assuming the games were
>identical.

No, the Science article just didn't say what Ray said it said.

-- Roy L


   
Date: 07 Oct 2007 13:08:30
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go



Roy Schmidt wrote:

>Although _Science_ is a famous publication, sometimes a lemon gets into
>their pages. Remember the "cold fusion" debacle? I am sure everyone would
>grant that chess is deeper than go-moku. Just because go-moku uses the same
>equipment as igo/weiqi/baduk does not make it the same game. The authors of
>that article just revealed their ignorance by assuming the games were
>identical.

It was our resident troll who confused the two. The chances of
the creators of Chinook mistaking Go-moku for Go are vanishingly
small.

--
Guy Macon
<http://www.guymacon.com/ >




   
Date: 07 Oct 2007 05:08:15
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

> "raylopez99" <[email protected]> wrote:
>> jb is one of the pioneers of chess programming and AI--and he agrees
>> with part of what I said.

"Roy Schmidt" <[email protected] > wrote:
> HAHAHAHAHAHAHAHA OOH HAHAHAHAHAHA
> Mr. Lopez, you have been trolled by the master troller (i.e., jb).


Killfiler "Roy Schmidt" hasn't researched any background material.
Though Roy won't need to pay attention to anything he killfiles ...



> And on top of that, you thumbed your nose at one of the top researchers in
> AI: http://www.recherche.enac.fr/~alliot


He merely neglected to acknowledge that "james" claimed to be that guy.



> Although _Science_ is a famous publication, sometimes a lemon gets into
> their pages. Remember the "cold fusion" debacle? I am sure everyone would
> grant that chess is deeper than go-moku. Just because go-moku uses the same
> equipment as igo/weiqi/baduk does not make it the same game. The authors of
> that article just revealed their ignorance by assuming the games were
> identical.


Par for the course, Roy hasn't read the _Science_ article nor much of
anything that its authors have written. Pretty damn sad if you ask me.




> Ta!


Smelly fart from Roy ?



> my reply-address is got-stoned at insightbb dot com


Thanks Roy, you're definitely a business role-model.




- regards
- jb

------------------------------------------------------------------
The Blackwater Massacre
http://www.wvwnews.net/story.php?id=1962
------------------------------------------------------------------



  
Date: 05 Oct 2007 08:45:28
From: james
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 a �crit :
> On Oct 4, 2:11 pm, james <[email protected]> wrote:
>> David Richerby a �crit :> C'mon, guys. Anything cross-posted between rec.games.chess.* and
>>> rec.games.go is almost certainly a troll. Just leave it alone
>>> already.
>>> Once one person's pointed out that Go and Go-Moku are totally separate
>>> games, there really isn't anything else to say.
>>> Dave.
>> Right.
>
> You missed jb's reply, which essentially agreed with me (see below).
> jb is one of the pioneers of chess programming and AI--and he agrees
> with part of what I said. Who are YOU? (a nobody, that's who).
>
http://www.recherche.enac.fr/~alliot


 
Date: 04 Oct 2007 22:34:16
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 9:09 am, [email protected] (-) wrote:
> raylopez99 <[email protected]> wrote:
> > Were you trying to say something, or just teach us definitions? =20
> > THe latter I presume.

Well jb, it seems you know this area better than I; my appologies but
of course this thread started out as a flame.

I did check out Mueller et al.'s works (three of them), thanks for the
cite, and it seems, much to my surprise and delight, that indeed alpha-
beta algorithm is used in GO, see "we incorporate a simple alpha-beta
algorithm" below (copy and paste from inside Adobe--the repetitions
are unintentional):

**
Reinf Reinfor orcement cement Lear Learning ning of Local Shape in the
Game of Go
Da David vid Silv Silver er, Richard Sutton, and tin M=A8uller
Department of Computing Science
Uni University ersity of Alberta

In this paper paper, we return to the strate strategy gy that has been
so
successful in other domains, and apply it to Go. We de develop elop
a systematic approach for representing intuiti intuitive Go kno knowl-
wledge
edge using local shape features. We evaluate aluate positions using
a linear combination of these features, and learn weights by
temporal dif difference ference learning and self-play play. Finally
Finally, we incor incor-
porate a simple alpha-beta search algorithm.
**

Essentially, there is a crack in the GO facade of impenetrability--if
trees exist in GO, GO can be solved like chess, since computers are
good at solving trees (much better than pattern recognition it seems).


**
On your point concerning "decision-complexity per move" my jury
is out.
There are several methods for obtaining a measure of this: the
phrase
refers to "state-of-the-art" heuristics. On balance, if
tournament prize
money were equal, number of rounds equal, number of players
equal,
then the fewer moves in Chess appear to argue for "higher
decision
complexity per move" because we acknowledge that Chess is
sharper.
We may also state that, beyond the opening, few Chess
continuations
appear obvious while, at times, Go presents rather obvious
urgencies.
**

"CHESS IS SHARPER"--this is what I was driving at. I've not played GO
except on a Palm with a small 9x9 board, only a few times, but I
somewhat got the feel each move is not absolutely critical, but it's
the accumulation of moves (of course the same is true in chess, but to
a lesser degree--think how often have you 'blown' a winning position
by making a one-move blunder, meaning every move 'packs more punch'
than in Go). But chess had a more tactical feel than (and chinese
chess has an even more tactical feel than chess).

Rest of your points are well taken.

It does seem that Go has more "complexity" for solving than chess--the
evolution arguments support this. But one could argue that passing
the Turning test by imitating a six year old human baby is also more
"complex" than Go--but I don't think, from a recreational point of
view, that imitating a six year old is fun (except in this NG as this
login), and thus I prefer chess over more "complex" games like GO or
imitating schoolchildren. I do have a 18?x18 board and stones that I
intend to at some point learn, but unfortunately there's not much
computer help in GO as there is in chess, which hinders learning the
game.

Flame on!

RL



  
Date: 05 Oct 2007 06:45:48
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

raylopez99 <[email protected] > wrote:
> Reinfororcement Learning of Local Shape in the Game of Go
> David Silver, Richard Sutton, and tin Muller
>
> [ ... ]
>
> Essentially, there is a crack in the GO facade of impenetrability--if
> trees exist in GO, GO can be solved like chess, since computers are
> good at solving trees (much better than pattern recognition it seems).


You're walking up on a discussion which periodically washes
over this newsgroup. A lot of people have "been there, done that."
Gaming analyzes strong and weak, and there are inroads by computers.
Yet, among the three scales for Go Ranks -- kyu, amateur dan, pro dan --
known computer programs have not yet graduated beyond the kyu scale.
Kyu level is through high-school, amateur dan is college level, and
pro scale is competitve play at a very high level, to teach college.
Despite some very intense computer science efforts by the best among
the best, the Go Programs do not even have a high school diploma.



> "CHESS IS SHARPER"--this is what I was driving at. I've not played
> GO except on a Palm with a small 9x9 board, only a few times, but I
> somewhat got the feel each move is not absolutely critical, but it's
> the accumulation of moves (of course the same is true in chess, but to
> a lesser degree--think how often have you 'blown' a winning position
> by making a one-move blunder, meaning every move 'packs more punch'
> than in Go). But chess had a more tactical feel than (and chinese
> chess has an even more tactical feel than chess).
>
> ... thus I prefer chess over more "complex" games like GO or
> imitating schoolchildren. I do have a 18x18 board and stones that I
> intend to at some point learn, but unfortunately there's not much
> computer help in GO as there is in chess, which hinders learning the
> game.


The Japan Go Academy (Nihon Ki-In) is sometimes rendered by
machine translators as "the Chessman Institute." In practice most
Asians are not quite so much interested in competitive distinctions
among their many varieties of cognitive-skill boardgames. Games
each exist in their own right -- forms of mental sports tial arts --
and obtain intrinsic respect in that regard. In abstract, computers
don't care much what sort of gaming problem they're really solving.
If one were to examine code the initial appearance of a game gets
lost pretty quickly among those artificial intelligence entanglements.



"raylopez99" <[email protected] > wrote in message
> jb is one of the pioneers of chess programming and AI--and he
> agrees with part of what I said. Who are YOU? (a nobody, that's who).


I have been internet-cited on a dispute regarding copyrights and
game programming, and am published for some math conjectures,
but I am not even a minor contributor for game programming A.I.

Luckily I obtained a copy of the _Science_ paper today. It is not
particularly newsworthy but reflects much of the same information as
was on record 3-to-5 years ago. Authors delayed publication while
they conducted lint-checking and sanity-checking, and wordsmithed
it for _Science_. The game of Go is mentioned nowhere in the article:
a sign to me which suggests that the authors intended to distance this
"solvability" concept from the Go game as much as they could. So
"raylopez99" breached the very wall that esteemed _Science_ authors
tried to carefully build, and paraded before our newsgroup luminaries
his unwarranted example of intellectual dishonesty and arrogance.

The _Science_ paper is surely a vellous coordination of geeks
arriving at "one of the longest running computations completed" in the
history of computer proofs. One wonders why it might be so difficult
to accept that best play on both sides produces a draw in Checkers.
Powerful forces sought to instill the notion that either one side or the
other "must have" an irrefutable advantage. Against those powerful
forces their comptuers labored day-and-night for more than 15 years
to establish these ceaseless facts of dynamical balance in Checkers.
According to the publication rules at _Science_ magazine us ordinary
folks must wait one year before electronic document text is released.
Only then may we ingest their stale report without a trip to the library.

More author-names are listed, than were supplied by the website
record of abstract: Jonathan Schaeffer, Neil Burch, Yngvi Bj�rnsson,
Akihiro Kishimoto, tin M�ller, Robert Lake, Paul Lu, Steve Sutphen.
A "google" search on everything these guys have done might save
all of us needless muck-churning and then we can carry-on with this
ionospheric discourse already so familiar to the researchers spotlighted.




- regards
- jb

-----------------------------------------------------------------------
New revelations in attack on American spy ship
http://www.chicagotribune.com/news/nationworld/chi-liberty_tuesoct02,0,66005.story?coll=chi_tab01_layout
-----------------------------------------------------------------------



 
Date: 04 Oct 2007 08:18:00
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 3:43 am, "Dirt At Your Face" <[email protected] >
wrote:

> Modern chess has no pedigree. The chess as we know it now matured from
> generations of "token" improvisation.
>

That's mighty stupid of you.

> Chess as purely tactical game has only strict moves-order-sequence. Familiarity
> with these sequence thru rigorous tactical training does not give you harder
> "per move". Tactic in Go is different as you still need to consult the
> surrounding conditions if your known "go sequence" still yield favorable result.

This would be true if chess was all book (I trust you know this term)
and GO had no book. But you own post admits that even Go has "opening
sequences" (book). Fischer, a famous chessmaster, complained of chess
book and invented a form of chess that helps eliminate book (Fischer
Random Chess), but most people do not believe, like you apparently do,
that chess is all book; nor does science say so.

>
> There is no obfuscation and confusion in go.

Of course there is--feigns are classic in game theory.


> Chess has literature made for specific readers where the game is popular.
> Likewise how much oriental books on go do you know?

None. I don't speak oriental. How many books do you go?

RL




 
Date: 04 Oct 2007 08:14:41
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 6:59 am, Guy Macon <http://www.guymacon.com/ > wrote:
> - wrote:
> Good point.
>
> Poker has three possible moves at each point. Roulette has 87.
> Does that make Poker a simpler or easier game?
>
> --

Since you don't know the answer (it's clear this question is not
retorical), I'll provide an answer as per Shannon's theories (Shannon
was also a chess player and chess programming pioneer).

Roulette has more data, but less information, because data is random
but information is ordered. So Poker is more complex.

RL




 
Date: 04 Oct 2007 08:08:19
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 12:50 am, [email protected] (-) wrote:

Erudite trash; rebutted below.

> > raylopez99 wrote:
> >> In an article in Science, 14 September 2007 p. 1519, written by the
> >> Chinook team that weakly solved the game of checkers (draughts), it's
> >> asserted essentially that chess is harder than Go.
>
> >> What is meant is that the "decision complexity" of chess (and
> >> checkers) is higher than for Go (actually Go-Moku was mentioned, which
> >> is not exactly the same as Go, but I take it the analogy holds for
> >> Go), while the game of Go has a higher "space complexity" (essentially
> >> non-repeating patterns--more such patterns in Go than Chess/Checkers).

> Jonathan Buss <[email protected]> wrote:
> > Go has at least the same decision complexity as chess, and possibly gre=
ater.
>
> Surely it's a mistake by "raylopez99" to draw inferences about Go fr=
om
> article's mention of Go-Moku. Go-Moku, without initial starting rul=
es,
> was "solved" by Victor Allis in 1993. A "definition of terms" avail=
able:
>
> "The decision complexity of a game is the number of leaf nodes in the
> smallest decision tree that establishes the value of the initial posit=
ion."
> http://en.wikipedia.org/wiki/Game-tree_complexity
>
> Once paring away the Greek, this sort of definition can follow from =
an
> alpha-beta simplification, or as by labelling massive branches a win=
for
> Player A or Player B. Game-tree complexity uses "full-width" search,
> a slightly different concept, though a key difference to keep in min=
d=2E


Thanks. Were you trying to say something, or just teach us
definitions? THe latter I presume.


>
> It's not easy to obtain the article, in Microsoft Territory, so I se=
ttled
> for an abstract, supporting online material, a few google references.
> Here's what Jonathan Schaeffer actually said, about Chess & Go:
>
> at: http://www.aaai.org/AITopics/html/go.html
>
> "If you want to understand intelligence, the game of Go is much
> more demanding ... It doesn't have the silver bullet: deep search.
> Chess has somewhat outlived its usefulness. It turned out to be
> easier than we thought.
> - Jonathan Schaeffer, in Chess - Man vs. Machine Plays Out
> http://www.aaai.org/AITopics/html/go.html#wired


Yeah, I think this same guy, who is prominent in AI research, said
"bridge" is the next chess. You wanna play 'bridge' with grandpa's
foursome, to test your mind?


>
> Our particular _Science_ article lists a -team- of author-names:
> Jonathan Schaeffer, Akihiro Kishimoto, tin Muller, Robert Lake ...
> Crucial to recognize tin Muller as a longstanding contributor to =
Go
> game algorithms, programming, and theorem-proving. That _Science_
> article was not intended to become some polemic for slandering Go.

No, but my polemic was. And what exactly is M. Muller's contribution
to go? Does GO have a algorithm? Or just lame pattern recognition?

>
> > (-) wrote:
> >> ... we could examine the sociological aspect: what led computer
> >> scientists first to attack Chess instead of Go? Was there a perception
> >> that Chess is more highly respected as a decision-rendering tool for
> >> purposes of conducting our state-of-affairs in World Diplomacy?
> "raylopez99" <[email protected]> wrote:
> > A short answer to a rather deep compound question: YES.
>
> Thank you for answering. Still, I'm rather curious how that percept=
ion
> plays out. In what way are Chess Players exerting a positive influe=
nce
> upon contemporary World Diplomacy? =20
>

Read up on the literature stoopid. "Pawns" as a metaphor for puppet
rulers is one.


> >> As the Science Magazine authors note, a solved game may fall due
> >> to low decision -OR- low space complexity. The decision side is more
> >> difficult to visualize, and to devise in hardware, however it remains
> >> true that the impenetrability of a game to machine analysis is the
> >> result of a combination of both and not simply one aspect or the other.
> > Agreed. But you have to acknowledge that chess has a higher decision
> > complexity per move. THis is what the paper said. It's pretty
> > obvious too--alpha-beta takes thought (whether in wetware or
> > software), while moving little white stones around is child's play
> > (pattern recognition, not unlike a baby recognizing her mother's
> > face).
>
> Let's follow-up with a definition for "space complexity." This=
is:
>
> "The state-space complexity of a game is the number of legal game
> positions reachable from the initial position of the game.[1]

Stay on topic shiitz for brains. The topic is decision complexity.


>
> "When this is too hard to calculate, an upper bound can often be
> computed by including illegal positions or positions that can never
> arise in the course of a game."
> - http://en.wikipedia.org/wiki/Game-tree_complexity
>

What a wiki-idiot.

> John Tromp and Gunnar Farneb=E4ck have conducted research
> on the state-space complexity of Go, periodically publishing their
> preliminary findings to rec.games.go for cases of smaller boards.
> I must wait on the text of the article, which I haven't before me, b=
efore
> commenting on your claims concerning what the paper said. =20

OK, please cross-post to rec.games.chess and/or email me. I'd like to
build, for fun, a simple "go" playing program, like I've done for
chess. BTW the paper I quoted was on weakly solving draughts
(checkers), not Go.


> We think
> that this _Science_ paper was comparing Go-Moku but not Go. The
> decision-complexity definition supplied by Wikipedia refers to the
> calculation of the value of a position. Because Go Positions must
> be calculated within narrow gins there is much opportunity for
> shifting flim-flams about each convergent _komi_ value of the game.
> Unlike Chess the calculation of a positional values in Go does not
> reduce to +/- for either player so quickly. So it is quite likely t=
hat
> decision-complexity for Go is also higher than would be for Chess.

"Quite likely" is an opinion. I beg to differ.


> When one game has both higher decision-complexity -AND- higher
> space-complexity then you may infer an overall ordering relation.
>

Right. So the jury is still out. I say GO is for simpletons. You're
entitled to your *opinion*, and, like an arse hole, everybody has an
opinion. Opinion is not fact.


> This result, however, is deceptive because it argues from the
> definition of a term, and these term-definitions do not lead us towa=
rd
> much solvability direction in Go. It would be more productive to no=
te
> a positional value distribution of candidate moves, perhaps bell-cur=
ve
> shaped, and whether there was any flattening near the selection for
> optimal move candidates. If, in Go, I can weed out candidate moves
> to two dozen and then posit a half-dozen good moves which continue
> forward from a position then perhaps my selection problem reduces
> to a likelihood of 1/4th probability. Let's say, though, in Chess I=
find
> six candidate moves after weeding, but owing to the "sharpness" of
> Chess only one of those moves is really "best" and "worth playing."
> In this sort of thought-experiment my selection problem is then 1/6t=
h,
> offering me less opportunity of finding a good forward continuance,
> meaning that a Chess task would be more difficult than the Go task.

Well said. Sorry for the previous insults if you wrote this
yourself. Again, I'm not familiar with GO and what algorithms they
use in GO (if any), but I'd be curious to find out. It sounds a
little like the game of LIFE, which is simple to program; a lot
simplier than chess. Ergo, GO is simple.


>
> In either case the machine task is to play fairly well though n=
ot
> brilliantly, just maintaining a balance, and then winning the endgam=
e=2E
> Whether it's Chess or Go, the machine is dominant for the endgame.
> Yet, for Go, perhaps it's a more difficult task just to keep a balan=
ce.
> Playing "approximately" would never suffice to raise the level of Go
> machine-play toward pro level. =20

No, I disagree; they said the same thing about chess. If GO is a game
of blunders, like chess is, then playing approximately is good enough.


>
> Playouts in MonteCarlo could be full-width or alpha-beta pruned.
> With additional pattern-recognition a pruning might reduce awesome
> decision-complexity into something more manageable: this was one
> of Frank de Groot's visions. Recall that the machine task is not to=
be
> an overwhelming player but simply to find humdrum game balance
> in order to apply CGT during _yose_. Prior to the CGT there may be
> some sort of extensible pre-CGT phase which initiates the lead-out.
> The problem concerns "sente" moves which are, typically, not in a
> general vicinity of the move recently played. The phrase "not in a
> general vicinity" tends to widen, not restrict, the candidate search.

Did you cut and paste this from somewhere or are you trying to say
something? What an idiot (sorry for my premature retraction of
calling you an idiot before).

>
> Beyond a threshold, more playouts do not improve move selection.
> The hardware I was attempting to describe is not a MC/UCT scheme,
> however. Would be hardware from something like Hogwarts, actually,
> where the particular configurations of switches, levers and tripwire=
s,
> were already calibrated for zapping electrons into proper pigeonhole=
s=2E
> Could be very fast, especially dart-like pieces that do not move aro=
und.
> The "moving around" aspects of Chess could later (future) prove to be
> more intractably tenacious than the "sitting around" aspects for Go.
>

Metaphysics. You'd make a good high priest, or maybe a theoretical
physicist.

RL



  
Date: 04 Oct 2007 16:09:37
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

raylopez99 <[email protected] > wrote:
> Were you trying to say something, or just teach us definitions?
> THe latter I presume.


Crucial to obtain definitions because you have been demonstrating
an abysmal ignorance of the various complexity measures and how they
each contribute to our understanding of what renders a game complex.



> (-) wrote:
>> Here's what Jonathan Schaeffer actually said, about Chess & Go:
>> at: http://www.aaai.org/AITopics/html/go.html
>> "If you want to understand intelligence, the game of Go is much
>> more demanding ... It doesn't have the silver bullet: deep search.
>> Chess has somewhat outlived its usefulness. It turned out to be
>> easier than we thought.
>> - Jonathan Schaeffer, in Chess - Man vs. Machine Plays Out
>> http://www.aaai.org/AITopics/html/go.html#wired

> Yeah, I think this same guy, who is prominent in AI research, said
> "bridge" is the next chess. You wanna play 'bridge' with grandpa's
> foursome, to test your mind?


I don't suppose this topic concerns what anybody wants to play.
We were attempting to discuss a relative complexity measure of games.
Not whether some game is easy to play but whether it is easy to win.



>> Our particular _Science_ article lists a -team- of author-names:
>> Jonathan Schaeffer, Akihiro Kishimoto, tin Muller, Robert Lake ...
>> Crucial to recognize tin Muller as a longstanding contributor to Go
>> game algorithms, programming, and theorem-proving. That _Science_
>> article was not intended to become some polemic for slandering Go.

> No, but my polemic was. And what exactly is M. Muller's contribution
> to go? Does GO have a algorithm? Or just lame pattern recognition?


Pattern recognition has been applied for Go, among many algorithms.
You are invited to "google" search on tin Muller for further details.



>> In what way are Chess Players exerting a positive influence
>> upon contemporary World Diplomacy?

> Read up on the literature stoopid. "Pawns" as a metaphor for
> puppet rulers is one.


Though, in many games, a "pawn" becomes the deciding piece.
Not exactly your "puppet" anymore ... :-)



>>>> As the Science Magazine authors note, a solved game may fall due
>>>> to low decision -OR- low space complexity. The decision side is more
>>>> difficult to visualize, and to devise in hardware, however it remains
>>>> true that the impenetrability of a game to machine analysis is the
>>>> result of a combination of both and not simply one aspect or the other.

>>> Agreed. But you have to acknowledge that chess has a higher decision
>>> complexity per move. THis is what the paper said. It's pretty
>>> obvious too--alpha-beta takes thought (whether in wetware or
>>> software), while moving little white stones around is child's play
>>> (pattern recognition, not unlike a baby recognizing her mother's
>>> face).

>> Let's follow-up with a definition for "space complexity." This is:
>> "The state-space complexity of a game is the number of legal game
>> positions reachable from the initial position of the game.[1]

> Stay on topic shiitz for brains. The topic is decision complexity.


The topic is what renders a game impenetrable, versus amenable, to
machine analysis. Chess has fallen by the wayside. Go has not (yet).
Earlier you "agreed" that it is a combination of both, not one aspect.

On your point concerning "decision-complexity per move" my jury is out.
There are several methods for obtaining a measure of this: the phrase
refers to "state-of-the-art" heuristics. On balance, if tournament prize
money were equal, number of rounds equal, number of players equal,
then the fewer moves in Chess appear to argue for "higher decision
complexity per move" because we acknowledge that Chess is sharper.
We may also state that, beyond the opening, few Chess continuations
appear obvious while, at times, Go presents rather obvious urgencies.

This consideration for "space complexity" explains why alpha-beta
pruning for Go is itself more intractable than for Chess, offering much
evidence for why a "decision complexity" in Go cannot be so reduced
as might be possible for Chess. Retrograde analysis in Chess/Checkers
allows for closure when approaching the endgames. Computer Go is
closed by Combinatorial Game Theory, not by any retrograde analysis.
Here also, CGT rapidly falls apart much sooner in retrograde direction.



>> "When this is too hard to calculate, an upper bound can often be
>> computed by including illegal positions or positions that can never
>> arise in the course of a game."
>> - http://en.wikipedia.org/wiki/Game-tree_complexity

> What a wiki-idiot.


Which argued in your favor. The "actual proportional" state-space
complexity of Go falls off more rapidly when excluding illegal positions.



> I'd like to build, for fun, a simple "go" playing program, like I've done
> for chess. BTW the paper I quoted was on weakly solving draughts
> (checkers), not Go.


Yes, and his "weak solution" resorted to mathematical arguments.
To whet your whistle, I built a working Chess program as early as 1973.
I do not yet have an adequate Go program.



>> We think
>> that this _Science_ paper was comparing Go-Moku but not Go. The
>> decision-complexity definition supplied by Wikipedia refers to the
>> calculation of the value of a position. Because Go Positions must
>> be calculated within narrow gins there is much opportunity for
>> shifting flim-flams about each convergent _komi_ value of the game.
>> Unlike Chess the calculation of a positional values in Go does not
>> reduce to +/- for either player so quickly. So it is quite likely that
>> decision-complexity for Go is also higher than would be for Chess.

> "Quite likely" is an opinion. I beg to differ.


And who might you be? While game complexity is not necessarily
locked in chronological discovery with computer science evolution
perhaps there is, nonetheless, a compelling argument for correlation.



>> When one game has both higher decision-complexity -AND- higher
>> space-complexity then you may infer an overall ordering relation.

> Right. So the jury is still out. I say GO is for simpletons. You're
> entitled to your *opinion*, and, like an arse hole, everybody has an
> opinion. Opinion is not fact.


The authors of the cited _Science_ article say Go has higher
decision complexity -AND- higher space-complexity overall. Your
discussion concerning "decision complexity per move" is the point
you wish to advance. I partially agree with you that Chess is sharper.
We all support bringing your supercomputer powers to a Go question.



>> In either case the machine task is to play fairly well though not
>> brilliantly, just maintaining a balance, and then winning the endgame
>> Whether it's Chess or Go, the machine is dominant for the endgame.
>> Yet, for Go, perhaps it's a more difficult task just to keep a balance.
>> Playing "approximately" would never suffice to raise the level of Go
>> machine-play toward pro level.

> No, I disagree; they said the same thing about chess. If GO is a game
> of blunders, like chess is, then playing approximately is good enough.


Unlike Chess, Go is not a game of blunders. The relative "dullness"
of Go means that compensation for mistakes does not always go away.



>> The problem concerns "sente" moves which are, typically, not in a
>> general vicinity of the move recently played. The phrase "not in a
>> general vicinity" tends to widen, not restrict, the candidate search.

> Did you cut and paste this from somewhere or are you trying to say
> something?


A centrality of sente, in Go, prevents reduction of decision-complexity,
because we reduce space-complexity for reducing decision-complexity.



- regards
- jb

-------------------------------------------------------------
Arrested For Reading The Constitution
http://www.prisonplanet.com/articles/october2007/021007_reading_constitution.htm
-------------------------------------------------------------



 
Date: 04 Oct 2007 13:10:13
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go




############################################################
# #
# TROLL FEEDING IS STRICTLY PROHIBITED. THIS MEANS YOU! #
# #
###############(\##########################/)###############
((_)) ((_))


 
Date: 04 Oct 2007 12:59:40
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go



raylopez99 wrote:

>Go has a smaller decision complexity per move than chess.
>One way of seeing this is to count the moves in an 'average'
>game of GO (to completion) versus an 'average' game of chess.
>In chess, good players finish in around 40 moves. I doubt
>the same can be said for Go. Chess packs more per move
>than Go.

Slighly better trolling, but still not very good.

So if I play you a game of high card (we each draw a single card
from a deck, high card wins all) that would have more "decision
complexity per move" than if we spent all night playing poker?

And if I started a chess game and resigned on the first move
that would have more "decision complexity per move" than if I
slugged it out for 60 or 80 moves?

-----------------------------------------------------------------------

Troll

1. v.,n. [From the Usenet group alt.folklore.urban] To utter a posting
on Usenet designed to attract predictable responses or flames; or, the
post itself. Derives from the phrase "trolling for newbies" which in
turn comes from mainstream "trolling", a style of fishing in which one
trails bait through a likely spot hoping for a bite.
The well-constructed troll is a post that induces lots of newbies and
flamers to make themselves look even more clueless than they already do,
while subtly conveying to the more savvy and experienced that it is in
fact a deliberate troll. If you don't fall for the joke, you get to be
in on it. See also YHBT.

2. n. An individual who chronically trolls in sense 1; regularly posts
specious arguments, flames or personal attacks to a newsgroup,
discussion list, or in email for no other purpose than to annoy someone
or disrupt a discussion. Trolls are recognizable by the fact that they
have no real interest in learning about the topic at hand - they simply
want to utter flame bait. Like the ugly creatures they are named after,
they exhibit no redeeming characteristics, and as such, they are
recognized as a lower form of life on the net, as in, "Oh, ignore him,
he's just a troll."

Some people claim that the troll (sense 1) is properly a narrower
category than flame bait, that a troll is categorized by containing
some assertion that is wrong but not overtly controversial.

The use of `troll' in either sense is a live metaphor that readily
produces elaborations and combining forms. For example, one not
infrequently sees the warning "Do not feed the troll" as part of
a followup to troll postings.

See also Kook.

-----------------------------------------------------------------------




--
Guy Macon
<http://www.guymacon.com/ >



 
Date: 04 Oct 2007 14:53:22
From: Robert Jasiek
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 07:15:26 -0700, raylopez99 <[email protected] >
wrote:
>In an article in Science, 14 September 2007 p. 1519, written by the
>Chinook team that weakly solved the game of checkers (draughts), it's
>asserted essentially that chess is harder than Go.
>
>What is meant is that the "decision complexity" of chess (and
>checkers) is higher than for Go [...]

From the followup posts I conclude that here "decision complexity"
(and "harder") shall mean "(average or maximal?) number of next legal
moves".

>What is the consequence of this?

The consequence is (BTW, not for the first time) that (some) articles
in Science about game complexity deceive their readers.

With measures of game complexity, one should estimate how hard it is
for a human and / or computer to find the best move(s) throughout the
game. Contrarily, the cited complexity measure does not do this
because it cannot be decided what is (one of) the best move(s) by
looking ahead only one move (unless we are already just one move
before the game end with whichever might be the next legal move).


  
Date: 04 Oct 2007 14:28:13
From: David Richerby
Subject: Re: Computer Science Experts: Chess 'harder' than Go
Robert Jasiek <[email protected] > wrote:
> The consequence is (BTW, not for the first time) that (some)
> articles in Science about game complexity deceive their readers.

Not really. The article said `Go-Moku' and the reader didn't know the
difference between Go-Moku and Go. Hardly _Science_'s fault in this
case.


Dave.

--
David Richerby Portable T-Shirt (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ fashion statement but you can take
it anywhere!


 
Date: 04 Oct 2007 10:48:06
From: David Richerby
Subject: Re: Computer Science Experts: Chess 'harder' than Go
C'mon, guys. Anything cross-posted between rec.games.chess.* and
rec.games.go is almost certainly a troll. Just leave it alone
already.

Once one person's pointed out that Go and Go-Moku are totally separate
games, there really isn't anything else to say.


Dave.

--
David Richerby Gigantic Toy (TM): it's like a fun
www.chiark.greenend.org.uk/~davidr/ child's toy but it's huge!


  
Date: 05 Oct 2007 00:54:00
From: Kenneth Sloan
Subject: Re: Computer Science Experts: Chess 'harder' than Go
newsgroups trimmed to r.g.c.misc

David Richerby wrote:
> C'mon, guys. Anything cross-posted between rec.games.chess.* and
> rec.games.go is almost certainly a troll. Just leave it alone
> already.
>
> Once one person's pointed out that Go and Go-Moku are totally separate
> games, there really isn't anything else to say.
>
>
> Dave.
>


You mean, like YOUR posting, which was cross-posted between
rec.games.chess.* and rec.games.go???

--
Kenneth Sloan [email protected]
Computer and Information Sciences +1-205-932-2213
University of Alabama at Birmingham FAX +1-205-934-5473
Birmingham, AL 35294-1170 http://www.cis.uab.edu/sloan/


   
Date: 05 Oct 2007 10:21:16
From: David Richerby
Subject: Re: Computer Science Experts: Chess 'harder' than Go
Kenneth Sloan <[email protected] > wrote:
> David Richerby wrote:
>> C'mon, guys. Anything cross-posted between rec.games.chess.* and
>> rec.games.go is almost certainly a troll. Just leave it alone
>> already.
>
> You mean, like YOUR posting, which was cross-posted between
> rec.games.chess.* and rec.games.go???

Individual cases are for the reader to decide.


Dave.

--
David Richerby Pointy-Haired Whisky (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a single-malt whisky that's completely
clueless!


  
Date: 04 Oct 2007 23:11:21
From: james
Subject: Re: Computer Science Experts: Chess 'harder' than Go
David Richerby a �crit :
> C'mon, guys. Anything cross-posted between rec.games.chess.* and
> rec.games.go is almost certainly a troll. Just leave it alone
> already.
>
> Once one person's pointed out that Go and Go-Moku are totally separate
> games, there really isn't anything else to say.
>
>
> Dave.
>
Right.


 
Date: 04 Oct 2007 00:44:35
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 10:02 pm, "Dirt At Your Face" <[email protected] >
wrote:

>
> http://img167.imageshack.us/img167/700/chessgeniuslt4.jpg
>
> I also have the image made into a nice wall-poster minus the text-message of
> course. How much would you reckon the poster would sell?

Of course, "minus" the text message, which belies your subliminal
message.

Face it--chess has pedigree. Chess is harder "per move". Chess
doesn't relying on obfuscation and confusion (asian specialties) to
make the game hard--it's simple Mandelbrot chaos theory embodied in a
few simple rules. Chess has literature and lots of people playing it,
even younger people today.

WHat does Go have? Nothing but apeish pattern recognition--you might
as well play bridge with the old folks. Further, I 'go' to the "GO"
board and it's pretty dead--dead board, dead game!

'Nuff said.

RL

"YOu can learn a lot from a dummy" <--public service message re
seatbelts in the USA, featuring a test crash dummy




  
Date: 04 Oct 2007 19:58:07
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Thu, 04 Oct 2007 00:44:35 -0700, raylopez99 <[email protected] >
wrote:

>On Oct 3, 10:02 pm, "Dirt At Your Face" <[email protected]>
>wrote:
>
>> http://img167.imageshack.us/img167/700/chessgeniuslt4.jpg
>>
>> I also have the image made into a nice wall-poster minus the text-message of
>> course. How much would you reckon the poster would sell?
>
>Of course, "minus" the text message, which belies your subliminal
>message.
>
>Face it--chess has pedigree.

About 1/5 of go's pedigree.

>Chess is harder "per move".

Translation: because it is so completely tactical, it is harder to
avoid losing a game of chess with any given move.

>Chess doesn't relying on obfuscation and confusion (asian specialties) to
>make the game hard--it's simple Mandelbrot chaos theory embodied in a
>few simple rules.

But far more complicated rules than go's.

>Chess has literature and lots of people playing it,
>even younger people today.

As does go.

>WHat does Go have? Nothing but apeish pattern recognition--you might
>as well play bridge with the old folks.

Obvious nonsense.

>Further, I 'go' to the "GO"
>board and it's pretty dead--dead board, dead game!

Brain-dead...

-- Roy L


  
Date: 04 Oct 2007 18:43:56
From: Dirt At Your Face
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 <[email protected] > typed:
> On Oct 3, 10:02 pm, "Dirt At Your Face" <[email protected]>
> wrote:
>
>>
>> http://img167.imageshack.us/img167/700/chessgeniuslt4.jpg
>>
>> I also have the image made into a nice wall-poster minus the text-message of
>> course. How much would you reckon the poster would sell?
>
> Of course, "minus" the text message, which belies your subliminal
> message.

I think I'll make the poster for mass production. Watch out for it in the coming
months... Would anyone buy it at US$26.00 as a go player?


>
> Face it--chess has pedigree. Chess is harder "per move". Chess
> doesn't relying on obfuscation and confusion (asian specialties) to
> make the game hard--it's simple Mandelbrot chaos theory embodied in a
> few simple rules. Chess has literature and lots of people playing it,
> even younger people today.

Modern chess has no pedigree. The chess as we know it now matured from
generations of "token" improvisation.

Chess as purely tactical game has only strict moves-order-sequence. Familiarity
with these sequence thru rigorous tactical training does not give you harder
"per move". Tactic in Go is different as you still need to consult the
surrounding conditions if your known "go sequence" still yield favorable result.

There is no obfuscation and confusion in go. Its beauty lies within your ability
to construct living shapes (and obstructing your opponent's in a way beneficial
only to your game) efficiently from the empty grid covering a space-size (or
capturing the enemy stones) only limited by your playing skill.

Chess has literature made for specific readers where the game is popular.
Likewise how much oriental books on go do you know?




 
Date: 04 Oct 2007 00:40:22
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 7:14 pm, Ben Finney <[email protected] >
wrote:
> raylopez99 <[email protected]> writes:
> > Go games take more moves than the average chess game to conclude (a
> > chess game takes about 40 moves to conclude).
>
> > Ergo, per move, Go is simplier.
>
> Chess and Go both take many more moves to conclude, on average, than
> the average Tic-Tac Toe game.
>
> Ergo, per move, Go and Chess are both simplier than Tic-Tac-Toe.
>

THis would be true, if the number of moves in Tic-Tac-Toe was not so
low. Go and Chess have roughly the same number of moves (or much
higher than Tic-Tac-Toe), so my "thought per moves" point was well
made.

Thanks for playing with the troll. As a public safety message
featureing a crash test dummy once said in the USA, "you can learn a
lot from an idiot".

RL



  
Date: 04 Oct 2007 19:52:02
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Thu, 04 Oct 2007 00:40:22 -0700, raylopez99 <[email protected] >
wrote:

>On Oct 3, 7:14 pm, Ben Finney <[email protected]>
>wrote:
>> raylopez99 <[email protected]> writes:
>> > Go games take more moves than the average chess game to conclude (a
>> > chess game takes about 40 moves to conclude).
>>
>> > Ergo, per move, Go is simplier.
>>
>> Chess and Go both take many more moves to conclude, on average, than
>> the average Tic-Tac Toe game.
>>
>> Ergo, per move, Go and Chess are both simplier than Tic-Tac-Toe.
>
>THis would be true, if the number of moves in Tic-Tac-Toe was not so
>low. Go and Chess have roughly the same number of moves

The number of moves in a typical go game is at least triple the number
in a typical chess game.

>(or much
>higher than Tic-Tac-Toe), so my "thought per moves" point was well
>made.

No, it was idiotic and wrong.

-- Roy L


 
Date: 03 Oct 2007 23:59:00
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 4, 12:44 am, [email protected] wrote:

> > Many years ago, I played chess with a fellow who
> >insisted that go was a far better game. He continued
> >to insist this night after night, which he wasted playing
> >chess, *not go*! (Presumably, he could find no
> >monkeys against which to play, and this is why he
> >continually pestered us to switch games.)
>
> Maybe he just wanted an opponent he could beat.

Well, in that case, he would have done better to
coerce some of the weaker chess players, boosting
his odds a bit.


> >> BTW, an old friend of mine was one of the early contributors to the
> >> Chinook research effort, and he long ago acknowledged that go is so
> >> much more complex than checkers that he could not even hazard a guess
> >> how much more complex.
>
> > A battleship is infinitely more complex than a car.
> >Therefore, we should all drive... battleships?
>
> Careful, you're giving the logical powers of chess players a bad name.

A lot of talk has centered around which game is the
more complex, chess or go. It seems to me that this
is very similar to asking: who was the faster draw,
Jesse James or Billy the Kid? Which would win in a
fight, a bear or a lion? My point is that complexity is
a poor measure of which is the better game; that such
an approach is rather silly.


> > Another drawback is the very limited playing
> >field -- just 64 squares, along with only two types of
> >men: expendable and not expendable. Kidding. But
> >there are only a relatively few types of men in chess.
> >The "terrain" is perfectly flat;
>
> Not so: the center is effectively the "high ground."


Ah. But then, it is also the high ground in tic-tac-toe;
I want complexly variable terrain, more like the real
world. In the game Stratego there are two obstacles
dividing the different sides from one another, creating
three different possible routes for attack or retreat. I
found even that to be a bit simplistic. In chess there
are no terrain-obstacles; it is wide open, plain and
simple. (One can create obstacles by locking the
pawns, but this requires both sides to cooperate.)



> Go's excellent handicap system effectively allows players of widely
> divergent skill to play on an equal footing without totally changing
> the nature of the game. The weaker player just gets a head start in
> the places where he puts his handicap stones.

An excellent aspect. In chess, if you force a top-GM
to give up his Queen, he will still be a tactical monster,
but all of his vast knowledge of the intricasies of the
chess openings, all of his preparation, goes up in smoke.
None of these guys got to the tip-top by just showing up
and winging it.


> >Where nobody has "prepared
> >lines" against almost any reasonable approach you
> >might try? I think so.
>
> In go, free placement of handicap stones allows the weaker player, if
> he so chooses, to challenge the stronger with an opening position
> unlike any he has seen before.

Chess is much more rigid in nature. Remove my
Queen, and I still "know" how to develop and castle.
I still "know" that it is best to develop in a certain
way, though now I must try to avoid even exchanges.
Handicapping destroys the opening book, except
when the handicap is small (like say, one pawn).



> >I'm tired of the 90+% of chess
> >that is chained to seeing all the tactics, merely
> >calculating captures and checks; give me more
> >deep strategy, more long-range planning and less
> >check-and-pea-pick-that-pawn.
>
> You are effectively asking that someone teach you how to play go.


No, I've read that game sucks. ; >D


> It's much freer than chess because it is not totally dominated by
> tactics, and there are consequently more good, playable moves to
> choose from. It likewise gives greater scope for self-expression, and
> a player's personal style is much easier to see in go than in chess.


Sounds good. As soon as I can find someone to
succeed me (I cannot be "replaced") as the world's
highest-rated player (on GetClub, that is), I might
have a look at go. "I lasted two hours against a
go master before blundering horribly at my second
turn. He replied immediately, capturing all twenty-
seven of the handicap stones I had placed, plus the
two I had played later. Realizing I was in a fix, I
asked for a rematch -- but this time at ninety stones
odds. I then set out to find -- and kill -- Edward
Lasker, for convincing me that this was an easy
game."


-- help bot



 
Date: 04 Oct 2007 13:02:17
From: Dirt At Your Face
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 <[email protected] > typed:
<snip nonsense >
>
> So, let me translate for you Go players: Go is a game that even a
> monkey can play, the moves require little or no thought, you can't
> make a "one-move blunder" easily, like in chess, and, consequently,
> because so many possible patterns exist, computers with present day
> hardware are "fooled" into playing weakly. However, once hardware
> improves, Go will be weakly solved (meaning with 99.9999999...%
> probability, like checkers) and it will be seen that Go players are
> nothing but egotistical naked apes. Watching and playing pachinko is
> more interesting than Go. The true cerebral game of skill is, will be
> and remains, international chess.

http://img167.imageshack.us/img167/700/chessgeniuslt4.jpg

I also have the image made into a nice wall-poster minus the text-message of
course. How much would you reckon the poster would sell?




 
Date: 03 Oct 2007 21:11:45
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 4:31 pm, [email protected] wrote:

> It is perhaps a significant datum that an apparently rabid chess
> partisan (most chess players are certainly intelligent and civilized
> enough to disown the sentiments raylopez99 expresses) would say that
> about go players, but go players do not say that about chess players.
> We simply observe that our game is better ;^)

Hey, if go is better than chess, then why does it take so
long to come up with a legal move? In chess, even a rank
beginner can find some blunder after only a minute or two
of miscalculation.

Many years ago, I played chess with a fellow who
insisted that go was a far better game. He continued
to insist this night after night, which he wasted playing
chess, *not go*! (Presumably, he could find no
monkeys against which to play, and this is why he
continually pestered us to switch games.)


> BTW, an old friend of mine was one of the early contributors to the
> Chinook research effort, and he long ago acknowledged that go is so
> much more complex than checkers that he could not even hazard a guess
> how much more complex.

A battleship is infinitely more complex than a car.
Therefore, we should all drive... battleships?

---------

One serious problem with chess is that it is played
from start to finish, which is to say that every game
begins with position x, and the first few moves can
be, and are to a certain extent, played by rote. (In
other games you determine your own starting position,
or it is generated such that no two games are exactly
alike.) Another drawback is the very limited playing
field -- just 64 squares, along with only two types of
men: expendable and not expendable. Kidding. But
there are only a relatively few types of men in chess.
The "terrain" is perfectly flat; one cannot develop new
weapons of mass destruction; and the enemy is
known in advance. These are all limitations which
give the game a rather narrow scope, making long
familiarity with the way the pieces move and work
together a huge part of success. No newcomer could
ever just come along and crush me at chess (I don't
count computers, dammit!), because I know so much
more about these peculiarities of the game. But
wouldn't it be nice to have a game where everyone is
on an equal footing? Where nobody has "prepared
lines" against almost any reasonable approach you
might try? I think so. I'm tired of the 90+% of chess
that is chained to seeing all the tactics, merely
calculating captures and checks; give me more
deep strategy, more long-range planning and less
check-and-pea-pick-that-pawn.


-- help bot





  
Date: 04 Oct 2007 16:16:16
From: gaga
Subject: Re: Computer Science Experts: Chess 'harder' than Go
> A battleship is infinitely more complex than a car.
> Therefore, we should all drive... battleships?

I live in Germany. majority if not all of battleships that were
registered here were sank very fast during traffic accidents. This makes
me believe that it is dangerous to drive (German) battleships. Besides
this being Germany i.e. a country of zillion laws would make a life of
battleship driver a hell.


gaga


  
Date: 04 Oct 2007 05:44:13
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 21:11:45 -0700, help bot <[email protected] >
wrote:

>On Oct 3, 4:31 pm, [email protected] wrote:
>
>> It is perhaps a significant datum that an apparently rabid chess
>> partisan (most chess players are certainly intelligent and civilized
>> enough to disown the sentiments raylopez99 expresses) would say that
>> about go players, but go players do not say that about chess players.
>> We simply observe that our game is better ;^)
>
> Hey, if go is better than chess, then why does it take so
>long to come up with a legal move?

It takes almost no time at all to come up with a legal move. It is
_illegal_ moves that take a while to come up.

>In chess, even a rank
>beginner can find some blunder after only a minute or two
>of miscalculation.

It is just as easy to find a bad move in go. Maybe easier, as there
are more bad choices.

> Many years ago, I played chess with a fellow who
>insisted that go was a far better game. He continued
>to insist this night after night, which he wasted playing
>chess, *not go*! (Presumably, he could find no
>monkeys against which to play, and this is why he
>continually pestered us to switch games.)

Maybe he just wanted an opponent he could beat.

>> BTW, an old friend of mine was one of the early contributors to the
>> Chinook research effort, and he long ago acknowledged that go is so
>> much more complex than checkers that he could not even hazard a guess
>> how much more complex.
>
> A battleship is infinitely more complex than a car.
>Therefore, we should all drive... battleships?

Careful, you're giving the logical powers of chess players a bad name.

> One serious problem with chess is that it is played
>from start to finish, which is to say that every game
>begins with position x, and the first few moves can
>be, and are to a certain extent, played by rote. (In
>other games you determine your own starting position,
>or it is generated such that no two games are exactly
>alike.) Another drawback is the very limited playing
>field -- just 64 squares, along with only two types of
>men: expendable and not expendable. Kidding. But
>there are only a relatively few types of men in chess.
>The "terrain" is perfectly flat;

Not so: the center is effectively the "high ground."

>one cannot develop new
>weapons of mass destruction; and the enemy is
>known in advance. These are all limitations which
>give the game a rather narrow scope, making long
>familiarity with the way the pieces move and work
>together a huge part of success. No newcomer could
>ever just come along and crush me at chess (I don't
>count computers, dammit!), because I know so much
>more about these peculiarities of the game. But
>wouldn't it be nice to have a game where everyone is
>on an equal footing?

Go's excellent handicap system effectively allows players of widely
divergent skill to play on an equal footing without totally changing
the nature of the game. The weaker player just gets a head start in
the places where he puts his handicap stones.

>Where nobody has "prepared
>lines" against almost any reasonable approach you
>might try? I think so.

In go, free placement of handicap stones allows the weaker player, if
he so chooses, to challenge the stronger with an opening position
unlike any he has seen before.

>I'm tired of the 90+% of chess
>that is chained to seeing all the tactics, merely
>calculating captures and checks; give me more
>deep strategy, more long-range planning and less
>check-and-pea-pick-that-pawn.

You are effectively asking that someone teach you how to play go.
It's much freer than chess because it is not totally dominated by
tactics, and there are consequently more good, playable moves to
choose from. It likewise gives greater scope for self-expression, and
a player's personal style is much easier to see in go than in chess.

-- Roy L


 
Date: 03 Oct 2007 15:55:18
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 12:08 pm, Juha Nieminen <[email protected] > wrote:
> raylopez99 wrote:
> > What is meant is that the "decision complexity" of chess (and
> > checkers) is higher than for Go (actually Go-Moku
>
> Saying that the decision complexity of chess is higher than for
> Gomoku is a *completely* different thing from saying that the
> decision complexity of chess is higher than for Go.
> Gomoku and Go are two completely different games. Gomoku is
> an enormously simpler game than Go in all possible ways.

OK, let's rephrase it--Go is simplier than chess, per move.

Go games take more moves than the average chess game to conclude (a
chess game takes about 40 moves to conclude).

Ergo, per move, Go is simplier.

Think of it this way: building a house piece by piece is simplier
than building it with prefabricated pieces. More technology to snap
together a prefabricated house, rather than build it with stones and
mud and straw.

Go is still a "primitive" board game, and it's like chess was (or
backgammon) when it first started 1500+ years ago.

Go is messier than chess to solve programmically, hence people mistake
this for 'hardness' and 'skill'. Hell, even getting a PC to talk is
"hard", but we know in fact how simple speech really is (and I'm not
talking about the Turing test).

Another analogy: Go is like bridge--is bridge "hard" because the best
software programs cannot play world calibre bridge? Or poker? I
think not.

Case closed, like my mind is on this subject.

RL




  
Date: 04 Oct 2007 12:14:08
From: Ben Finney
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 <[email protected] > writes:

> Go games take more moves than the average chess game to conclude (a
> chess game takes about 40 moves to conclude).
>
> Ergo, per move, Go is simplier.

Chess and Go both take many more moves to conclude, on average, than
the average Tic-Tac Toe game.

Ergo, per move, Go and Chess are both simplier than Tic-Tac-Toe.

--
\ "I'd like to see a nude opera, because when they hit those high


  
Date: 04 Oct 2007 01:47:51
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 15:55:18 -0700, raylopez99 <[email protected] >
wrote:

>On Oct 3, 12:08 pm, Juha Nieminen <[email protected]> wrote:
>> raylopez99 wrote:
>> > What is meant is that the "decision complexity" of chess (and
>> > checkers) is higher than for Go (actually Go-Moku
>>
>> Saying that the decision complexity of chess is higher than for
>> Gomoku is a *completely* different thing from saying that the
>> decision complexity of chess is higher than for Go.
>> Gomoku and Go are two completely different games. Gomoku is
>> an enormously simpler game than Go in all possible ways.
>
>OK, let's rephrase it--Go is simplier than chess, per move.
>
>Go games take more moves than the average chess game to conclude (a
>chess game takes about 40 moves to conclude).
>
>Ergo, per move, Go is simplier.

By that criterion, tic-tac-toe is more complex than chess because it
takes only half a dozen moves.

>Think of it this way: building a house piece by piece is simplier
>than building it with prefabricated pieces. More technology to snap
>together a prefabricated house, rather than build it with stones and
>mud and straw.

Not analogous.

>Go is still a "primitive" board game, and it's like chess was (or
>backgammon) when it first started 1500+ years ago.

Nonsense.

>Go is messier than chess to solve programmically, hence people mistake
>this for 'hardness' and 'skill'. Hell, even getting a PC to talk is
>"hard", but we know in fact how simple speech really is (and I'm not
>talking about the Turing test).

Wrong again. We are beginning to understand just how hard speech is.

>Another analogy: Go is like bridge--is bridge "hard" because the best
>software programs cannot play world calibre bridge? Or poker?

Yes.

>I think not.

You don't know just how true that is.

>Case closed, like my mind is on this subject.

Yep.

-- Roy L


  
Date: 04 Oct 2007 01:11:56
From: Harry Fearnley
Subject: Re: Computer Science Experts: Chess 'harder' than Go

raylopez99 (who cannot spell "Ruy") wrote:

> OK, let's rephrase it--Go is simplier than chess, per move.
>
> Go games take more moves than the average chess game to conclude (a
> chess game takes about 40 moves to conclude).
>
> Ergo, per move, Go is simplier.

Make up your mind -- do you mean Go, or the far superior game
"Ergo", so keted as "Thus"?

> Think of it this way: building a house piece by piece is simplier
> than building it with prefabricated pieces. More technology to snap
> together a prefabricated house, rather than build it with stones and
> mud and straw.

OK -- I'll play with the troll.

Using an analogy, I am pleased to report that I have found a
conclusive proof of your assertion:

A clock is an object.
A brick is an object.
All objects are of equal value.
A clock has many components, and a brick only one.
To have more components is to be simpler.
Therefore, a clock is simpler than a brick.
Q.E.D. (just to prove I know some Latin)

From this I can further safely conclude that a brick will offer
_you_ more intellectual stimulation than a clock.


Please advise me on another, related, matter that is troubling
me. I want to make a cup of coffee, easily. Which should I
start with?:

a) Some instant coffee granules, and some water.

or

b) A collection of many different molecules; a printout of
the DNA of a coffee plant together with instructions on
how to use this; a well-equipped laboratory; a team of
biochemists; ...

--
Harry Fearnley -- bestiaries, seki, ... http://www.goban.demon.co.uk


 
Date: 03 Oct 2007 15:48:26
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 2:57 pm, james <[email protected] > wrote:

> All you say is completely wrong, and I know well enough Jonhathan
> Schaeffer and his team's publications and work to be absolutely sure
> that no one of the University of Alberta's CS could have said such a
> thing. Go is much more complex than chess or checkers. That is the
> scientific fact.

Yeah, on the net I know Steven Hawking, and he says Go sucks.


> The only part of truth is that some Go endgames are, in a way, easier to
> solve than chess endgames, because they fit inside John Conway's "nimber
> theory". See Berlekamp's book "Mathematical Go"

OK, conceeded that Go is simplier. Thanks.




  
Date: 04 Oct 2007 01:43:30
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 15:48:26 -0700, raylopez99 <[email protected] >
wrote:

>Yeah, on the net I know Steven Hawking, and he says Go sucks.

Lie.

>> The only part of truth is that some Go endgames are, in a way, easier to
>> solve than chess endgames, because they fit inside John Conway's "nimber
>> theory". See Berlekamp's book "Mathematical Go"
>
>OK, conceeded that Go is simplier. Thanks.

Lie.

-- Roy L


 
Date: 03 Oct 2007 15:45:54
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 1:49 pm, Ken Blake <[email protected] > wrote:
> On Wed, 03 Oct 2007 13:31:23 -0700, help bot <[email protected]>
> wrote:
>
> > For those of us who don't know Go from Go-Moku,
> > what is the difference, in layman's terms?
>
> Go-Moku is really just a fancy (although much more difficult) version
> of Tic-Tac-Toe. The object is to get five in a row, but it's played on
> a regular 19x19 Go board.
>
> Go is an entirely different, and much more difficult, game.
>


Thanks Ken. I've never played Go except against a s/w program, but
have played go-moku. But I'll go out on a limb and say Go has a
smaller decision complexity per move than chess. One way of seeing
this is to count the moves in an 'average' game of GO (to completion)
versus an 'average' game of chess. In chess, good players finish in
around 40 moves. I doubt the same can be said for Go. Chess packs
more per move than Go.

So, in short, GO is for monkeys.

Sorry, them'z da facts.

RL



  
Date: 04 Oct 2007 01:54:11
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 15:45:54 -0700, raylopez99 <[email protected] >
wrote:

>On Oct 3, 1:49 pm, Ken Blake <[email protected]> wrote:
>> On Wed, 03 Oct 2007 13:31:23 -0700, help bot <[email protected]>
>> wrote:
>>
>> > For those of us who don't know Go from Go-Moku,
>> > what is the difference, in layman's terms?
>>
>> Go-Moku is really just a fancy (although much more difficult) version
>> of Tic-Tac-Toe. The object is to get five in a row, but it's played on
>> a regular 19x19 Go board.
>>
>> Go is an entirely different, and much more difficult, game.
>
>Thanks Ken. I've never played Go except against a s/w program, but
>have played go-moku. But I'll go out on a limb and say Go has a
>smaller decision complexity per move than chess.

That's not out on a limb. That's a sawed-off limb.

>One way of seeing
>this is to count the moves in an 'average' game of GO (to completion)
>versus an 'average' game of chess. In chess, good players finish in
>around 40 moves. I doubt the same can be said for Go. Chess packs
>more per move than Go.

And tic-tc-toe more than chess....?

ROTFL!!

>So, in short, GO is for monkeys.
>
>Sorry, them'z da facts.

Troll meter's pinned...

-- Roy L


 
Date: 03 Oct 2007 15:41:08
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 10:35 am, [email protected] (-) wrote:
> > raylopez99 wrote:
> >> ... --even though, for each move, less thought is involved.
> Michael Alford <[email protected]> wrote:
> > utter nonsense
>
> Agree. We don't need to discuss "what is thought" in order
> to build a game-playing device.
>
> >> From the author's paper, page 1519: "Checkers [and by
> >> extention, chess, mentioned in the article--RL] ...requires extensive
> >> skill [<--note that phrase--RL] to make strong move choices...and
> >> [has] moderate space complexity... All the games solved thus far have
> >> either low decision complexity (Qubic and Go-Moku), low space
> >> complexity (Nine Men's Morris...Awari).. or both (Connect Four)."
>
> >> [ ... ]
> > Total bullshit.
>
> First of all, in order to discuss the Science Magazine article we
> should have the author names cited.

It's J. Schaeffer and some dudes named "et al." They've been around
for a while (since 1989).


> Next, at risk of treading upon
> "sacred temple territory" of these Holy Go Spaces, the considerations
> expressed are sound. Frank de Groot has also been here, a prophetic
> champion of "gate arrays." There exists a hardware solution to the
> problem of exponential fan-out, albeit that the hardware solution is
> still elusive. Disinterest by much of a computer artificial intelligence
> community should not be misinterpreted to indicate severe impossibility.
> Instead we could examine the sociological aspect: what led computer
> scientists first to attack Chess instead of Go? Was there a perception
> that Chess is more highly respected as a decision-rendering tool for
> purposes of conducting our state-of-affairs in World Diplomacy?

A short answer to a rather deep compound question: YES.

>
> As the Science Magazine authors note, a solved game may fall due
> to low decision -OR- low space complexity. The decision side is more
> difficult to visualize, and to devise in hardware, however it remains
> true that the impenetrability of a game to machine analysis is the
> result of a combination of both and not simply one aspect or the other.

Agreed. But you have to acknowledge that chess has a higher decision
complexity per move. THis is what the paper said. It's pretty
obvious too--alpha-beta takes thought (whether in wetware or
software), while moving little white stones around is child's play
(pattern recognition, not unlike a baby recognizing her mother's
face).

> The fact that many millenia of human history had elapsed prior to atomic
> weapons does not invalidate the reality that atomic weapons exist today.
> When the hardware devices are invented people may come to regard
> that scientific team with awe and respect, as was accorded to Einstein.

Yes, though I'll add that Bohr, Einstein, Dirac and others gained awe
and respect by doing 'thought experiments' before the hardware was
even invented. And Leo Szilard is the real unsung hero for the atomic
bomb (http://www.jewishvirtuallibrary.org/jsource/biography/
Szilard.html)

>
> Go Players unprepared for the prospect that their favorite sanctuary
> will someday be invaded, and overrun by electronic hoardes, share an
> immaturity aspect which the Chess Community has long ago shed. It
> will be interesting to observe how many Go Players continue once those
> hardware devices gain an advantage: this would surely be a test of faith
> once the rubber meets the road. Who prefers not to live to see that day?

I do. The Cyborgs are my friends. When they colonize the earth they
promise to keep me well fed and taken care of as their pet, if I
promote their cause now, which I am doing with this post.

RL



  
Date: 04 Oct 2007 01:50:57
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 15:41:08 -0700, raylopez99 <[email protected] >
wrote:

>But you have to acknowledge that chess has a higher decision
>complexity per move. THis is what the paper said.

Lie.

>It's pretty
>obvious too--alpha-beta takes thought (whether in wetware or
>software), while moving little white stones around is child's play
>(pattern recognition, not unlike a baby recognizing her mother's
>face).

The stones don't move around in go, and pattern recognition is a
higher cognitive function than alpha-beta.

-- Roy L


 
Date: 03 Oct 2007 15:33:16
From: raylopez99
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 8:05 am, Guy Macon <http://www.guymacon.com/ > wrote:
> Trollometer
> 0 1 2 3 4 5 6 7 8 9 +10dB +20dB
>


  
Date: 03 Oct 2007 15:46:03
From: Michael Alford
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 wrote:
> On Oct 3, 8:05 am, Guy Macon <http://www.guymacon.com/> wrote:
>> Trollometer
>> 0 1 2 3 4 5 6 7 8 9 +10dB +20dB
>>


 
Date: 03 Oct 2007 23:57:06
From: james
Subject: Re: Computer Science Experts: Chess 'harder' than Go
Gomoku has nothing in common with Go. It is in western countries a
pencil and paper game played by lot of young students.
Programming GoMoku is a very simple task, and programs have been
extremely strong for years.

All you say is completely wrong, and I know well enough Jonhathan
Schaeffer and his team's publications and work to be absolutely sure
that no one of the University of Alberta's CS could have said such a
thing. Go is much more complex than chess or checkers. That is the
scientific fact.

The only part of truth is that some Go endgames are, in a way, easier to
solve than chess endgames, because they fit inside John Conway's "nimber
theory". See Berlekamp's book "Mathematical Go"

http://www.amazon.com/Mathematical-Go-Chilling-Gets-Point/dp/1568810326/ref=sr_1_9/103-8091652-3116645?ie=UTF8&s=books&qid=1191448274&sr=8-9



raylopez99 a �crit :
> In an article in Science, 14 September 2007 p. 1519, written by the
> Chinook team that weakly solved the game of checkers (draughts), it's
> asserted essentially that chess is harder than Go.
>
> What is meant is that the "decision complexity" of chess (and
> checkers) is higher than for Go (actually Go-Moku was mentioned, which
> is not exactly the same as Go, but I take it the analogy holds for
> Go), while the game of Go has a higher "space complexity" (essentially
> non-repeating patterns--more such patterns in Go than Chess/Checkers).
>
> What is the consequence of this? It's harder for a computer to
> "solve" the game of Go only because there is more "stuff" the computer
> has to consider--even though, for each move, less thought is
> involved. From the author's paper, page 1519: "Checkers [and by
> extention, chess, mentioned in the article--RL] ...requires extensive
> skill [<--note that phrase--RL] to make strong move choices...and
> [has] moderate space complexity... All the games solved thus far have
> either low decision complexity (Qubic and Go-Moku), low space
> complexity (Nine Men's Morris...Awari).. or both (Connect Four)."
>
> So, let me translate for you Go players: Go is a game that even a
> monkey can play, the moves require little or no thought, you can't
> make a "one-move blunder" easily, like in chess, and, consequently,
> because so many possible patterns exist, computers with present day
> hardware are "fooled" into playing weakly. However, once hardware
> improves, Go will be weakly solved (meaning with 99.9999999...%
> probability, like checkers) and it will be seen that Go players are
> nothing but egotistical naked apes. Watching and playing pachinko is
> more interesting than Go. The true cerebral game of skill is, will be
> and remains, international chess.
>
> That's what SCIENCE sez, sucker. Don't blame me for reporting the
> FACTS.
>
> RL
>


 
Date: 03 Oct 2007 21:31:27
From:
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 07:15:26 -0700, raylopez99 <[email protected] >
wrote:

>In an article in Science, 14 September 2007 p. 1519, written by the
>Chinook team that weakly solved the game of checkers (draughts), it's
>asserted essentially that chess is harder than Go.

That is, inevitably, a flat-out lie.

>What is meant is that the "decision complexity" of chess (and
>checkers) is higher than for Go (actually Go-Moku was mentioned, which
>is not exactly the same as Go, but I take it the analogy holds for
>Go),

Obviously, it does not. The only similarities between go and go-moku
are the occurrence of the "go" syllable in both their names, and the
nature of the equipment and moves (i.e., placing a stone on a vacant
intersection of a square grid). The similarity of decision complexity
between go and go-moku is substantially less than the similarity
between chess and giveaway chess.

>while the game of Go has a higher "space complexity" (essentially
>non-repeating patterns--more such patterns in Go than Chess/Checkers).
>
>What is the consequence of this? It's harder for a computer to
>"solve" the game of Go only because there is more "stuff" the computer
>has to consider--even though, for each move, less thought is
>involved.

That is also clearly just false. Professional go games routinely
feature longer periods of thought between moves than top-level chess
games. Half an hour or more is not at all uncommon. The World Chess
Championship time control of 40 moves in the first two hours of a
player's time -- an average of three minutes per move -- would be
considered almost a "speed go" pace by top professional go players.

>From the author's paper, page 1519: "Checkers [and by
>extention, chess, mentioned in the article--RL] ...requires extensive
>skill [<--note that phrase--RL] to make strong move choices...and
>[has] moderate space complexity... All the games solved thus far have
>either low decision complexity (Qubic and Go-Moku), low space
>complexity (Nine Men's Morris...Awari).. or both (Connect Four)."

Note that none of this even refers to go at all.

>So, let me translate for you Go players: Go is a game that even a
>monkey can play,

None has ever done so, so that is obviously another lie.

>the moves require little or no thought,

As long as you don't care if they are good or bad.

>you can't make a "one-move blunder" easily, like in chess,

One move blunders are about as common in go as in chess.

>and, consequently,
>because so many possible patterns exist, computers with present day
>hardware are "fooled" into playing weakly. However, once hardware
>improves, Go will be weakly solved (meaning with 99.9999999...%
>probability, like checkers)

There is effectively no chance whatever that go will be weakly solved
by hardware improvements.

>and it will be seen that Go players are
>nothing but egotistical naked apes.

It is perhaps a significant datum that an apparently rabid chess
partisan (most chess players are certainly intelligent and civilized
enough to disown the sentiments raylopez99 expresses) would say that
about go players, but go players do not say that about chess players.
We simply observe that our game is better ;^)

>Watching and playing pachinko is more interesting than Go.

Probably true if you do not know how to play go, like raylopez99 or a
monkey.

>The true cerebral game of skill is, will be
>and remains, international chess.

Anyone even slightly knowledgeable about both games acknowledges that
go requires at least as much intellectual skill as chess.

>That's what SCIENCE sez, sucker. Don't blame me for reporting the
>FACTS.

You did not report facts. You just lied about what a scientific paper
said. Simple.

BTW, an old friend of mine was one of the early contributors to the
Chinook research effort, and he long ago acknowledged that go is so
much more complex than checkers that he could not even hazard a guess
how much more complex.

-- Roy L


 
Date: 03 Oct 2007 14:08:06
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 1:29 pm, Harry Sigerson <[email protected] > wrote:

> Note that last bit. They are games, with a set of fairly narrow rules
> limiting their function to their gaming surfaces. Clever games sure but still
> only games.

Chess is only a game? Preposterous. That would mean
that all of us chess players -- though thinking we are doing
something very clever -- are merely wasting our time. That
can't be... no! No, no, no, no, no! Gasp...


> As far as cerebration is concerned they can neither of them be compared to
> *top level* 'play' in the fields of say mathematics, or music, or literature --

Already been done, my man. The quick and dirty
sumy is that some fields have a more natural
audience (music), more real-world value (math), or
are of no value whatever to illiterates (lit.). Recorded
chess games are a bit like art, though ignorants
cannot properly appreciate them.


> and I'm sure you can quote some more of the same (as long as you don't include
> religion).

Hang religion; if God had meant for man to believe
in fictitious deities, he would have made us all
gullible idiots. If we had been meant to fly, we'd
obviously have been born with beaks and feathers.


> Undoubtedly, clever people in any one or more than one of those fields would
> be likely to be good at playing Go or Chess or Draughts but in the spirit of
> there only being so many days in the year, they might not have played at all or
> on the other hand decide to try for chess grandmastership.

Nonsense. Many in the field of math *are* excellent
chess players, having spend a great deal of time on
the game. But there are always a few who strive but
fail, in spite of any accomplishments in other fields.
Look at Dr. Einstein and Dr. Blair, for instance. ; >D


> So the
> initial path to getting even reasonably good at the game is to learn the basic,
> long establish, even fashionable openings.

Um, no. That is a well-worn path into a deep, dark
forest, leading nowhere. Stick with learning the fun-
damentals, and focus heavily on tactics; this will get
you up to speed the quickest, and if you make a few
errors in the openings, that will simply serve to help
you learn them, not by rote, but by real understanding.


> Somehow even the most ordinary low-level player still feels that he has a
> handle on the game

How could you possibly know how I feel?


-- help bot





  
Date: 04 Oct 2007 13:21:41
From: Harry Sigerson
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 14:08:06 -0700, help bot <[email protected] > wrote:

Dear Help Bot,
an enjoyable reply; I'll pick up and run a little with it; in what I hope
will be the same light-hearted spirit.


> Chess is only a game? Preposterous. That would mean
> that all of us chess players -- though thinking we are doing
> something very clever -- are merely wasting our time. That
> can't be... no! No, no, no, no, no! Gasp...
>
Inference, the very bugbear of discussion; though the 'Gasp...' does tell me
the souffl� is still up.
In saying that Chess and Go are only games, the 'only' part isn't
derogation. Of all the games that have been thought up through the ages both of
those have hung about persistently. That, /only/ just points out that they are
passtimes; clever ones, with more to offer than conkers.
{Conkers, here, was frowned on by Health & Safety a few years ago; to the
extent that in one town, roadside chestnut trees were cut down to prevent
accidents. I don't know if was to prevent horse-chestnut gathering kids being
knocked down by motor cars or if H&S figured the little dears might knock each
others' eyes out playing such rough games. But pass-times are _all_ they are;
good ones yes, but most of us don't put meat in the stew by being good Chess or
Go players.


> Already been done, my man. The quick and dirty
> sumy is that some fields have a more natural
> audience (music), more real-world value (math), or
> are of no value whatever to illiterates (lit.).
>
There's that divide and rule thing in action.
Take the middle one first, mathematics. [I'll infer possibly wrongly that
your 'real-world value' relates to financial gain] Most who have a little bit of
maths tucked safely up their sleeve will be among the first to tell of their
enjoyment in its acquisition. While continuing to feed the family by being a good
butcher or /whatever/. I'd say that enjoyment came repeatedly each time the light
went on for them while trying to grasp some particular facet of the subject.
The music one, is almost simpler in that you can't turn your ears off; you
can look away from a tv set but the ears keep on hearing the crap ads. If you
live alone you can dump the goggle box but few do.
Literature, "...written works, especially those regarded as having artistic
merit." is of course founded on the spoken language. If the crap movies that they
show on our television is anything to go by then one might be forgiven for some
anxiety for Lit's future; though that would be ill founded.
People always have and always will want to hear stories, good stories.
The present day stories are told by moving pictures, to millions at a time
with minimal recourse to language; it's all there in the coloured pictures. Even
to the bullets battering through the baddy's sternum and, nowadays, you can even
watch them squirting out through his back.
Who needs literature in such an society; as for those 'illiterates', now
there's a very flexible decation line all ready and renewed for addressing.
/Flashing digression alarm noticed/


> Recorded
> chess games are a bit like art, though ignorants
> cannot properly appreciate them.
>
Well, that now is both an evaluation and an observation and both counts are
definitely in the eyes of the beholder and why shouldn't' they be. Each of us is
alone, no matter which of the many of our institutions we use, consciously or
unconsciously to ignore this fact; be it riage, family, religion, the
workplace, freemasonry, whatever - to use an American catch-all word. I cannot
let others think with my brain - even if I wanted to - nor can I think with
theirs and long may that remain so. The old-fashioned way is still the best,
speaking to each other while limiting the use of the put-down pedal.
But any societal body can and all societies should, bring up their
individuals to know this simple fact and provide the necessary 'enlightenment'
arena that makes music, maths and literature the norm.

> Hang religion; if God had meant for man to believe
> in fictitious deities, he would have made us all
> gullible idiots.
>
Nice one, but again language-of-longstanding imposes the word 'God' on the
writer and the reader both, as a matter of course.
Your play on the name recalls a cartoon dating from the early days of the
wee Morris Mini-Minor, that first east/west engined car.
The cartoonist has you looking at the backend of a large car of the Rolls or
Bently variety, being driven sedately along by an old buffer, his wife sitting
upfront with him.
They are being overtaken by a Mini going hell-for-leather, up in the air a
couple of feet above the tarmac (asphaltic concrete) as both cars breast a rise.
The old gent says...
"If the good God had intended little cars to go that fast he'd have given
them bigger wheels."

> If we had been meant to fly, we'd
> obviously have been born with beaks and feathers.
>
Now there's a thought. Though I still remember a tv movie of an American who
built a plane that flew round the world 'in a wanny' as they say in these hyar
peurts. Twin fuselage P38-like, it had two props of the push-me pull-me sort,
back and front of the central cabin. I know that round the world flight's been
done again using a jet engine but that first one for me is still the best
remembered.

> Nonsense. Many in the field of math *are* excellent
> chess players, having spend a great deal of time on
> the game. ...
>
Which paraphrases what I said in the paragraph that won the 'nonsense'
medal.

> ...But there are always a few who strive but
> fail, in spite of any accomplishments in other fields.
> Look at Dr. Einstein and Dr. Blair, for instance. ;>D
>
I can't recall the names in this possibly apocryphal tale about Albert. It
seems that he had a friend who was a good or even great violinist and Einstein
had some pretensions as a player. They played together.
The friend, irritated at some point in their duet admonishes,
"No, no, no, Albert! It's 'One two three; One two three,...'

To make comparisons across such a boundary as that between Einsteinian
mathemtics and anything else he did is normal; celebrity has to pay its dues but
it is to some extent pointless.
There is a play in which ilyn Monroe really wants to meet Albert and does
eventually meet him in his hotel room, in New York I think; now that was an
interesting juxtaposition of celebrity. A sort of brains and blondery situation
but we each know which of them we'd prefer to meet.
I don't know who Dr Blair is; let's hope it's not 'wor Tony' the teurreurist
twacker and good Bush-buddy.

> Um, no. That is a well-worn path into a deep, dark
> forest, leading nowhere.
>
That's good to know; though I get the impression there are a lot of players
well-treed in that deep dark.


> Stick with learning the fun-
> damentals, and focus heavily on tactics; this will get
> you up to speed the quickest, and if you make a few
> errors in the openings, that will simply serve to help
> you learn them, not by rote, but by real understanding.
>
Sounds like good advice and of much the same calibre as that given to newbie
Go-ers; to play lots of games as quick as you like and keep on doing just that.


> > Somehow even the most ordinary low-level player still feels that he has a
> > handle on the game
>
> How could you possibly know how I feel?
>
<smile > I'm not sure if that was wonder at my perspicuity or a quibble at
such temerity.

Harry



 
Date: 03 Oct 2007 13:31:23
From: help bot
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Oct 3, 10:05 am, Guy Macon <http://www.guymacon.com/ > wrote:
> Trollometer
> 0 1 2 3 4 5 6 7 8 9 +10dB +20dB
>


  
Date: 03 Oct 2007 13:49:51
From: Ken Blake
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 13:31:23 -0700, help bot <[email protected] >
wrote:

> For those of us who don't know Go from Go-Moku,
> what is the difference, in layman's terms?



Go-Moku is really just a fancy (although much more difficult) version
of Tic-Tac-Toe. The object is to get five in a row, but it's played on
a regular 19x19 Go board.

Go is an entirely different, and much more difficult, game.

--
Ken Blake
Please Reply to the Newsgroup


 
Date: 03 Oct 2007 22:08:05
From: Juha Nieminen
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 wrote:
> What is meant is that the "decision complexity" of chess (and
> checkers) is higher than for Go (actually Go-Moku

Saying that the decision complexity of chess is higher than for
Gomoku is a *completely* different thing from saying that the
decision complexity of chess is higher than for Go.
Gomoku and Go are two completely different games. Gomoku is
an enormously simpler game than Go in all possible ways.


 
Date: 03 Oct 2007 13:48:51
From: Jonathan Buss
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 wrote:
> In an article in Science, 14 September 2007 p. 1519, written by the
> Chinook team that weakly solved the game of checkers (draughts), it's
> asserted essentially that chess is harder than Go.
>
> What is meant is that the "decision complexity" of chess (and
> checkers) is higher than for Go (actually Go-Moku was mentioned, which
> is not exactly the same as Go, but I take it the analogy holds for
> Go), while the game of Go has a higher "space complexity" (essentially
> non-repeating patterns--more such patterns in Go than Chess/Checkers).

Go has at least the same decision complexity as chess, and possibly greater.

Go-moku, on the other hand, is a different story. It's a completely
different game than go, despite similar equipment.

The Chinook team is quite aware of both of the above. Errors to the
contrary in the article, if indeed any exist, would be the fault of the
editing staff and/or the reader.

Jonathan Buss


  
Date: 04 Oct 2007 07:50:23
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

> raylopez99 wrote:
>> In an article in Science, 14 September 2007 p. 1519, written by the
>> Chinook team that weakly solved the game of checkers (draughts), it's
>> asserted essentially that chess is harder than Go.
>>
>> What is meant is that the "decision complexity" of chess (and
>> checkers) is higher than for Go (actually Go-Moku was mentioned, which
>> is not exactly the same as Go, but I take it the analogy holds for
>> Go), while the game of Go has a higher "space complexity" (essentially
>> non-repeating patterns--more such patterns in Go than Chess/Checkers).

Jonathan Buss <[email protected] > wrote:
> Go has at least the same decision complexity as chess, and possibly greater.


Surely it's a mistake by "raylopez99" to draw inferences about Go from
article's mention of Go-Moku. Go-Moku, without initial starting rules,
was "solved" by Victor Allis in 1993. A "definition of terms" available:

"The decision complexity of a game is the number of leaf nodes in the
smallest decision tree that establishes the value of the initial position."
http://en.wikipedia.org/wiki/Game-tree_complexity

Once paring away the Greek, this sort of definition can follow from an
alpha-beta simplification, or as by labelling massive branches a win for
Player A or Player B. Game-tree complexity uses "full-width" search,
a slightly different concept, though a key difference to keep in mind.

It's not easy to obtain the article, in Microsoft Territory, so I settled
for an abstract, supporting online material, a few google references.
Here's what Jonathan Schaeffer actually said, about Chess & Go:


at: http://www.aaai.org/AITopics/html/go.html

"If you want to understand intelligence, the game of Go is much
more demanding ... It doesn't have the silver bullet: deep search.
Chess has somewhat outlived its usefulness. It turned out to be
easier than we thought.
- Jonathan Schaeffer, in Chess - Man vs. Machine Plays Out
http://www.aaai.org/AITopics/html/go.html#wired


Our particular _Science_ article lists a -team- of author-names:
Jonathan Schaeffer, Akihiro Kishimoto, tin Muller, Robert Lake ...
Crucial to recognize tin Muller as a longstanding contributor to Go
game algorithms, programming, and theorem-proving. That _Science_
article was not intended to become some polemic for slandering Go.



> (-) wrote:
>> ... we could examine the sociological aspect: what led computer
>> scientists first to attack Chess instead of Go? Was there a perception
>> that Chess is more highly respected as a decision-rendering tool for
>> purposes of conducting our state-of-affairs in World Diplomacy?

"raylopez99" <[email protected] > wrote:
> A short answer to a rather deep compound question: YES.


Thank you for answering. Still, I'm rather curious how that perception
plays out. In what way are Chess Players exerting a positive influence
upon contemporary World Diplomacy?



>> As the Science Magazine authors note, a solved game may fall due
>> to low decision -OR- low space complexity. The decision side is more
>> difficult to visualize, and to devise in hardware, however it remains
>> true that the impenetrability of a game to machine analysis is the
>> result of a combination of both and not simply one aspect or the other.

> Agreed. But you have to acknowledge that chess has a higher decision
> complexity per move. THis is what the paper said. It's pretty
> obvious too--alpha-beta takes thought (whether in wetware or
> software), while moving little white stones around is child's play
> (pattern recognition, not unlike a baby recognizing her mother's
> face).


Let's follow-up with a definition for "space complexity." This is:

"The state-space complexity of a game is the number of legal game
positions reachable from the initial position of the game.[1]

"When this is too hard to calculate, an upper bound can often be
computed by including illegal positions or positions that can never
arise in the course of a game."
- http://en.wikipedia.org/wiki/Game-tree_complexity


John Tromp and Gunnar Farneb�ck have conducted research
on the state-space complexity of Go, periodically publishing their
preliminary findings to rec.games.go for cases of smaller boards.
I must wait on the text of the article, which I haven't before me, before
commenting on your claims concerning what the paper said. We think
that this _Science_ paper was comparing Go-Moku but not Go. The
decision-complexity definition supplied by Wikipedia refers to the
calculation of the value of a position. Because Go Positions must
be calculated within narrow gins there is much opportunity for
shifting flim-flams about each convergent _komi_ value of the game.
Unlike Chess the calculation of a positional values in Go does not
reduce to +/- for either player so quickly. So it is quite likely that
decision-complexity for Go is also higher than would be for Chess.
When one game has both higher decision-complexity -AND- higher
space-complexity then you may infer an overall ordering relation.

This result, however, is deceptive because it argues from the
definition of a term, and these term-definitions do not lead us toward
much solvability direction in Go. It would be more productive to note
a positional value distribution of candidate moves, perhaps bell-curve
shaped, and whether there was any flattening near the selection for
optimal move candidates. If, in Go, I can weed out candidate moves
to two dozen and then posit a half-dozen good moves which continue
forward from a position then perhaps my selection problem reduces
to a likelihood of 1/4th probability. Let's say, though, in Chess I find
six candidate moves after weeding, but owing to the "sharpness" of
Chess only one of those moves is really "best" and "worth playing."
In this sort of thought-experiment my selection problem is then 1/6th,
offering me less opportunity of finding a good forward continuance,
meaning that a Chess task would be more difficult than the Go task.

In either case the machine task is to play fairly well though not
brilliantly, just maintaining a balance, and then winning the endgame.
Whether it's Chess or Go, the machine is dominant for the endgame.
Yet, for Go, perhaps it's a more difficult task just to keep a balance.
Playing "approximately" would never suffice to raise the level of Go
machine-play toward pro level. So, for that thought-experiment,
we were comparing a professional Chess program with only an
amateur Go program, i.e. merely an apples and oranges model.



> The Cyborgs are my friends. When they colonize the earth they
> promise to keep me well fed and taken care of as their pet, if I
> promote their cause now, which I am doing with this post.


"Thinking" or having "causes" ... were these machine characteristics?
A banker tried to tell me the purpose of my savings account. Instead
I withdrew my money and found another bank where I intend purpose.


"Complexity" is a shifting target, of course. The term follows
from state-of-the-art theory in both algorithms and hardware, always
relying upon the known rather than the knowable (of high controversy).
Find a new, previously unthought-of, secret technique and then the
deck is shuffled. The "implicate ordering" of what was complex and
what was simple could be recast. The "mobile emitter" for holographic
doctors was the product of a -future-future- technology time-warped
back into the context of the -future- Star Trek Voyager sci-fi fantasy.
Where the "mobile emitter" could operate was, of course, something
also extensible, later allowing holographic doctors into a transporter.
Yet, the "mobile emitter" could not be duplicated: so it was unique.

Playouts in MonteCarlo could be full-width or alpha-beta pruned.
With additional pattern-recognition a pruning might reduce awesome
decision-complexity into something more manageable: this was one
of Frank de Groot's visions. Recall that the machine task is not to be
an overwhelming player but simply to find humdrum game balance
in order to apply CGT during _yose_. Prior to the CGT there may be
some sort of extensible pre-CGT phase which initiates the lead-out.
The problem concerns "sente" moves which are, typically, not in a
general vicinity of the move recently played. The phrase "not in a
general vicinity" tends to widen, not restrict, the candidate search.

Beyond a threshold, more playouts do not improve move selection.
The hardware I was attempting to describe is not a MC/UCT scheme,
however. Would be hardware from something like Hogwarts, actually,
where the particular configurations of switches, levers and tripwires,
were already calibrated for zapping electrons into proper pigeonholes.
Could be very fast, especially dart-like pieces that do not move around.
The "moving around" aspects of Chess could later (future) prove to be
more intractably tenacious than the "sitting around" aspects for Go.





- regards
- jb

-------------------------------------------------------------
U.S. Labs Mishandling Deadly Germs
http://apnews.myway.com/article/20071002/D8S16GAG0.html
-------------------------------------------------------------



   
Date: 04 Oct 2007 13:59:55
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go



- wrote:

> Surely it's a mistake by "raylopez99" to draw inferences about Go from
> article's mention of Go-Moku. Go-Moku, without initial starting rules,
> was "solved" by Victor Allis in 1993. A "definition of terms" available:
>
> "The decision complexity of a game is the number of leaf nodes in the
> smallest decision tree that establishes the value of the initial position."
> http://en.wikipedia.org/wiki/Game-tree_complexity
>
> Once paring away the Greek, this sort of definition can follow from an
> alpha-beta simplification, or as by labelling massive branches a win for
> Player A or Player B. Game-tree complexity uses "full-width" search,
> a slightly different concept, though a key difference to keep in mind.

Good point.

Poker has three possible moves at each point. Roulette has 87.
Does that make Poker a simpler or easier game?

--
Guy Macon
<http://www.guymacon.com/ >



 
Date: 03 Oct 2007 15:05:46
From: Guy Macon
Subject: Re: Computer Science Experts: Chess 'harder' than Go



Trollometer
0 1 2 3 4 5 6 7 8 9 +10dB +20dB


 
Date: 03 Oct 2007 07:44:43
From: Michael Alford
Subject: Re: Computer Science Experts: Chess 'harder' than Go
raylopez99 wrote:
> In an article in Science, 14 September 2007 p. 1519, written by the
> Chinook team that weakly solved the game of checkers (draughts), it's
> asserted essentially that chess is harder than Go.
>
> What is meant is that the "decision complexity" of chess (and
> checkers) is higher than for Go (actually Go-Moku was mentioned, which
> is not exactly the same as Go, but I take it the analogy holds for
> Go), while the game of Go has a higher "space complexity" (essentially
> non-repeating patterns--more such patterns in Go than Chess/Checkers).
>
> What is the consequence of this? It's harder for a computer to
> "solve" the game of Go only because there is more "stuff" the computer
> has to consider

this is true

--even though, for each move, less thought is
> involved.

utter nonsense

From the author's paper, page 1519: "Checkers [and by
> extention, chess, mentioned in the article--RL] ...requires extensive
> skill [<--note that phrase--RL] to make strong move choices...and
> [has] moderate space complexity... All the games solved thus far have
> either low decision complexity (Qubic and Go-Moku), low space
> complexity (Nine Men's Morris...Awari).. or both (Connect Four)."
>
> So, let me translate for you Go players: Go is a game that even a
> monkey can play, the moves require little or no thought, you can't
> make a "one-move blunder" easily, like in chess, and, consequently,
> because so many possible patterns exist, computers with present day
> hardware are "fooled" into playing weakly. However, once hardware
> improves, Go will be weakly solved (meaning with 99.9999999...%
> probability, like checkers) and it will be seen that Go players are
> nothing but egotistical naked apes. Watching and playing pachinko is
> more interesting than Go. The true cerebral game of skill is, will be
> and remains, international chess.
>
> That's what SCIENCE sez, sucker. Don't blame me for reporting the
> FACTS.
>
> RL
>

Total bullshit.


  
Date: 03 Oct 2007 18:29:09
From: Harry Sigerson
Subject: Re: Computer Science Experts: Chess 'harder' than Go
On Wed, 03 Oct 2007 07:44:43 -0700, Michael Alford <[email protected] > wrote:

> > That's what SCIENCE sez, sucker. Don't blame me for reporting the
> > FACTS.
> >
> > RL
> >
>
> Total bullshit.
>
Michael,
It's a succinct reply but it's not much of a riposte to a quote from a
science mag and a 'translation' which has the 'monkey' comment; the 'one-move-
blunder' misconception; and is rounded off with a 'miasma', centred on another
misconception - that of relative 'cerebral skills' for the two games.
Note that last bit. They are games, with a set of fairly narrow rules
limiting their function to their gaming surfaces. Clever games sure but still
only games.
As far as cerebration is concerned they can neither of them be compared to
*top level* 'play' in the fields of say mathematics, or music, or literature --
and I'm sure you can quote some more of the same (as long as you don't include
religion).
Undoubtedly, clever people in any one or more than one of those fields would
be likely to be good at playing Go or Chess or Draughts but in the spirit of
there only being so many days in the year, they might not have played at all or
on the other hand decide to try for chess grandmastership.
Almost certainly they would not have recourse to words like 'sucker'; now
where did a word like that come from - and why is it still hanging around? COED's
11 Edition has it as...

> sucker
> 3 informal a gullible or easily deceived person. �(a sucker for) a person >
especially susceptible to or fond of (a specified thing).
> 4 North American informal: an unspecified person or thing.

It's as overused as 'awesome'.
========
Chess' I was introduced to by an uncle, when I 12 or so. I really like the
board and the chessmen and of course the neatly set-out rules on the different
and very specific ways in which the chessmen are allowed to move.
With its small area of no-man's-land between the opposing teams and
initially, the clear constraints on the movement of each side's backrow, players
have to make the best fist they can out of opening pathways for those major
pieces. This is fine. It's a game. Both sides have the same basic problem. So the
initial path to getting even reasonably good at the game is to learn the basic,
long establish, even fashionable openings.
Somehow even the most ordinary low-level player still feels that he has a
handle on the game; even if that handle for him is that his men, in general, are
moving towards the opponent's and in particular for his pawns. But that's fine
it's the rules and they're not difficult to learn.
========
Go/Wieqi/Baduk: the basic rules of play are ever so much simpler than those
for chess. Though the players sit on opposite sides of the goban, it seems to me
that you could introduce an new rule requiring the goban to be rotated 90� every
say five moves and it would make no difference to the players' conception of the
play.
But it's very doubtful if the ordinary newbie player really has a handle on
the game. Knowledge about what is happening within the time-flow of the play, I
can only imagine as being the preserve of players who have played many hundreds
of games. Even then when listening to the comments made in Chat areas of internet
clients by audiences following live games, the handle-grasp to me seems, shall I
say 'variable'. I'm not dogmatic about this as I still, even after watching so
many games, am very much in the dark.
That is clearly not the case with the great and the good Go players. Each
knows what they are about and what their opponent is about with each stone
placement. Not necessarily in an exact way but in a possible 'pressure wave'
manner. The winning is effected, not in the slightly narrow but somehow
circumscribed set of moves of a chess game. You somehow don't feel surprised that
Chess brings to the fore so many of the disagreeable traits of competitive
intercourse. A sort of, Chess is to Go, as karate is to judo; yet in all four the
rules are set out and followed on pain of disqualification.
The "one-move blunder" comment in particular is amusing. True in chess the
o-m-b is the devoutly to be wished thing that you want to happen to your
opponent; which if anything highlights a certain narrowness in the game.
In Go, as far as I can see, a good move - if recognised by an opponent as
being 'good' - is a threat, a foundation, his potential nemesis, which has to be
treated and treated in such a way that at some time, either immediately or at
some time during the rest of the game its 'goodness' is nullified.
For real newbie players, watching a well-matched game between two
outstanding professional players, may not be the way to detect the ebb and flow
of such a game of Go. For them that detection might only happen with replaying
the game later. That flow will be more immediately clear to co-equals in the
audience and lets say, a little less so to the higher level amateur watchers. It
is not immediately clear to me without work.
With chess on the other hand there is firstly the initial audience
recognition of whichever particular opening gambit has been played. Then the game
arrives at the inevitable one-move-blunder. That's pretty much always the way of
it; only rarely is there a great comeback. Even then you sometimes think, if one
of those puts in an appearance, that it is was more a happy chance than
happenstance. Of course the chess players will deny this. On the other hand that
ebb and flow to me seems the very soul of Go.
Then what do I know.

Harry.


> and remains, international chess.



  
Date: 03 Oct 2007 17:35:55
From: -
Subject: Re: Computer Science Experts: Chess 'harder' than Go

> raylopez99 wrote:
>> ... --even though, for each move, less thought is involved.

Michael Alford <[email protected] > wrote:
> utter nonsense


Agree. We don't need to discuss "what is thought" in order
to build a game-playing device.



>> From the author's paper, page 1519: "Checkers [and by
>> extention, chess, mentioned in the article--RL] ...requires extensive
>> skill [<--note that phrase--RL] to make strong move choices...and
>> [has] moderate space complexity... All the games solved thus far have
>> either low decision complexity (Qubic and Go-Moku), low space
>> complexity (Nine Men's Morris...Awari).. or both (Connect Four)."
>>
>> [ ... ]

> Total bullshit.


First of all, in order to discuss the Science Magazine article we
should have the author names cited. Next, at risk of treading upon
"sacred temple territory" of these Holy Go Spaces, the considerations
expressed are sound. Frank de Groot has also been here, a prophetic
champion of "gate arrays." There exists a hardware solution to the
problem of exponential fan-out, albeit that the hardware solution is
still elusive. Disinterest by much of a computer artificial intelligence
community should not be misinterpreted to indicate severe impossibility.
Instead we could examine the sociological aspect: what led computer
scientists first to attack Chess instead of Go? Was there a perception
that Chess is more highly respected as a decision-rendering tool for
purposes of conducting our state-of-affairs in World Diplomacy?

As the Science Magazine authors note, a solved game may fall due
to low decision -OR- low space complexity. The decision side is more
difficult to visualize, and to devise in hardware, however it remains
true that the impenetrability of a game to machine analysis is the
result of a combination of both and not simply one aspect or the other.
The fact that many millenia of human history had elapsed prior to atomic
weapons does not invalidate the reality that atomic weapons exist today.
When the hardware devices are invented people may come to regard
that scientific team with awe and respect, as was accorded to Einstein.

Go Players unprepared for the prospect that their favorite sanctuary
will someday be invaded, and overrun by electronic hoardes, share an
immaturity aspect which the Chess Community has long ago shed. It
will be interesting to observe how many Go Players continue once those
hardware devices gain an advantage: this would surely be a test of faith
once the rubber meets the road. Who prefers not to live to see that day?




- regards
- jb


-------------------------------------------------------------------
Tamiflu mutates in sewage systems ...
http://afp.google.com/article/ALeqM5johTPaZnrC32MSU6WyDgsQL4KcGA
-------------------------------------------------------------------