How not to be thought of as an arse
richard mullens
Concentrate on these who have the motivation and the money to hurt you.
Bully not those who would spread windoze software on p2p sites
Frank de Groot
> There are many other rapacious capitalists on this group.
>
> Concentrate on these who have the motivation and the money to hurt you.
> Bully not those who would spread windoze software on p2p sites
If any of my bullies spread my software on a p2p group, I will spread ALL of
my bullies limbs over the pavement.
How about that?
richard mullens
Wrong target.
Just like the RIAA and others you would try to go for the weak and vulnerable who would never purchase your software in the
first place.
If, and I haven't looked, Moyogo were available on p2p sites, and it is as good as you make out, then anyone with a serious
intent to study would very likely buy the supported version.
These would be your friends who would spread the word.
Instead, your ravings put people off the idea.
Frank de Groot
> If, and I haven't looked, Moyogo were available on p2p sites, and it is as
> good as you make out, then anyone with a serious intent to study would
> very likely buy the supported version.
> These would be your friends who would spread the word.
>
> Instead, your ravings put people off the idea.
I know. But I can't help that most people are stupid.
Not that I'm too worried - my sales are slowly increasing, month-by-month.
And my market is SE Asia. Those 4d players here that are bashing me because
their ego feels threatened, are considered hopeless beginners in my target
market. Wannabe's. Fake Go teachers. Disgruntled scriptkiddies.
In the west, half of all Go players are computer programmers and most of
them would like to drink my blood out of sheer envy.
This won't be the case in SE Asia. And if I sell 15 copies/month here, I'll
sell 100 copies/month there. Eat your heart out.
I will simply focus my efforts on that market, and forget about you guys
here.
You simply do not matter in the big scheme of things.
You have no sense of honor, you are without culture and most of you are a
disgrace to the Go community.
(I'm talking about 90% of all contributors to rec.games.go)
But you're right, my competitors deserve to be crushed after what they've
done to me, and I'll make sure that happens.
Jonathan Schaeffer
>
> I know. But I can't help that most people are stupid.
> Not that I'm too worried - my sales are slowly increasing, month-by-month.
Hi Frank
So have you considered to GPL your software ? I can't see anywhere on
your product page a reference to any license. It would be *great* and
maybe one could work on availability for other patforms, etc. If the
software is Good, there would be a programmers community, contributors etc.
jo
Frank de Groot
> So have you considered to GPL your software ? I can't see anywhere on your
> product page a reference to any license. It would be *great* and maybe one
> could work on availability for other patforms, etc. If the software is
> Good, there would be a programmers community, contributors etc.
There are several problems to GPL-ing my software.
1. It's in Pascal
2. It uses a huge number of very expensive (of course Copyrighted) 3rd party
libraries
3. My life goal was to become an independent Go software developer and build
a very strong Go program at all costs (all I need is money to eat and a roof
over my head in a cheap country)
But if I have to, in order to destroy my enemies, I might even put my Go
software for free on the internet, one day, just to cost MasterGo and
SmartGo sales.
But this will only be many years from now. It depends on how my current
customers defend my software on public fora. If my customers do not care, I
might diversify my development efforts into non-Go related fields (there's
much more, money to be made there but it's boring for me)
As soon as I decide to do that, I'll add as much SmartGo-like functionality
and throw my software for free on the internet.
But this may never happen, or it will take a decade or so, who knows.
Dav...@gmail.com
> their ego feels threatened, are considered hopeless beginners in my target
> market. Wannabe's. Fake Go teachers. Disgruntled scriptkiddies.
>
> In the west, half of all Go players are computer programmers and most of
> them would like to drink my blood out of sheer envy.
> This won't be the case in SE Asia. And if I sell 15 copies/month here, I'll
> sell 100 copies/month there. Eat your heart out.
First, what the hell do you know about 4d players?
Second you cant see the future, and dont pretend to. If your so adament
about go in SE asia, then please learn to speak/write mandarin, and
join a chinese go discussion group.
Frank de Groot
> First, what the hell do you know about 4d players?
I know that 4d is ridiculously low.
Any stupid kid can become 4d in a few years.
Vit Brunner is living proof of that.
> Second you cant see the future, and dont pretend to. If your so adament
> about go in SE asia, then please learn to speak/write mandarin,
I eat mandarins, does that qualify?
> and
> join a chinese go discussion group.
I don't think you understand.
All I need is to get a few more Korean & Japanese translation strings, a
Japanese & Korean cardboard box and start selling to Japanese/Korean
retailers.
Then I'll laugh in your faces.
You can delay me a little, but ultimately I'll prevail.
Vit Brunner
> I know that 4d is ridiculously low.
You are not the first 20kyu to think so. Many 20kyu people (usually
after having seen Hikaru no Go) even think that they will become
professionals in two years and win a major title in four years.
> Any stupid kid can become 4d in a few years.
> Vit Brunner is living proof of that.
Yeah, right.
Frank de Groot
> You are not the first 20kyu to think so.
I am not a 20 Kyu.
I am a 25 Kyu.
Robert Jasiek
Why not spend 10 - 20 hours of playing and being taught to become 15
kyu?
--
robert jasiek
Mef
Frank de Groot wrote:
> <Dav...@gmail.com> wrote
>
> > First, what the hell do you know about 4d players?
>
> I know that 4d is ridiculously low.
> Any stupid kid can become 4d in a few years.
> Vit Brunner is living proof of that.
>
I don't recall offhand where I read it, but I'm pretty sure I've heard
that of the people who start playing, and at least keep with it for
some time, about 1 in 10 will make shodan. I don't think any
experienced go player would call 4d "ridiculously low."
Cheers,
Mef
Frank de Groot
> Why not spend 10 - 20 hours of playing and being taught to become 15
> kyu?
To be honest I don't know my rating.
I do not consider it important.
My last game I won from a regular club player, no idea what her rank was.
She used to wipe the board with me but after a handful of games I started
winning.
I have no time to play Go now but I am taking lessons in practicing global
concepts in 2007, likely from A. Dinerchtein.
My upcoming Go playing program needs some detectors for its learning stage
and I need to be sure I am not overseeing things.
Frank de Groot
>some time, about 1 in 10 will make shodan. I don't think any
> experienced go player would call 4d "ridiculously low."
I'm looking at it from an absolute POV.
From the POV Go programs will play at, one day.
A dozen stones stronger than 4d.
Mef
You're assuming that it's even possible to get a dozen stones stronger
than a 4d. Most theorists predict that current top professionals are
about 3-4 stones from perfect play. 4d is about 4 stones away from
professional play, if we even say there is as much as 6 stones between
a 4d and a top pro that still leaves your perfect playing computer only
10 stones stronger....and that's assuming your computer is perfect.
Cheers,
Mef
Frank de Groot
> You're assuming that it's even possible to get a dozen stones stronger
> than a 4d. Most theorists predict that current top professionals are
> about 3-4 stones from perfect play.
They're full of shit.
> 4d is about 4 stones away from
> professional play, if we even say there is as much as 6 stones between
> a 4d and a top pro that still leaves your perfect playing computer only
> 10 stones stronger....and that's assuming your computer is perfect.
That's assuming your theorists did not make a 60% error in their
guesstimation.
Mef
The strength to stone correlation is not linear, so the guesstimation
of perfect play strength vs. stones will be reasonably forgiving to
error. Also I think the assumption that programs will achieve perfect
play in any timespan that concerns me is a considerably larger
assumption. Even with all this aside, I would hardly consider within
12 stones of perfect play "ridiculously low," which was the original
issue at hand.
Denis Feldmann
>
>
>> You're assuming that it's even possible to get a dozen stones stronger
>> than a 4d. Most theorists predict that current top professionals are
>> about 3-4 stones from perfect play.
>
>
> They're full of shit.
>
As usual, the expert has spoken.
>> 4d is about 4 stones away from
>> professional play, if we even say there is as much as 6 stones between
>> a 4d and a top pro that still leaves your perfect playing computer only
>> 10 stones stronger....and that's assuming your computer is perfect.
>
>
> That's assuming your theorists did not make a 60% error in their
> guesstimation.
Your estimation, of course, being necessarily much better
>
>
Frank de Groot
> assumption. Even with all this aside, I would hardly consider within
> 12 stones of perfect play "ridiculously low," which was the original
> issue at hand.
I have to concede that my "ridiculously low" statement was a little
influenced by the various statements Mr. Brunner made in regard to me.
But in my opinion, 9p players are very far from perfect play (half a dozen
stones perhaps).
Because computers will be able to use even more of an advantage in Go than
they now do with Chess - they'll set up very complex, deep tactical traps
and the nature of Go is such that this is a disadvantage for the human.
Frank de Groot
>> They're full of shit.
>>
>
> As usual, the expert has spoken.
Indeed.
You forgot that my profession is comp. Go engine design and I am required to
read all relevant academic publications with regard to these matters.
You don't have to have wings on your back in order to know something about
aviation.
> Your estimation, of course, being necessarily much better
Of course my estimation is better.
I'm not a theorist.
We have a saying: "Those who can, do, those who can't, teach".
I would like to add: "Those who can't teach, become theorists" (a bit like
jb).
Vit Brunner
> "Mef" <mwille...@gmail.com> wrote
> > some time, about 1 in 10 will make shodan. I don't think any
> > experienced go player would call 4d "ridiculously low."
> From the POV Go programs will play at, one day.
> A dozen stones stronger than 4d.
How can you be so sure? (note for people who like twisting my
sentences: I didn't say that the programs won't be that strong, I just
don't understand how someone can be so sure)
Frank de Groot is ridiculously incompetent programmer (from the POV of
very advanced AI that will be able to create programs hundreds times
faster and dozen times better then Frank).
Mef
Frank de Groot wrote:
> But in my opinion, 9p players are very far from perfect play (half a dozen
> stones perhaps).
>
> Because computers will be able to use even more of an advantage in Go than
> they now do with Chess - they'll set up very complex, deep tactical traps
> and the nature of Go is such that this is a disadvantage for the human.
I'll agree that pros are far from perfect play...however when you look
at that board with 6 black stones on it, and think about the ability of
current pros...I don't think any computer would stand a chance.
Denis Feldmann
And I would add that Frank's opinions about the nature of Go are, shall
we say, maybe a little bit lacking in personal experience (and if he has
been reading comp.go, he should realize results so far tend to prove
that "experts" (in go programmation) are not necessarily perfectly
accurate. Professional goexperts could be nearer the mark)
Frank de Groot
> I'll agree that pros are far from perfect play...however when you look
> at that board with 6 black stones on it, and think about the ability of
> current pros...I don't think any computer would stand a chance.
But a computer would turn it into a most complex tactical mess, designed to
be as hard as possible for human players.
No striving for balance, but attack, attack attack and always designed to be
as complex as possible for humans.
I have a hunch that high-level comp. Go play will be designated by Pro
players as "incredibly vulgar" etc.
Frank de Groot
> And I would add that Frank's opinions about the nature of Go are, shall we
> say, maybe a little bit lacking in personal experience (and if he has been
> reading comp.go, he should realize results so far tend to prove that
> "experts" (in go programmation) are not necessarily perfectly accurate.
> Professional goexperts could be nearer the mark)
You're bulshitting illogically again.
The fact that you have a low opinion of some folks posting on some group,
have nothing to do with the validity of my statements.
Go cultivate some mushrooms or something.
Stop burping here when you're drunk.
Frank de Groot
> How can you be so sure?
After several days of being occupied by troll-hunting, I now am in "the
zone" again, so I'm afraid I don't have time for little kiddies this
evening.
Rest assured that the things I am sure about, I have reason to be sure
about.
In this case, I don't think you will be able to follow my reasoning. You're
a little kiddie, after all.
PS
Beware of the Brno gypsies.
I hear they usually don't attack the locals, but you never know.
Bantari
Mef wrote:
> I don't recall offhand where I read it, but I'm pretty sure I've heard
> that of the people who start playing, and at least keep with it for
> some time, about 1 in 10 will make shodan.
You probably read it on the GoDiscussions forum because there was a
thread discussing just that recently.
My impression is, however, that most people (90%?) fail to reach shodan
not because it is such an advanced level and you need to be in the
upper 10% with your IQ to reach it. More lilely, to me, is the
explanation that to reach shodan you need to invest some extra time and
effort in studying Go, and 90% of the people who pick up the game are
unwilling (or unable - work, kids, etc) to do that.
____________
-Bantari
http://www.bantari.net/
Robert Jasiek
>Most theorists predict that current top professionals are
>about 3-4 stones from perfect play.
I say: 16-17 stones.
--
robert jasiek
Frank de Groot
> I say: 16-17 stones.
I'm not sure whether you're joking but in fact I think that you are in the
correct order of magnitude here.
I just did not want to sound too extreme, claiming two stones above the
average theorists' estimation already got me a bunch of flames by the usual
suspects.
Bantari
What is the basis for your estimate?
____________
-Bantari
http://www.bantari.net/
Mef
I'm curious as to how you arrived at that, what makes you think the gap
is so large?
Mef
This may have to do with the fact that you are not an experienced go
player, so your opinions on the matter of go theory are not exactly
held in high regard. Mr. Jasiek however is an experienced go player.
One would expect him to have a much more sound basis for his conclusion.
Robert Jasiek
<fr...@moyogo.com> wrote:
>> I say: 16-17 stones.
>I'm not sure whether you're joking
I claimed these numbers here already years ago. I am not joking. From
studying quite some local positions with kos, I have learned just how
little we amateur or professional players understand about ko fights
and tactics. Analysing such positions for Japanese rules requires
perfect play and I could notice just how easy it is to make reading
mistakes and oversights even in just the latest endgame. A board with
only 17 handicap stones leaves enough room to contruct similar local
ko shapes until the middle game. If you are forced to play an
unintended ko, the best you can hope for is an exchange. Locally, you
have already failed by overlooking the possibility of a ko shape there
at all. Of course, the omniscient player has more strategic means
available like complex cutting as a global strategy. Even if each
potential cut is just a binary question like "Defend or let the cut be
possible?", 20 cuts like that create a major decision complexity of
2^20, which is far beyond human reading during the game. Combine this
with kos - and the top 9p player is hopelessly lost if he took just a
few handicap stones. With many handicap stones, there is at least a
chance that he might get away with a "connect your handicap stones"
strategy, as you would recommend it to a kyu player in a 9-stone game.
--
robert jasiek
Robert Jasiek
>What is the basis for your estimate?
The actual numbers are wild guesses from looking at some positions
with that many handicap stones. For the principle rough magnitude, see
my other reply.
--
robert jasiek
Vit Brunner
> > about 3-4 stones from perfect play.
>
> I say: 16-17 stones.
I find it ridiculous to talk about this without mentioning time limits
or what "perfect play" means (it could mean always playing objectively
best possible move or trying to trick the opponent abusing his weak
points, I think there is quite a big difference between the two).
Frank de Groot
> This may have to do with the fact that you are not an experienced go
> player, so your opinions on the matter of go theory are not exactly
> held in high regard. Mr. Jasiek however is an experienced go player.
> One would expect him to have a much more sound basis for his conclusion.
People should stop judging things they don't understand on the basis of "We
do not respect you".
It is my firm conviction based on the FACTS and LOGIC that the perfect
player is many stones above 9p and not just a few, as "the theorists" claim.
Funny how time and time again the opinions-based-on-emotions of the trolls
here are kicked from under their petty little monkey-ego's.
If someone would face me with a modicum of respect, I would explain myself,
instead of calling peopel names as they deserve.
I am amazed that anybody thinks that pro players are anything else but
utterly incompetent against a perfect player, and I am amazed that people
think that a perfect player is a few stones above 9p.
From my research and logic reasoning, I conclude that a 9p has no clue what
he's doing, sorry to say.
I am not a go player but the ecidence is clear. Robert explained it, and I
might have read his explanation a year ago and have used it as part of my
conviction.
Mef
In theory time limit would be irrelevant (though in practice they would
be significant). However deciding between the objective best move and
the subjective "most likely to win against this opponent" is an
interesting question...
Frank de Groot
> I find it ridiculous to talk about this without mentioning time limits
Utter nonsense.
Think about what you just said and then shut up.
> or what "perfect play" means (it could mean always playing objectively
> best possible move or trying to trick the opponent abusing his weak
> points, I think there is quite a big difference between the two).
It does not matter. Point is, it's theoretically possible to play a dozen
stones stronger than current 9p, against a current 9p.
This is where the understanding of the 25 Kyu is better than the
understanding of the 4d.
And this is what counts, in making a strong Go engine, not how good the
programmer plays.
Bantari
Robert Jasiek wrote:
> Of course, the omniscient player has more strategic means
> available like complex cutting as a global strategy.
Snipping the rest of the explanation (you *are* joking, right?) one
could argue that an omniscient (?) player would have no need for
strategy at all. Or am I misinterpreting your definition of
"omniscient"?
__________
-Bantari
http://www.bantari.net/
Bantari
Mef wrote:
> However deciding between the objective best move and
> the subjective "most likely to win against this opponent" is an
> interesting question...
In an even game, I am not sure the difference is meaningless.
Objectively "best-move" would win, and so would a
"best-move-just-against-this-player". At least as far as "perfect play"
is concerned.
Practically?...
In handi game, i think you are right. It is interesting. By observing
my own handi games, I see that I play different than I would in even
games (or low handi games). I try to be more "tricky" and often choose
sequences which are more difficult to answer correctly. I know that
this is probably not the "right" approach, but it seems to give me the
best results against handi games.
In even games, when I play seriously, I try to always go for what I
believe is the best move. Although even then, when I know that a
particular player is uncomfortable with a particular aspect of the
game, I try to use this knowledge to my advantage.
A good example is the anti-compute strategies, I think. :)
_________________
-Bantari
http://www.bantari.net/
Frank de Groot
> Snipping the rest of the explanation (you *are* joking, right?) one
> could argue that an omniscient (?) player would have no need for
> strategy at all. Or am I misinterpreting your definition of
> "omniscient"?
He isn't joking.
I was afraid to say someting similar but I agree with him.
A perfect player is more than a dozen stones stronger than a 9p.
I can't explain it as well as Robert but he is right.
I am glad he came here to explain his opinion.
It is a very important issue.
I never thought too much about the Ko-factor,, I thought "tactics" but
Ko-issues are the ultimate tactical nightmare.
Dav...@gmail.com
defeated by an intuitive sense of things like danger, shape, influence,
urgency and 'walking'. Are you saying that this go playing program
(hypothetical?) is going to have intuition?
On Dec 14, 1:35 pm, "Frank de Groot" <f...@moyogo.com> wrote:
> "Mef" <mwille...@gmail.com> wrote
>
> > assumption. Even with all this aside, I would hardly consider within
> > 12 stones of perfect play "ridiculously low," which was the original
> > issue at hand.I have to concede that my "ridiculously low" statement was a little
Frank de Groot
> urgency and 'walking'. Are you saying that this go playing program
> (hypothetical?) is going to have intuition?
Intuition is the result of an algorithm, both in humans and in computers.
Read some relevant literature, like "On Intelligence" by Jeff Hawkins.
Don't always presume you are God's gift to computer science and that your
discussion opponent is a moron (as you called me).
I may be a moron, but I'm a moron with a few hundred kilo science
books/publications on my bookshelf, and the intellect to make sense of it
all.
That is not boasting, that is pre-empting retorts like: "books mean nothing
if you don't understand them".
I don't know who/what you are, but all you've shown here is infantile
incomprehension combined with a huge ego.
Just shut the fuck up already.
Nobody cares about your opinion.
Go play with your comic books or your camera or whatever kids do to
entertain themselves these days.
-
> clue what he's doing, sorry to say.
Even more so, what the 9k concludes that the 9p is doing ?
"Robert Jasiek" <jas...@snafu.de> wrote in message
> ... A board with only 17 handicap stones leaves enough room to
> contruct similar local ko shapes until the middle game.
The discussion was about who wins by territory, not whether
it is possible to make a living shape nor what a 5d thinks of a 9p.
"Frank de Groot" <fr...@moyogo.com> wrote in message
> A perfect player is more than a dozen stones stronger than a 9p.
Not more than 4 or 5 stones, according to the theoretical math.
Robert neglected to compute the square root of his estimates.
- regards
- jb
-----------------------------------------------------------------
http://www.jacquielawson.com/viewcard.asp?code=ER13610552
-----------------------------------------------------------------
Frank de Groot
> Robert neglected to compute the square root of his estimates.
No he didn't.
Read his explanation, oh armchair clueless Go theorist.
Michael Alford
>
> "Frank de Groot" <fr...@moyogo.com> wrote in message
>> A perfect player is more than a dozen stones stronger than a 9p.
>
>
> Not more than 4 or 5 stones, according to the theoretical math.
> Robert neglected to compute the square root of his estimates.
>
>
>
> - regards
> - jb
>
Grist for the mill: I've forgotten where I read this, but in an
interview, a 9p was asked how much handicap he'd want playing against
God. The 9p said 3, unless he was betting his life, in which case he'd
ask for 4 :)
Michael
Dav...@gmail.com
On Dec 14, 4:30 pm, "Frank de Groot" <f...@moyogo.com> wrote:
> <Davo...@gmail.com> wrote
>
> > urgency and 'walking'. Are you saying that this go playing program
> > (hypothetical?) is going to have intuition?Intuition is the result of an algorithm, both in humans and in computers.
If this were true, then compscience would be much nearer to true AI
than it is now. But given that I'm not a programmer/inteligence
theorist I'm very willing to see some proof of this 'intuitional
algorythm'. Untill then I'm going to have to assume that humans are
capable of something very finite that computers just cannot do.
> Don't always presume you are God's gift to computer science and that your
> discussion opponent is a moron (as you called me).
I never claimed to know anything big about computer science. In fact,
the closest thing to proramming I've ever done was setting up pipes. If
you dont want to be regarded as a moron refrain from straw man argument
tactics (note you saying I assumed I was gods gift)
> I may be a moron
Close frank, your almost right... May be a moron is only an inch away
from the truth; you are a moron. Your also arrogant, I remember reading
somewhere on your blog that the doctor treating your heart was all but
useless, and that you only needed her as a leaping off point to make
your own treatments. You've also cited yourself as a scientist, english
teacher, professional photographer and god knows however many more
things you've claimed to be here.
>but I'm a moron with a few hundred kilo science
> books/publications on my bookshelf, and the intellect to make sense of it
> all.
They've also taught an ape to use sign language; congradualtions.
> That is not boasting, that is pre-empting retorts like: "books mean nothing
> if you don't understand them".
Pre-empting against arguments people havent made is another straw man
tactic.
> I don't know who/what you are, but all you've shown here is infantile
> incomprehension combined with a huge ego.
Actually, I asked you a question about something you claimed to know
about. What you've done is presented yourself as someone who knows
about computer go, and then resorted to name calling when someone asked
you about the topic. Because of this, we can just infer that you done
actually know what your talking about and that you lied about any
understandings you have.
> Just shut the fuck up already.
> Nobody cares about your opinion.
Actually, given that my internet access is _still_ running fine I'm
going to speak here as much as I want. Also, judging by the responces
I've generated with my topics, people do care about what I have to
say... Wether they agree or not is a different matter, but to anyone
who wants to argue with me and (unlike you) wont just start calling
people names (and threaten to release their army of jypsies) is going
to be recieved well by me.
No one cares about my oppinion? Then why were you among the FIRST to
reply to my questions about copyright? Dont worry frank, no one expects
you to answer that, we all know you dont like logic.
> Go play with your comic books or your camera or whatever kids do to
> entertain themselves these days.
How about playing go? Or have you forgotten that this is R.G.G. and not
a public forum for your own
ranting/lying/insults/threats/biggotry/arrogance/etc.
Dav...@gmail.com
calling you arrogant =D)
I'm not really strong (6k) but from my modest understanding of go, I'm
under the impression that perfect play is impossible on the go board...
I imagine that every position can be attacked effectively, provided the
person doing the attacking is clever enough to see it. Likewise, each
position can be defended in the same respect... Is there any reason why
this standpoint is refutable?
On Dec 14, 6:03 pm, Michael Alford <m...@aracnet.com> wrote:
> - wrote:
>
> > "Frank de Groot" <f...@moyogo.com> wrote in message
> >> A perfect player is more than a dozen stones stronger than a 9p.
>
> > Not more than 4 or 5 stones, according to the theoretical math.
> > Robert neglected to compute the square root of his estimates.
>
> > - regards
> > - jbGrist for the mill: I've forgotten where I read this, but in an
Reinhold Burger
On Thu, 14 Dec 2006, Michael Alford wrote:
> Grist for the mill: I've forgotten where I read this, but in an
> interview, a 9p was asked how much handicap he'd want playing against
> God. The 9p said 3, unless he was betting his life, in which case he'd
> ask for 4 :)
Hello Michael:
For the record, the 9p seems to be Otake Hideo; the statement appears
in "The Magic of Go", column 157, by Rob van Zeijst. You can find it
through the "browsable archive" at
http://shinbo.free.fr/index.php?page=tmog&lang=en
Quote:
----------------------------------------------------------------------
How strong would the perfect go player be compared to the reigning
champions of our time? This may seem like a ridiculous question, but
when asked, Hideo Otake, 9th-dan, a former Meijin titleholder and
still one of the strongest players in Japan, replied as follows:
"Playing God, I think I would probably have a good chance playing with
a handicap of three stones, but if I had to stake my life on it, I
would probably take four stones. If I still lost, that would be too
bad." With his honor as a professional at stake, Otake would rather
risk his life than take more than a four-stone handicap.
----------------------------------------------------------------------
The claim is also mentioned in Senseis Library (which cites the Magic
of Go column as its source):
http://senseis.xmp.net/?DivineGoProblem
Reinhold
Michael Alford
Thanks for the info.
-
"Dav...@gmail.com" <Dav...@gmail.com> wrote:
> I'm not really strong (6k) but from my modest understanding of go,
> I'm under the impression that perfect play is impossible on the go
> board... I imagine that every position can be attacked effectively,
> provided the person doing the attacking is clever enough to see it.
> Likewise, each position can be defended in the same respect... Is
> there any reason why this standpoint is refutable?
Again, Go is not about its deconstructed notions (position,
handicap, _ko_, life and death, etc.) but about the final score
which is based upon area, or upon territory and prisoners. It
is not possible to play perfectly, of course, but this one is not a
discussion about what is possible. None of the pros have any
idea of how well God (perfect player) would play.
---------------------------------------
> "-" <jazze...@hotmail.com> wrote:
>> Robert neglected to compute the square root of his estimates.
"Frank de Groot" <fr...@moyogo.com> wrote in message
> No he didn't.
> Read his explanation, oh armchair clueless Go theorist.
I did read again, but still found no reference to square roots.
Robert is out there flapping & yapping about _ko_ deconstructions.
------------------------------------------------------
"Reinhold Burger" <rfbu...@cs.uwaterloo.ca> wrote:
> For the record, the 9p seems to be Otake Hideo; the statement
> appears in "The Magic of Go", column 157, by Rob van Zeijst.
> You can find it through the "browsable archive" at
> http://shinbo.free.fr/index.php?page=tmog&lang=en
> The claim is also mentioned in Senseis Library (which cites
> the Magic of Go column as its source):
> http://senseis.xmp.net/?DivineGoProblem
With four, Otake Sensei would have a 50-50 chance of being dead,
if he were so silly as to bet his life upon the outcome against God.
- regards
- jb
--------------------------------------------------------------------
Danish road safety councils solves the speeding problem
http://www.speedbandits.dk/
--------------------------------------------------------------------
Robert Jasiek
wrote:
>I find it ridiculous to talk about this without mentioning time limits
Let us be generous: 10 hours per player plus byoyomi.
>or what "perfect play" means
This is a well-established term: The absolute value of the score in
favour of the currently moving player is chosen maximally.
--
robert jasiek
Robert Jasiek
> Not more than 4 or 5 stones, according to the theoretical math.
Are these just extrapolations of rating statistics?
--
robert jasiek
Robert Jasiek
>I play different than I would in even
>games (or low handi games). I try to be more "tricky" and often choose
>sequences which are more difficult to answer correctly. I know that
>this is probably not the "right" approach, but it seems to give me the
>best results against handi games.
>
>In even games, when I play seriously, I try to always go for what I
>believe is the best move. Although even then, when I know that a
>particular player is uncomfortable with a particular aspect of the
>game, I try to use this knowledge to my advantage.
I tried to define the devil (in contrast to god) earlier:
http://groups.google.de/group/rec.games.go/msg/8a2b269c9510d8c8?dmode=source&hl=de
The "definition" was a bit rough, but one could improve it.
--
robert jasiek
Robert Jasiek
>Snipping the rest of the explanation (you *are* joking, right?) one
>could argue that an omniscient (?) player would have no need for
>strategy at all.
His strategy is "reading and evaluating the complete game tree".
Reference to ko and cutting is just my naive approximation in terms of
us players' thinking to what might happen in practical application of
his strategy.
--
robert jasiek
Max
Frank de Groot wrote:
> "richard mullens" <mullensd...@ntlworld.com> wrote
>
> > There are many other rapacious capitalists on this group.
> >
> > Concentrate on these who have the motivation and the money to hurt you.
> > Bully not those who would spread windoze software on p2p sites
>
>
> If any of my bullies spread my software on a p2p group, I will spread ALL of
> my bullies limbs over the pavement.
> How about that?
1) Illegal
2) Immoral
3) Bullshit
T Mark Hall
<jas...@snafu.de> writes
>On 14 Dec 2006 12:13:53 -0800, "Vit Brunner" <vit.b...@gmail.com>
>wrote:
>>I find it ridiculous to talk about this without mentioning time limits
>
>Let us be generous: 10 hours per player plus byoyomi.
But who would have the courage to tell God that He's lost on time?
>
>>or what "perfect play" means
>
>This is a well-established term: The absolute value of the score in
>favour of the currently moving player is chosen maximally.
>
Best wishes.
--
T Mark Hall
http://www.gogod.demon.co.uk/index.htm
http://www.gogod.demon.co.uk/NewInGo/NewInGo.htm
http://www.gogod.demon.co.uk/CoffeeBreak/CBProblem.htm
Denis Feldmann
against any position (in the actual play conditions, of course, i.e it
would not necessarily be the same move if played early, or after hours
of thought), and he minimaxes (is ther such a verb?) on that. (if the
rules are not perfectly defined, he may, as a last trick, use any
ambiguities at his advantage too, of course) On the other hand, he is
still playing fair (the Devil is a gentleman), ie not striking the
player with fits of dumbness, for instance :-)
On the other hand, I bet not only my life, but anything else, that even
a beginner knowing a few things about the rules (and the easy strategy)
cannot lose (even against the devil above) on 10 stones handicap with
free placement (the simple mathematical proof will be given later, if
necessary; note I didn't claim to be able to win, but only not to lose :-))
Denis Feldmann
>
> On the other hand, I bet not only my life, but anything else, that even
> a beginner knowing a few things about the rules (and the easy strategy)
> cannot lose (even against the devil above) on 10 stones handicap with
> free placement (the simple mathematical proof will be given later, if
> necessary; note I didn't claim to be able to win, but only not to lose :-))
Sllght precision to the above : this is only true in Japanese-style
rules ie without superko (why?). With superko rules, I am not so sure :-)
Frank de Groot
> If this were true, then compscience would be much nearer to true AI
> than it is now.
Utter bullshit.
You are lying.
It is a lie that all algorithms to do anything have already been invented.
>But given that I'm not a programmer/inteligence
> theorist I'm very willing to see some proof of this 'intuitional
> algorythm'.
Do you need proff the Earth is round too?
You freak me out!
If you are too stupid to understand something, how is proof going to
convince you?
That's like trying to prove to a NeoCon that labal warming exists.
He'll just say that the proof is invalid because he does not understand it
or he does not want to understand it.
> Untill then I'm going to have to assume that humans are
> capable of something very finite that computers just cannot do.
Neurons can be simulated, never heard of that?
> I never claimed to know anything big about computer science. In fact,
> the closest thing to proramming I've ever done was setting up pipes.
You should have chosen a career in unclogging drainage pipes, not lecturing
comp. sci Go AI folks.
> Close frank, your almost right... May be a moron is only an inch away
> from the truth; you are a moron. Your also arrogant, I remember reading
> somewhere on your blog that the doctor treating your heart was all but
> useless, and that you only needed her as a leaping off point to make
> your own treatments.
Exactly.
The bith told me that Borrelia Burgdorferi is a virus.
When I insisted it was a spirochete (bacterium), she started shouting at me
that she had been a cardiologist for 30 years and I should shut up.
If it wasn't so sad, it would have been funny.
Yes, I know much more about cardiac problems than cardiologists and yes, I
know much more about infectious diseases than Infectious disease
specialists, litte snotnosed boy.
You have a problem with that? You want to challenge my knowlegde instead of
calling me arrogant? Want to discuss things like modern diagnostic methods
for diagnosis & treatment of chrnoic infectious diseases? Novel antifungal
synergetic pharmaceutical compounds to treat systemic afungal infections?
Cochleates? Industrial synthesis of organic chemicals? Interpretation of
ECG's? The neurological factors of arrythmia's? (neurologically mediated
coronary vasospasms, tenth cranial nerve hyperstimulation etc.).
All the bith thought I needed was "Living yoghurt". While I was on the
intensive care. On oxygen. With three different arrythmia's.
Oh - she also said she was "very worried" and that it could be "serious".
She said: "I do not know yet what it is by try Biola yoghurt".
I have to come back to her today, report on the effect of the yoghurt, LOL.
I'll tell her it's the medulla oblongata, chronic inflammation by Bb, Nervus
Vagus hyperstimulation.
She'll tell me I'm crazy and that it's yoghurt-deficciency, LOL.
Listen little buddy - by B12 levels are normal, even high-scale normal.
My gut flora is normal too. I did not end up on the IC due to a yoghurt
deficciency, and Bb is NOT a virus.
> You've also cited yourself as a scientist, english
> teacher,
for a long time, I made money being a full-time ENglish teacher.
Calling me a liar again, chump?
> professional photographer
I made good cash selling my photo's of my travels to stock photograhpy
databases (agencies, they want 50 pictures, all need to be perfect in terms
of exposure time, contrast, focus, motion blur and frame selection).
I was a professional photographer starting at 16 years old, when I shot 360
pictures of a NATO excercise. We got arrested twice, running under landing
helicopters. I went into armored personell carriers, made closeups of comms
equipment, the FAL rifle, LAWs and a scene with a naked POW. A major
demanded the film in our camera's. I refused and kept filming. That's how
big sissies they were already then.
Oh - You never guess who paid for my photographs.
All I can say is that I had *very* good contact with Radio Moscow :-)
Frigging baby..
You think everybody is a snotnosed little punk as you?
You know how I avoided military service?
I said I would immediately execute the three highest-ranking officers in my
vicinity, as soon as they would give me the three live rounds.
For some reason I managed to persuade the shrink that it was better not to
lock me up.
And I am not afraid of people with guns either. Like I said, I have smuggled
people over the iron curtain.
I hate snotnosed little punks telling me what to do, that's all. Especially
snotnosed little punks with guns.
> and god knows however many more
> things you've claimed to be here.
Calling me a liar is not going to make you look bigger, little boy.
If you want to do something useful, go do it, don't go telling those you
actually do it "liars".
> Actually, given that my internet access is _still_ running fine I'm
Little boy, you crawled in the dust here and took everything back and
apologized, I immediately removed everything about you in my blog, what are
you trying to say?
> people names (and threaten to release their army of jypsies) is going
One gypsie will do.
Frank de Groot
> Grist for the mill: I've forgotten where I read this, but in an
> interview, a 9p was asked how much handicap he'd want playing against God.
> The 9p said 3, unless he was betting his life, in which case he'd ask for
> 4 :)
This only shows that he does not understand the game.
Frank de Groot
Litle babyfaced bitch - why did you quote a long posting w/o adding
anything?
Don't you know that's a violation of TOS?
And you are inaccurate.
It should be called "military espionage".
I was fifteen and a half years old - a minor.
LOL
Just trying to get my pre-emptive revenge on my country's army.
They wanted to brainwash me against the Russians - I thought it only fair to
provide the Russians with the latest info on the largest NATO excercise in
my country.
Nothing special - I'm sure the Russians had their info, and much better at
that - but I resent little snotnosed punks kidnapping and brainwashing me
into becoming a babykiller, don't you?
I can defend myself very well, thank you, and the Russians were our allies,
my country just didn't understand that at the time.
My country told me "we are going to lock you up with a bunch of morons in
military barracks, abuse you a little and brainwash you into becoming an
obeying killing machine".
So I did what I had to do to get my pre-emptive revenge, just in case.
I hate being threatened by bullies, especially with guns.
Max
Frank de Groot wrote:
> "Mef" <mwil...@gmail.com> wrote
>
> > This may have to do with the fact that you are not an experienced go
> > player, so your opinions on the matter of go theory are not exactly
> > held in high regard. Mr. Jasiek however is an experienced go player.
> > One would expect him to have a much more sound basis for his conclusion.
>
>
> People should stop judging things they don't understand on the basis of "We
> do not respect you".
>
> It is my firm conviction based on the FACTS and LOGIC that the perfect
> player is many stones above 9p and not just a few, as "the theorists" claim.
>
No, your conviction is based on your ego. You don't have facts and
logic on your side. That "we" don't respect you is a consequence of the
fact that you talk a bunch of garbage and fail to back it up. The fact
is that noone knows how many stones a pro could take against perfect
play, but some people have made reasonable theories from seeing certain
patterns. You have simply made an assumption.
> Funny how time and time again the opinions-based-on-emotions of the trolls
> here are kicked from under their petty little monkey-ego's.
>
> If someone would face me with a modicum of respect, I would explain myself,
> instead of calling peopel names as they deserve.
>
> I am amazed that anybody thinks that pro players are anything else but
> utterly incompetent against a perfect player, and I am amazed that people
> think that a perfect player is a few stones above 9p.
>
You are amazed and chagrined that people don't find you amazing.
> From my research and logic reasoning, I conclude that a 9p has no clue what
> he's doing, sorry to say.
> I am not a go player but the ecidence is clear. Robert explained it, and I
> might have read his explanation a year ago and have used it as part of my
> conviction.
What evidence? Have you yet written a program that can defeat a 1 dan
amateur, let alone one of these 4 dans you turn your nose up at?
Considering the lack of reasoning you've exhibited in the field (like
arguing that whether a stone is alive has no bearing on influence), I
see no reason to put any weight on your empty theories.
Max
> "Denis Feldmann" <denis.feldm...@club-internet.fr> wrote
>
>
> >> They're full of shit.
> >>
> >
> > As usual, the expert has spoken.
>
> Indeed.
> You forgot that my profession is comp. Go engine design and I am required to
> read all relevant academic publications with regard to these matters.
>
Yet you can't even read a memo or a post to usenet without making
egregious errors. Never mind the fact that the academic publications
show that we don't know enough to be certain about any claims of this
order.
> You don't have to have wings on your back in order to know something about
> aviation.
>
No, you need information about aerodynamics. To make adequate
predictions of performance (particularly at extremes not previously
travelled at) one needs wind tunnels, computer simulations, or working
prototypes. In the world of computer go there is no equivalent as yet.
> > Your estimation, of course, being necessarily much better
>
> Of course my estimation is better.
> I'm not a theorist.
> We have a saying: "Those who can, do, those who can't, teach".
>
> I would like to add: "Those who can't teach, become theorists" (a bit like
> jb).
In this matter you are a theorist. Theory is all we have at the moment.
That you don't even understand that puts your theories in question.
Max
> "Vit Brunner" <vit.b...@gmail.com> wrote
>
> > I find it ridiculous to talk about this without mentioning time limits
>
> Think about what you just said and then shut up.
>
> > or what "perfect play" means (it could mean always playing objectively
> > best possible move or trying to trick the opponent abusing his weak
> > points, I think there is quite a big difference between the two).
>
> It does not matter. Point is, it's theoretically possible to play a dozen
> stones stronger than current 9p, against a current 9p.
>
But I thought you weren't a theorist?
> This is where the understanding of the 25 Kyu is better than the
> understanding of the 4d.
>
> And this is what counts, in making a strong Go engine, not how good the
> programmer plays.
No, what counts is results. Results do not back you up on this.
Max
> > at that board with 6 black stones on it, and think about the ability of
> > current pros...I don't think any computer would stand a chance.
>
>
> But a computer would turn it into a most complex tactical mess, designed to
> be as hard as possible for human players.
> No striving for balance, but attack, attack attack and always designed to be
> as complex as possible for humans.
>
> I have a hunch that high-level comp. Go play will be designated by Pro
> players as "incredibly vulgar" etc.
Keyword="hunch". No actual facts or logic.
Max
Frank de Groot wrote:
> "Jonathan Schaeffer" <jonathan....@in2p3.fr> wrote
>
> > So have you considered to GPL your software ? I can't see anywhere on your
> > product page a reference to any license. It would be *great* and maybe one
> > could work on availability for other patforms, etc. If the software is
> > Good, there would be a programmers community, contributors etc.
>
> There are several problems to GPL-ing my software.
>
> 1. It's in Pascal
>
?...
Why?
> 2. It uses a huge number of very expensive (of course Copyrighted) 3rd party
> libraries
>
> 3. My life goal was to become an independent Go software developer and build
> a very strong Go program at all costs (all I need is money to eat and a roof
> over my head in a cheap country)
>
> But if I have to, in order to destroy my enemies, I might even put my Go
> software for free on the internet, one day, just to cost MasterGo and
> SmartGo sales.
Wouldn't hurt the sellers of MasterGo in the slightest. It's not a
major source of income for them. I can't say for sure, but I suspect
the same holds true for SmartGo.
>
Max
Frank de Groot wrote:
> "David Gillen" <Bel...@RedBrick.DCU.IE> wrote
>
> Litle babyfaced bitch - why did you quote a long posting w/o adding
> anything?
>
> Don't you know that's a violation of TOS?
>
> And you are inaccurate.
>
> It should be called "military espionage".
>
> I was fifteen and a half years old - a minor.
> LOL
>
> Just trying to get my pre-emptive revenge on my country's army.
> They wanted to brainwash me against the Russians - I thought it only fair to
> provide the Russians with the latest info on the largest NATO excercise in
> my country.
>
> Nothing special - I'm sure the Russians had their info, and much better at
> that - but I resent little snotnosed punks kidnapping and brainwashing me
> into becoming a babykiller, don't you?
>
You would much rather be one on your own.
> I can defend myself very well, thank you, and the Russians were our allies,
> my country just didn't understand that at the time.
> My country told me "we are going to lock you up with a bunch of morons in
> military barracks, abuse you a little and brainwash you into becoming an
> obeying killing machine".
>
The irony is, in Russia you would have gotten yourself locked away by
now. It will probably take a few more years where you are now.
> So I did what I had to do to get my pre-emptive revenge, just in case.
> I hate being threatened by bullies, especially with guns.
Yes, so unlike everyone else in that respect. But the effect is always
more pronounced in those who are themselves bullies. Particularly those
trying to cover up their own weakness and helplessness.
Frank de Groot
>> 1. It's in Pascal
>>
> ?...
> Why?
Guess!
(If you have to ask, you won't understand the answer)
> Wouldn't hurt the sellers of MasterGo in the slightest.
Every penny counts.
> major source of income for them. I can't say for sure, but I suspect
> the same holds true for SmartGo.
Now he lives off his millions, but in the future he hopes to live off
SmartGo. The market is a multi-mllion USD/year one, in SE Asia.
His personal fortune has diminished greatly over the past years (I refer to
the tax-evasion article in the newspapers to which I link in my blog about
Mr. Kierulf in which his personal finances were discussed).
Frank de Groot
> Keyword="hunch". No actual facts or logic.
Says who? "Max the genius"?
You seem to have forgotten that I am the Go software innovator here, buddy
:-)
Frank de Groot
>
> The irony is, in Russia you would have gotten yourself locked away by
> now. It will probably take a few more years where you are now.
Why?
You just said it never happened.
How can I be locked away for a "lie"?
Make up your mind, little kiddie.
Frank de Groot
> You have simply made an assumption.
That's the difference between myself and people like you.
I make correct assumptions.
That's why I invented the only working pattern system as applied to Go,
while every high-ranking player told me it would never work.
Who's the fool now? Every sinlge dan-ranking player who told me to shut up
and leave this field or learn Go better, first.
Little kiddies should keep their drooling traps shut and suck on their
lollies or their mommie's breast, not lecture me what to do.
> What evidence? Have you yet written a program that can defeat a 1 dan
> amateur, let alone one of these 4 dans you turn your nose up at?
OT
> Considering the lack of reasoning you've exhibited in the field (like
> arguing that whether a stone is alive has no bearing on influence), I
> see no reason to put any weight on your empty theories.
Good :-)
I am 100% reliant on that, to keep being without competition in this field.
My rescue is the boundless arrogance of language-fanatics, Go-elitists and
otherwise severly mentally impaired persons.
A pitty you're an anonymous little fool, otherwise we could make a bet :-)
Frank de Groot
> But I thought you weren't a theorist?
Better leave thinking to horses, they have larger heads than you do.
> No, what counts is results. Results do not back you up on this.
ROTFL
Go scientists worldwide have scrambled to copy my research...
It's that you keep on accusing me of being a failure in the Go science field
that I keep pointing this out, but you'll call me a "liar" or "arrogant",
LOL.
Frank de Groot
> The commerical libraries which you use as part of the program are only
> available for pascal?
There are no, or very, very few commercial libraries for Pascal only.
The vendors make a C++ version and port to Pascal, or vice versa.
Frank de Groot
> Inaccurate, I didn't say anything. It was simply a quote of what you said.
> Are
> you saying what you said is inaccurate.
For this lie, I will NEVER remove my blog posting on you.
Have it your way.
The title of your post was "military trespassing".
This is not something you quoted from me, this is your own allegation.
I said it was inaccurate.
You now claim I am a liar and that you never wrote a posting that you
entitled "military tespassing"-
Therefore you need a little more punishment.
I'll see what I can do to expand on my blog posting on you, and I might keep
it on top regularly in 2007.
If you are so desperate in calling me a liar, if you do not learn from your
mistakes, I'm more than happy to oblige.
Let's dig up some more personal data of yours and throw it all over the
Internet!
When do you people learn?
<sigh>
I do not like to be called a liar when I am not, or a criminal when I am
not.
Get it into your skull, all of you.
If you don't apologize for calling me a liar just now, I'll expand my blog
posting with personal facts about your wife Karen.
She is co-responsible for not restraining her cybercriminal spouse.
Calling someone a liar on Usenet is a criminal offence. It's libel.
The punishment is severe, and entails having one's private life spilled out
on my blog.
You are a recidivist criminal. You have no right to privacy any more.
Frank de Groot
> Frank. I apologise
OK so I will remove everything I posted on my blog about you now.
Frank de Groot
> Okay, so still none the wiser why pascal is your language of choice for
> moyogo.
Because XML won't do the trick.
Mef
Frank de Groot wrote:
> "Max" <maxd...@gmail.com> wrote
>
> > You have simply made an assumption.
>
> That's the difference between myself and people like you.
>
> I make correct assumptions.
> That's why I invented the only working pattern system as applied to Go,
> while every high-ranking player told me it would never work.
>
Pattern systems are in the realm of computer science, strength of go is
not applicable. In spite of that, I'm suprised that anyone said you
would not be able to devise a pattern matching program, especially if
you are as experienced a programmer as they said you are. What I
would find much more likely would be the claim that your pattern
matching would not necessarily translate into a strong go playing
program, which thus far has been true. As far as I know, no strong go
playing program has come from your pattern matching programs.
> Who's the fool now? Every sinlge dan-ranking player who told me to shut up
> and leave this field or learn Go better, first.
Save comments like this for when you have a dan level go program. So
far you have made advances in software, but not go or go AI.
Cheers,
Mef
Frank de Groot
> program, which thus far has been true. As far as I know, no strong go
> playing program has come from your pattern matching programs.
> Save comments like this for when you have a dan level go program. So
> far you have made advances in software, but not go or go AI.
Fine. Just remember that when I have my dan-level program, that I will go
through a lot of effort to expose you and others as arrogant
know-it-betters'.
Only fair. You call me deluded as to my own profession, if I have the
opportunity to expose your allegations as being false, I will spare no
effort either :-)
Vit Brunner
> For this lie, I will NEVER remove my blog posting on you.
On Dec 15, 4:59 pm, Frank de Groot wrote:
> OK so I will remove everything I posted on my blog about you now.
I wonder why the word "NEVER" was in caps...
Frank de Groot
> I wonder why the word "NEVER" was in caps...
I wonder, is your sister still a virgin?
Mef
Frank de Groot wrote:
> "Mef" <mwil...@gmail.com> wrote
>
> > program, which thus far has been true. As far as I know, no strong go
> > playing program has come from your pattern matching programs.
>
> > Save comments like this for when you have a dan level go program. So
> > far you have made advances in software, but not go or go AI.
>
>
> Fine. Just remember that when I have my dan-level program, that I will go
> through a lot of effort to expose you and others as arrogant
> know-it-betters'.
>
I'm not saying it can't be done, I'm not saying that your method won't
work. I'm just saying I'll believe it when I see it, and highly doubt
I'll be seeing it in the near future. Many people have tried this and
failed, and I'll be amazed if you manage to succeed before becoming a
reasonably strong go player (or at least have significant help from
one).
> Only fair. You call me deluded as to my own profession, if I have the
> opportunity to expose your allegations as being false, I will spare no
> effort either :-)
My allegations? I have alleged that thus far you have not made
advancement in the field or go AI. As far as I know that's true. What
will happen in the future? Who knows.
Cheers,
Mef
Bantari
Robert Jasiek wrote:
> His strategy is "reading and evaluating the complete game tree".
This is not "strategy", this is reading.
You don't need *any* strategic principles as I understand the term to
look up the whole tree and count the final scores at the end of each
branch.
It sort-of confirms what i always said - that "strategy" is nothing but
a crutch for those of us who are humans and thus are unable to read
deep enough. :)
> Reference to ko and cutting is just my naive approximation in terms of
> us players' thinking to what might happen in practical application of
> his strategy.
Right.
There are three problems I see with that, though.
Here they are:
---( 1 )---
It assumes that "perfect play" or "God" if you wish, would not really
make the best possible move in any situation. Instead, she would make
a move that leads to the largest confusion of her opponent. It has not
been show that this is the same as "perfect play".
For example - I can win against computers usually by making moves I
would never dare making against human players of the same nominal
strength. This is not because I play any closer to "perfect play", but
because I use a strategy that works against one particular opponent.
This is the basis of your argument in favor of top pros being far away
from "perfect player", right?
If this is so, I think we should say that the top pros might be long
way away from a "most tricky player". It is to be shown that such
player is the same as "perfect player". No?
Definition #1:
A "perfect play" is a play conducted by a "perfect player".
Definition #2:
A "perfect play" is a play which would lead to maximizing one's end
score assuming one is playing against another "perfect player".
Definition #3:
A 'most tricky player" is a player that makes moves leading to
maximizing her end score against a player which is assumed to have very
specific and well-known weaknesses. The play is heavily based on
exploiting these weaknesses, but might not necessarily be identical to
"perfect play".
---( 2 )---
Since we do not really know how close the pros are to "perfect play"
(in the sense of making the best move rather than the most complicated
move), such "strategy" might or might not be good. In short, we do not
know how far the "perfect player" diverges from the "most tricky
plyer".
In other words, if the top pros' moves are close to being perfect, they
should stand up against the "most tricky player" at some predetermined
handicap. It is not yet clear to me what this handicap is and how we
can determine it without actually knowing both the perfect play and the
most tricky play.
In yet other words, how many clearly exploitable weaknesses does a top
pro have?
Which brings me to point #3.
---( 3 )---
If the "perfect player" really strive to make the perfect move in each
situation, assuming that the opponent is another "perfect player", and
so such confuson-tactics with creating complicated positions would be
fruitless, how close would such play approximate the play of
present-day top pros? I guess this is what most of us intuitively
think of as a "perfect player" or "god". This assumption is based on
the premise that a "perfect player" is a player that makes an
objectively "perfect move" which should be the same in a position
without any regard as to who she plays against.
Which closes the circle and brings us back to point #1 and my
introduction of the term "most tricky player".
*****
What do you think?
__________________
-Bantari
http://www.bantari.net/
Denis Feldmann
> "Mef" <mwil...@gmail.com> wrote
>
>> program, which thus far has been true. As far as I know, no strong go
>> playing program has come from your pattern matching programs.
>
>> Save comments like this for when you have a dan level go program. So
>> far you have made advances in software, but not go or go AI.
>
>
> Fine. Just remember that when I have my dan-level program, that I will go
> through a lot of effort to expose you and others as arrogant
> know-it-betters'.
>
No problem. Still want to take that bet?
Michael Alford
> "Michael Alford" <ma...@aracnet.com> wrote
>
>> Grist for the mill: I've forgotten where I read this, but in an
>> interview, a 9p was asked how much handicap he'd want playing against God.
>> The 9p said 3, unless he was betting his life, in which case he'd ask for
>> 4 :)
>
>
> This only shows that he does not understand the game.
As Mr Burger pointed out, the 9p is Otake Hideo, saying he does not
understand the game is absurd, I will leave it at that.
The other day, Frank, I was chatting with some people about you, MoyoGo,
and rgg, and one of the people, a regular contributor to the group, who
knows programming and plays Go and has reviewed MoyoGo, called you a
"brilliant" programmer. Yeah, he said brilliant. So I take him at his
word on this, you really know what you're doing, as far as programming
goes. He then pointed out that in these threads on rgg, you argue from a
CS point of view, others argue from the point of view of Go players, and
both sides are "right", but will never, in this environment, get
together, and he described it as "sad". So all these arguments are comme
ci comme ca. What I would like to point out is that as good as you may
be at programming, you are not a Go player, and this does matter, how
can you invent algorithms for things you don't understand? You may not
like me saying this, but Go is a very deep game, there are levels and
levels of understanding, and until you learn to play and come to
appreciate these levels, you simply have no clue as to what you want or
need to program. I know several people that have MoyoGo, I know one of
the people that did one of the language translations, and all of them
are impressed by it, so it's a good learning and reference tool, well
worth the investment, but play? That's not going to happen, not unless
you learn to play*. I'm a low amateur dan, and no matter how good you
are at programming, if you don't understand the game, your program will
never beat me, let alone a strong player, no matter how sophisticated
your programming. My own opinion is that Go is a creative art form, and
computers will never play it well, but I wish you luck in trying to
prove me wrong :) Merry Yule and merry programming!
Best regards,
Michael
* for instance, how could you invent an algorithm for "using thickness"
if you don't know what it is or how to use it?
Bantari
> As Mr Burger pointed out, the 9p is Otake Hideo, saying he does not
> understand the game is absurd, I will leave it at that.
This is exactly what Isaid when Robert made similar claims about top
pros in general. Do not remember getting much support then.... :)
________________
-Bantari
http://www.bantari.net/
Frank de Groot
> As Mr Burger pointed out, the 9p is Otake Hideo, saying he does not
> understand the game is absurd, I will leave it at that.
He's extremely strong, that does not mean he understands how much stronger a
perfect player is, and that means he does not completely understand all
aspects of the game.
> be at programming, you are not a Go player, and this does matter, how can
> you invent algorithms for things you don't understand?
Because I do understand them. Not as a Go player but as a comp. Go
scientist/researcher.
I make software that learn Go from pro games.
> like me saying this, but Go is a very deep game, there are levels and
> levels of understanding,
Only from the human POV.
The same has been said of chess until someone put a ferw utterly simple
heuristinc in a hyper-fast tree searcher with clever pruning and voila,
Kasparov had a hard time defeating the thing. Go ain't any different, only
different way of approaching the problem.
> and until you learn to play and come to appreciate these levels, you
> simply have no clue as to what you want or need to program.
Utter nonsense, and utterly laughable.
It's fully possible for a 15 Kyu to make a 9p Go program. It's near-insane
to believe otherwise and you have no arguments for it except a gross
misunderstanding as what Go is.
Go is a state-machine.
>I know several people that have MoyoGo, I know one of the people that did
>one of the language translations, and all of them are impressed by it, so
>it's a good learning and reference tool, well worth the investment, but
>play? That's not going to happen,
I will drag you through the mud of your hyperinflated elitist ego and will,
as long as I live, demand your grovelling apologies.
By that time, you will be to ashamed to use your own name here.
> not unless you learn to play*.
You are severely mistaken.
Learning to play is a severe handicap and I will never do that.
I am a computer scientist and I have no time for games.
> I'm a low amateur dan,
Congratulations buddy. It means nothing to me.
I haven't got the slightest respect for even a 9p.
In my opinion, they're trained monkeys (no offence).
OK - let's be fair - they are talented, nice guys with enormous skill,
aquired over a long time.
But it's irrelevant to me. I need to know the concepts relevant to Go, not
to play Go.
I need to know which concepts EXIST. I already do. Don't get all elitist on
me.
"influence". Nobody can define it. I will define it, in my software, and it
will LEARN what influence is.
And finally, it will "know" what influence is, better than you.
Not just my program, others' as well.
Sorry, but Go is a state machine.
> and no matter how good you are at programming, if you don't understand the
> game, your program will never beat me,
Keep dreaming. Threatening to your ego, isn't it?
A 5 USD pocket calculator can now beat chess grandmasters and the same will
happen to you, Michael, with Go.
I will be the one doing it. As payment for your elitist egomania.
> let alone a strong player, no matter how sophisticated your programming.
Keep repeating this and you'll convince yourself.
It's a ridiculous "magical thinking" conviction but hey, each his religion.
> My own opinion is that Go is a creative art form,
Go is a state machine.
> and computers will never play it well,
Seek help.
> * for instance, how could you invent an algorithm for "using thickness" if
> you don't know what it is or how to use it?
I know exactly what thickness is and it's near-trivial to program it.
I have never spent even one day in trying to make a computer play Go, and my
Go STUDY program plays at the same strength as GNU Go, in a blitz game.
That's because my Joseki pattern-library turned out to be able to do that
LOL.
Wait until I build a Go playing program :-)
I know *exactly* how it should be done.
I mean into the finest detail.
And I'll drag your and everyone else's eelitist "Go is a creative art" go
through the mud when I'm done with it.
Go is a state machine.
Dav...@gmail.com
shut up...
at the very least, you will be behind a text edditor coding, and wont
be able to bother us ;)
Mef
Frank de Groot wrote:
> Only from the human POV.
> The same has been said of chess until someone put a ferw utterly simple
> heuristinc in a hyper-fast tree searcher with clever pruning and voila,
> Kasparov had a hard time defeating the thing. Go ain't any different, only
> different way of approaching the problem.
>
Chess benefits from the idea of positional chess, where you can quickly
evaluate a position independentally from future outcomes. If you can
find a way to effectively do this with go that will be an amazing feat,
since as it stands it's hard to even get quantifiable reasoning out of
strong players as for why one position feels better than a similar
positon....
>
> Go is a state-machine.
>
True, however, arriving at the position evaluation as a state function
has been difficult in the past.
>
> Learning to play is a severe handicap and I will never do that.
>
I'm not sure agree with this. One would not expect to solve a
difficult math problem without learning math, nor would one expect to
sure a disease without at least a fundamental grounding in medicine.
In trying to solve go, one would think it important to understand how
go works.
>
> But it's irrelevant to me. I need to know the concepts relevant to Go, not
> to play Go.
Wouldn't these be one and the same? Wouldn't knowing the concepts and
how they relate equate to being able to play?
>
> > * for instance, how could you invent an algorithm for "using thickness" if
> > you don't know what it is or how to use it?
>
> I know exactly what thickness is and it's near-trivial to program it.
> I have never spent even one day in trying to make a computer play Go, and my
> Go STUDY program plays at the same strength as GNU Go, in a blitz game.
>
> That's because my Joseki pattern-library turned out to be able to do that
> LOL.
>
> Wait until I build a Go playing program :-)
>
We are and have been waiting precisely for this. I don't think anyone
has wanted to be proved more wrong than the people you're arguing with
on this forum. You think that we will be mad or embarrassed if you make
a strong program. On the contrary, we would probably be thankful that
such a breakthrough had occured in the game of go. One must understand
though that we are realists and you saying that your revolutionary
method will be 10 stones stronger than anything on the market right now
sounds to us like saying your program will correctly guess next week's
lottery numbers.
Cheers,
Mef
Denis Feldmann
>
[cut]
>> and no matter how good you are at programming, if you don't understand the
>> game, your program will never beat me,
>
> Keep dreaming. Threatening to your ego, isn't it?
> A 5 USD pocket calculator can now beat chess grandmasters
As usual, you are badly misinformed. Or, in your so nice terminology,
lying. And your belief that Go is "like" chess, in this simplistic
analysis ("finite-state") is also grossly wrong, as there are, say
,10^120 games to consider in chess, against 10^700 in Go. If all this is
so easy, why not try for something much more useful, like good
language-translators, or good "mathematical machines" (and become rich
through the Clay prizes)? After all, mathematics is a huge game, but
also a finite-state one...
[cut]
>
> I know exactly what thickness is and it's near-trivial to program it.
Hubris. You know that word? And the Greek proverb which goes with it?
>
Frank de Groot
> at the very least, you will be behind a text edditor coding, and wont
> be able to bother us ;)
It's the opposite.
I can't work because YOU are bothering me.
How do you expect me to be able to concentrate on my work when you post "I
have aquired a pirate copy and will give it to anyone who asks"?
Of course I won't be able to prove I'm right when you keep harassing me.
Michael Alford
> Go is a state machine.
Interesting way to approach it, good luck.
Frank, I have no ill will toward you or your product, only some
curiosity as to the outcome. I am past 60 years of age, if MoyoGo can
beat me playing before I die, I will apologize to you here on rgg for
saying it couldn't be done :)
Cheers,
Michael
Bantari
> Only from the human POV.
> The same has been said of chess until someone put a ferw utterly simple
> heuristinc in a hyper-fast tree searcher with clever pruning and voila,
> Kasparov had a hard time defeating the thing. Go ain't any different, only
> different way of approaching the problem.
True.
But looking at the Kasparov-DeepBlue games (or even at the recent
Kramnik loss to Deep Fritz), while computers are now stronger than the
best chess players, I still don't have an impression that they are
"much" stronger.
For example, I don't think Deep Fritz could give Kramnik a handicap
(like a pawn or two first moves, or whatever). And chess has a much
smaller game tree than Go. And it has been worked on for more years by
more people. And, and, and...
Having said that, I agree that eventually, Go will be cracked as well
(although I hope it never will.) We can wait and see when this will
happen and who will create the breakthrough algorithms.
I hope it is you.
Best of luck.
________________
-Bantari
http://www.bantari.net/
Frank de Groot
> Chess benefits from the idea of positional chess, where you can quickly
> evaluate a position independentally from future outcomes. If you can
> find a way to effectively do this with go that will be an amazing feat,
I know *exactly* how to do this.
I even learned some maths to write down the proof.
> since as it stands it's hard to even get quantifiable reasoning out of
> strong players as for why one position feels better than a similar
> positon....
Exactly.
This is why their opinions/advice is worth exactly jack shit.
Because they don't know anything about this topic (Go engine design)
They are "laypeople". Uninformed. (clueless).
> True, however, arriving at the position evaluation as a state function
> has been difficult in the past.
So?
Is that supposed to be an argument?
>> Learning to play is a severe handicap and I will never do that.
> I'm not sure agree with this. One would not expect to solve a
> difficult math problem without learning math,
Wrong analogy.
I am solving the flying problem without learning to fly myself.
> In trying to solve go, one would think it important to understand how
> go works.
I do know how Go works.
I know the rules. That's enough.
I do not need to learn the patterns myself and do tree-search myself.
We have computers to do that nowadays.
> Wouldn't these be one and the same? Wouldn't knowing the concepts and
> how they relate equate to being able to play?
No. They are unrelated.
Frankly I'm getting tired of this discussion.
You don't get it, that's clear.
> We are and have been waiting precisely for this. I don't think anyone
> has wanted to be proved more wrong than the people you're arguing with
> on this forum. You think that we will be mad or embarrassed if you make
> a strong program. On the contrary, we would probably be thankful that
> such a breakthrough had occured in the game of go.
I personally could not care less, except that it might make me financially
independent.
I deliberately waited trying until 2007 because I needed that much time to
design the Go engine.
> One must understand
> though that we are realists and you saying that your revolutionary
> method will be 10 stones stronger than anything on the market right now
> sounds to us like saying your program will correctly guess next week's
> lottery numbers.
Whatever. Just stop harassing me and wait and see.