at the bottom of my Go page http://tromp.github.io/go.html, which also
contains an sgf link.
Direct link to image: http://tromp.github.io/img/WO5lives.png
Enlarging the board to 29x29 allows for a much better final (I hope)
look, close to my first attempt.
-John
___
> The hunt for the simplest possible ko gadget continues...
Latest attempt at the usual place:
>>> at the bottom of my Go page http://tromp.github.io/go.html, which also
>>> contains an sgf link.
>>> Direct link to image: http://tromp.github.io/img/WO5lives.png
Unfortunately not as pretty as the
> Hopefully fixed now.
Nope. Still no good. White can play O13, M11, or Q11 instead of recapturing ko.
The hunt for the simplest possible ko gadget continues...
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/list
> I assume after white plays at U22
This fails the goal of saving White O5, as Black will just P5, and
White's only means of escape,
with the ladder M3 N4 N5 M5 M6, fails when Black directs it to N17.
___
Computer-go mailing list
Computer-go@computer-go.
>>> Direct link to image: http://tromp.github.io/img/WO5lives.png
Might be useful for go event organizers in need of arrow signs...
regards,
-John
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-
>> I have attempted to reduce this y || (x && z) problem to the minimum
>> number of stones
>> at the bottom of my Go page http://tromp.github.io/go.html, which also
>> contains an sgf link.
>> Direct link to image: http://tromp.github.io/img/WO5lives.png
>
> Unfortunately, my ko gadgets don't work
> I have attempted to reduce this y || (x && z) problem to the minimum
> number of stones
> at the bottom of my Go page http://tromp.github.io/go.html, which also
> contains an sgf link.
> Direct link to image: http://tromp.github.io/img/WO5lives.png
Unfortunately, my ko gadgets don't work properl
> If we call the three kos x,y,z from top to bottom, then a succesfull
> White ladder amounts to
> (x || y) && (y || z). Which is equivalent to y || (x && z).
> So with y currently false, and White unable to flip it, White should
> take the bottom ko to make z true.
> Black can the make x false, bu
P hard instances. However, without a proof this
> assumption is still as valid as (1).
>
> I am curious what's John Tromp opinion on the above.
I spent some time thinking about the loss-less-ladder problem, that
asks if Black can capture
a given white group in a ladder without lo
On Tue, Jun 19, 2018 at 12:03 PM, Marcel Crasmaru wrote:
>> White can start one ladder as a ko threat to take back the middle ko, and
>> black will then take the top ko.
> I claim that White cannot use the ladders as a ko thread because:
> - if W plays R4 as a ko threat then B responds with S4
On Tue, Jun 19, 2018 at 3:52 AM, Marcel Crasmaru wrote:
> I've eventually managed to create a problem that should show a full
> reduction from a Robson problem to Go - I hope is correct.
>
> The Problem:
> https://drive.google.com/file/d/1tmClDIs-baXUqRC7fQ2iKzMRXoQuGmz2/view?usp=sharing
> Black
On Mon, Jun 18, 2018 at 10:24 PM, Álvaro Begué wrote:
> I don't think ko fights have anything to do with this. John Tromp told
> me that ladders are PSPACE complete: https://tromp.github.io/lad.ps
Ko fights are needed to take Go problems beyond PSPACE.
For Japanese rules they su
On Mon, Jun 18, 2018 at 10:30 PM, Marcel Crasmaru wrote:
>> FWIW, first-capture go (i.e. winner is first one to make a capture) should
>> not be PSPACE-complete.
>
> Actually this is not obvious.
>
> If you are able to replace the White Choice gadget shown at page V in
> this paper: https://tromp
dear David,
> To quote from: http://tromp.github.io/go/legal.html
>
> It should come as no surprise that L19, viewed as a position, is itself
> illegal.
>
> In this absolute form this statement got disproved in my German Go Forum
> article at
> http://www.dgob.de/yabbse/index.php?topic=5935.msg216
> You can also start with 9x9 go. That way games are shorter, and you probably
> don't need 1600 network evaluations per move to do well.
Bonus points if you can have it play on goquest where many
of us can enjoy watching its progress, or even challenge it...
regards,
-John
__
> Shouldnt that number at most be 722^#positions? Since adding a black or a
> white stone is something fundamentally different?
The upper bound of 361^L(19,19) games is from Theorem 7 on page 31 of
http://tromp.github.io/go/gostate.pdf, where you will find a proof.
As the paragraph preceding that
> And what is the connection between the number of "positions" and the number
> of games
The number of games is at most 361^#positions.
> or even solving games? In the game trees we do not care about
> positions, but about situations.
We care about lots of things, including intersections, stones
> Under which ruleset is the 3^(n*n) a trivial upper bound for the number of
> legal positions?
Under all rulesets.
> Unless we talk about simply the visual aspect
Yes, we do.
> but then this has
> absolutely nothing to do with the discussion abour solving games.
If you want the notion of "pos
> There is a definition of “brute force” on Wikipedia.
https://en.wikipedia.org/wiki/Brute-force_search explains it as
"systematically enumerating all possible candidates for the solution".
There is nothing systematic about the pseudo random variation
selection in MCTS; it may not even have suffi
hi Ingo,
>> “Pair Go” — A game where one Chinese pro will play
>> against another...except they will both have their own
>> AlphaGo teammate, alternating moves, to take the concept
>> of ‘learning together’ quite literally.
>
> Will the pro players in these games see the evaluations
> of AlphaGo?
>> (Japanese rules are not *that* hard. IIRC, Many Faces, and all other
>> programs, including my own, scored in them
>
> There is a huge difference between doing some variation of territory
> scoring and implementing Japanese rules. Understanding this difference
> will get you some way to understa
hi Bo,
> Let me know if there is any silly mistakes :)
You say "the perfect policy network can be
derived from the perfect value network (the best next move is the move
that maximises the value for the player, if the value function is
perfect), but not vice versa.", but a perfect policy for both
1. An intended play must be legal -- no playing on top of a stone hoping
it 'falls' to the neighbor positions.
> The point of the rule is ease of implementation for computer programs,
> to promote adoption. A program that already plays Go will probably keep
> tabs on legal plays, not eve
>> On frisbee Go itself I used the following definition:
>> 1. An intended play must be legal -- no playing on top of a stone hoping
>> it 'falls' to the neighbor positions.
>
> Accepted.
What's the point of this rule?
I feel it is an unnecessary restriction, similar to the no-suicide rule,
and w
dear Go researchers,
>> > Found a 582 move 3x3 game...
>> Can you give us sgf?
>
> I took the effort of trying to format the 582 game in a more insightful way.
> I ended up with lines of positions that mostly add stones, only starting
> a new line when a capture of more than 1 stone left at most 4
dear Aja,
> AlphaGo is getting stronger and stronger. I hope you all will enjoy watching
> the games.
Could you tell us if Alpha Go is able to come up with that most famous of moves:
http://senseis.xmp.net/?EarReddeningMove
Or is it so strong that it found an even better move:-?
regards,
-John
dear Ingo,
>>> ... (1 + delta)^(m*n).
>>
>> This is true, and a delta > 2 follows from a Theorem in an
>> upcoming paper by Matthieu Walraet and myself.
>
> Do you mean (1+delta) > 2, or really (1+delta) > 3?
Oops; I mean delta >= 1, so the base of the exponent is at least 2.
(1+delta) is necess
dear Nick,
> There's an assumption implicitly made here, which does not accord with my
> experience of frisbee Go: that the player will always aim at an
> intersection.
>
> Suppose I want to play on either of two adjacent points, and I don't care
> which. If I aim for one of them, I will land on o
dear Darren, Ingo,
> Again by random sampling?
Yes, nothing fancy.
> Are there certain moves(*) that bring games to an end earlier, or
> certain moves(*) that make games go on longer? Would weighting them
> appropriately in your random playouts help?
You could try to weigh moves by how likely t
I don't remember if there was consensus, but can repeat my previous thoughts:
> 1. What happens with plays unintentionally on top of stones or out of
> bounds?
Converted to involuntary pass.
Note that a throw must have some positive probability of converting into
a legal move. This way, infinitel
> The longest I've been able to find, by more or less random sampling,
> is only 521 moves,
Found a 582 move 3x3 game...
regards,
-John
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
dear Go researchers,
Finding the maximum length of a Go game, if we measure length
by number of (non-pass) moves, is equivalent to finding the longest
simple path in the game graph.
For 2x2 this can easily be brute forced, and one finds a maximum
length of 48 moves (so a single game can visit at
On Wed, Jan 27, 2016 at 1:46 PM, Aja Huang wrote:
> We are very excited to announce that our Go program, AlphaGo, has beaten a
> professional player for the first time. AlphaGo beat the European champion
> Fan Hui by 5 games to 0.
It's interesting to go back nearly a decade and read this 2007 art
For those of you who missed it, chess grandmaster Hikaru Nakamura,
rated 2787, recently played a match against the world's top chess program
Komodo, rated 3368. Each of the 4 games used a different kind of handicap:
Pawn and Move Odds
Pawn Odds
Exchange Odds
4-Move Odds
As you can see, handicaps
> You must be kidding about Lee Sedol.
> ...
> So he was by far the biggest fish Google could ever catch for that
> game, for Go insiders as well as for people outside the Go scene.
Well said, Marc.
In terms of name recognition and domination in the past decade,
who else but Lee Sedol should be p
dear Robert,
>> It will never be known since there's not enough space in the known
>> universe to write it down. We're talking about a number with over
>> 10^100 digits.
>
> How do you know that an implicit expression (of length smaller than 10^80)
> of the number does not exist? :)
Of course an
dear Robert,
> The number G19 of legal games under a given go ruleset is unknown.
It will never be known since there's not enough space in the known
universe to write it down. We're talking about a number with over
10^100 digits.
> For positional
> superko (prohibition of recreation of the same
> A member of the German forum said, that a French Go player reported on
> Facebook, that Fan Hui lost 5 out of 5 games to the Google Go engine.
To ask the obvious:
Were these even or handicap games?
-John
___
Computer-go mailing list
Computer-go@compu
I foresee a future where we watch Google vs Facebook matches with
human professionals providing commentary on their superiors :-)
Interesting times we live in!
-John
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman
dear Mark,
> Well, although Dr. Tromp seems rather modest about this result, I haven't
> heard of anyone else doing similarly interesting work on the theoretical
> foundations of the game.
There is a lot of other interesting research beyond counting things.
Just to name a few there's rule explora
dear Erik,
> I was wondering if there is an efficient way to find the number of unique
> positions with symmetrical positions excluded.
It's roughly L19/16.
That's slightly short, but will be correct in the first 85 or so digits.
You just need to correct for the positions with rotational and/or
Wow, Robert, so many questions!
Many of which I have no idea how to answer:-(
> You must have needed 15 or 20 years of research to find the result?
Very intermittently though. If it were all continuous, it may be
several months of Go research, several more months of article editing,
and a few yea
> shows how these 57 positions form 13 equivalence classes with respect
> to mirroring/reflection which further reduces to 7 classes when
> considering color symmetry as well.
Correction: that should be 8 (not 7) classes for all symmetries.
-John
___
Co
dear Erik,
> Does the number include symmetrical positions (rotations / mirroring / color
> reversal)?
Yes, of course.
This is also apparent from the table at the bottom listing 57 legal
2x2 positions. Figure 4 on page 5 of our paper
http://tromp.github.io/go/gostate.pdf
shows how these 57 posi
It's been a long journey, and now it's finally complete!
http://tromp.github.io/go/legal.html
has all the juicy details...
regards,
-John
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
Our paper "Combinatorics of Go" has some results on this,
in a rule system allowing suicide.
See http://tromp.github.io/go/gostate.pdf,
in particular Section 7 on Hamiltonian games.
-John
On Thu, Nov 19, 2015 at 10:13 AM, Marc Landgraf wrote:
> Hi,
> there is a question that lately crossed my m
On Thu, Nov 12, 2015 at 9:07 AM, Nick Wedd wrote:
> I was thinking about the ko rule for frisbee ko, and realised it leads to
> problems.
>
> 1. Black takes a ko, White tries to make a ko threat, but accidentally
> retakes the ko. What should happen?
This was already covered by having any ill
>> Would the game end after two unintentional passes?
> Good point. In principle I would say so.
That makes little sense to me.
IMO, the principled rule is that two consecutive intentional passes
end the game.
To make sure that infinitely long games have 0 probability,
we must then require that
> By the way: It would also be necessary to decide about
> the eps for the event. Natural candidates would be
> eps=0.1 or eps=0.125.
I would say the 2 most interesting choices are 1/8 or 1/4.
The latter guarantees you miss your aim by distance 1,
while the former gives you an even chance to hit i
On Wed, Oct 7, 2015 at 7:44 PM, Petr Baudis wrote:
> On Wed, Oct 07, 2015 at 02:29:27PM +0200, Erik van der Werf wrote:
>> A measure that I find reasonable is a limit on number of threads x
>> clock frequency.
> I'm not sure this would work well. The #playouts difference between
> an old Bulld
> I have just been told by a colleague that Edouard Rodrigues solved hex
> mathematically. I was very surprised because I had never heard about it.
>
> The web site with the proof and optimal strategy is there:
> http://jeudhex.com/?page_id=17
Perhaps he found a winning strategy for an unrestrict
dear Robert,
> How much computation time do you expect to reveil the complete exact 19x19
> number? Or is more research necessary before I may ask this?
this computation of the 64 least significant bits was 1/9 of the total
effort needed. each such job contributes 64 bits to the answer.
regards,
See my updated webpage at
http://tromp.github.io/go/legal.html
regards,
-John
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
On Wed, Mar 11, 2015 at 9:21 AM, folkert wrote:
> Alvaro, Urban,
>
> thanks!
>
> I've got an additional question.
> It may be obvious but it is written a bit ambiguous imho on
> senseis.xmp.net:
>
> "A player's score is the number of points of her color, plus the number
> of empty points that reac
669723114288829212892740188841706543509937780640178732810318337696945624428547218105214326012774371397184848890970111836283470468812827907149926502347633
More details at http://tromp.github.io/go/legal.html,
including a call for volunteers to contribute computing power for
determining what we all
After some more tinkering, I put two new versions of Go Rules in Haskell
on my go page at
http://www.cwi.nl/~tromp/go.html
The simpler one is annotated with the 10 articles of the rules, while
the fancier one is parametrized by board topology (like templates in C++).
Yesterday, I discovered a
Claus Reinke wrote:
>As for me, i'm really NOT interested in knowing "what langage is good for go
>programming". That's simply not a question i can ask myself, nor anyone else.
> This question doesn't make any sense for me. Still if someone can get the
>"standard light playout" right in less than 1
On Jan 3, 2008 10:46 AM, Don Dailey <[EMAIL PROTECTED]> wrote:
> Yes, the KGS rules gives only 1 chance to agree. At one point KGS
> allowed this to happen repeatedly, but it cause some bots to infinite
> loop on the server when they disagreed. So I think it's better than
> nothing, but imper
On Dec 19, 2007 1:00 PM, Jeff Nowakowski <[EMAIL PROTECTED]> wrote:
> On Tue, 2007-12-18 at 15:04 -0500, John Tromp wrote:
> >
> > See the Haskell implementation of my connect-4 solver, Fhourstones, at
> > http://www.cwi.nl/~tromp/c4/fhour.html
>
> You say on th
On Dec 18, 2007 3:03 PM, Chris Fant <[EMAIL PROTECTED]> wrote:
> On Dec 18, 2007 2:21 PM, Harald Korneliussen <[EMAIL PROTECTED]> wrote:
> > I'd like to know how well MoGo would have played if you let it think
> > for a week for every move. Only it seems to me that is not possible,
> > because I do
> But I have to admit, I don't know exactly how I'd go about
> implementing a transposition table in Haskell :-/ Perhaps I'll try for
See the Haskell implementation of my connect-4 solver, Fhourstones, at
http://www.cwi.nl/~tromp/c4/fhour.html
regards,
-John
On Nov 16, 2007 10:05 AM, Chris Fant <[EMAIL PROTECTED]> wrote:
> > > Neat. Was the 15-bit version for 81 values or 361? At the risk of
> > > putting my foot in my mouth, I don't think there exist 361 15-bit
> > > numbers that satisfy minimum requirements (if the floating-point
> > > average of any
On Nov 14, 2007 9:02 PM, Eric Boesch <[EMAIL PROTECTED]> wrote:
> The "if average is in my original code_value set" seems like a
> bottleneck here, requiring about #bits (i.e. about 9, since 361 is a
> 9-bit number) operations no matter how you do it as far as I can tell
> (unless you use a stupid
On Nov 14, 2007 5:03 PM, Imran Hendley <[EMAIL PROTECTED]> wrote:
>
> On Nov 14, 2007 3:19 PM, John Tromp <[EMAIL PROTECTED]> wrote:
> > On Nov 14, 2007 2:00 PM, John Tromp <[EMAIL PROTECTED]> wrote:
> >
> > > My solution doesn't make use of that
On Nov 14, 2007 2:00 PM, John Tromp <[EMAIL PROTECTED]> wrote:
> My solution doesn't make use of that, and satisfies the stronger property:
> 0 <= a_i <= 4 and sum a_i * n_i is in 1*nums union 2*nums union 3*nums
> union 4*nums
> => only one a_i is nonzero.
that
On Nov 14, 2007 1:44 PM, Lavergne Thomas <[EMAIL PROTECTED]> wrote:
> Let elaborate a little more on this. We want one number for each cells :
> nums = {n1, n2, n3, ..., n81}
>
> And we want the following properties :
>
> for any a, b in nums :
> (a + b) / 2 is in nums --> a == b
> > Yes, you can generalize pseudoliberties by extending them
> > with another field, such that if the (summed) pseudoliberty field
> > is between 1 and 4, then the other (summed) field will tell you if all
> these
> > are coming from a single true liberty.
>
> Can you elaborate on this?
Let me po
On Nov 13, 2007 2:48 PM, Petr Baudis <[EMAIL PROTECTED]> wrote:
> I'm now somewhat torn. The speedup from using pseudo-liberty counts
> could be huge, estimating from my profiling. On the other hand, it would
> be very useful to still be able to quickly check if a group is in atari
> - it looks li
On Nov 13, 2007 2:15 PM, Don Dailey <[EMAIL PROTECTED]> wrote:
> How does the speed of the Haskell version compare to the C and Java
> version? The Haskell web site now brags about how fast Haskell is.
Not too well:-(
Fhourstones in Haskell runs more than 10 times slower than the C version...
On Nov 13, 2007 11:10 AM, William Harold Newman
<[EMAIL PROTECTED]> wrote:
> On Mon, Nov 12, 2007 at 04:41:35PM -0500, Chris Fant wrote:
> > I would like some language recommendations. Requirements:
> Among the languages I know something about (which excludes D and C#,
> for example)...
> techni
On 11/12/07, Don Dailey <[EMAIL PROTECTED]> wrote:
> Ok, on 2x2 I get a consistent result now that I implemented PSK. It
> gives the same result with SSK too. It's a 1 point win for the first
> player. I'm not sure this is in agreement with other peoples
> findings. But it appears to be c
hi Nick,
On 11/12/07, Nick Wedd <[EMAIL PROTECTED]> wrote:
> The results of yesterday's KGS bot tournament are now available at
> http://www.weddslist.com/kgs/past/32/index.html.
>
> Your comments and corrections will be appreciated as usual.
You wrote:
"In the round 8 game between FirstGoBot and
> I just ran my perm application for 4x4 and it's reporting
> 43,046,721 unique board states and took 2m6.980s. Will try for 5 and 6.
seems you're computing 3**(n*n)
3**16 = 43046721
3**25 = 847288609443
3**36 = 150094635296999121
don't you want to exclude illegal positions?
-john
On 10/29/07, Christoph Birk <[EMAIL PROTECTED]> wrote:
> On Mon, 29 Oct 2007, Jacques Basaldúa wrote:
> > This can also be done by the programmers. E.g. If CrazyStone is too strong,
> > Rèmi can introduce a CrazyStoneH3 which passes 3 times
> > at the beginning. But not at the first move, to avoid
On 8/16/07, Darren Cook <[EMAIL PROTECTED]> wrote:
> Apologies for the off-topic post, but I know lots of people here are
> interested in statistics and algorithms.
>
> Calculating the mean of a stream of numbers [1] is easy: just keep track
> of the sum and the count, and divide at the end. But wh
hi Sylvain,
> Figure 3 in your UCT paper shows the accuracy of different simulation
policies.
> Could you repeat these experiments for accuracy of win/loss determination
only?
Actually the labelled positions are rather end game positions, and are
labelled as 0/1 (loss/win). So we already are i
hi Sylvain & David,
Figure 3 in your UCT paper shows the accuracy of different simulation policies.
Could you repeat these experiments for accuracy of win/loss determination only?
So for each test position, you determine if it's won or lost under perfect play,
and then see how close each policy g
On 5/25/07, Don Dailey <[EMAIL PROTECTED]> wrote:
Is there some kind of bet on this?When did that happen? What is
the bet exactly?
Somewhere around 2000, I claimed I would not be beaten by a computer
under match conditions (eg. 10 games at 1hr per side + byo-yomi)
within 10 years. Which D
On 5/24/07, Darren Cook <[EMAIL PROTECTED]> wrote:
P.S. John, it says the new algorithm can topple strong players - shall
we just believe them and say I won that bet? We don't really need to
play the games out to prove it do we ;-).
On 9x9 they definitely can. I've lost a few games myself to th
Question for native English speakers: do you think this technique is best
described by "progressive unpruning" or "progressive widening"?
I'm no native speaker, but I think using the word "selectivity" may be
most descriptive.
Does "regressive selectivity" sound too weird ?
regards,
-John
_
On 5/19/07, Thomas Wolf <[EMAIL PROTECTED]> wrote:
Here is another Amsterdam paper on Go, although about life & death
and not full game playing.
I may be missing the obvious, but in Section 4.2, Diagram 13,
isn't Black 10 a basic ko violation?
regards,
-John
___
On 5/18/07, Rémi Coulom <[EMAIL PROTECTED]> wrote:
My idea was very similar to what you describe. The program built a
collection of rules of the kind "if condition then move". Condition
could be anything from a "tree-search rule" of the kind "in this
particular position play x", or general rule
On 5/18/07, Peter Drake <[EMAIL PROTECTED]> wrote:
It took me a long time to get around my mental block and accept the advice
of everyone here, but your intuition is correct: superko is so rare, and so
expensive to detect, that you should NOT check for it on every move.
In dimwit, we check for
The real name of your bot, if different
dimwit
* The names of the authors of all parts of your Go-playing
program. I will expect you to have the consent of all these people for
your program to be entered in this tournament.
Alvaro Begue and John Tromp
* whether you want to enter the
On 4/8/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
The factorial of 81 is about 10^140. The number of legal positions may be
it may be 103919148791293834318983090438798793469
regards,
-John
___
computer-go mailing list
computer-go@computer-go.or
On 4/5/07, Chaslot G (MICC) <[EMAIL PROTECTED]> wrote:
The workshop will be held on Friday 15. - Sunday 16. June 2007.
Must be a leap Saturday...
regards,
-John
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mail
dear Nick,
>> I am considering postponing the MAY KGS bot tournament from Sunday
>> MAY 6th (in the UK, this is the "Spring Bank Holiday") to Sunday MAY
>> 13th.Will this inconvenience anyone?
It might be a nice thing to watch on my birthday:-)
I might even participate...
regards,
-John
As far as I know, pseudo-liberties are only used for detecting a
capture or detecting atari. If this method you suggest has some value
beyond that, then I'm interested to learn more about it. But the
I have a nice mathematical puzzle for you.
Fix some k, say, 81.
What is the smallest range N
On 3/29/07, Christoph Birk <[EMAIL PROTECTED]> wrote:
On Thu, 29 Mar 2007, Jim O'Flaherty, Jr. wrote:
> What's a pseudo-liberty?
> And how can there be more of them than there are empty intersections
> (81) on the board?
It is the sum of all stone's liberties in a group; ignoring common
liberti
On 3/29/07, Weston Markham <[EMAIL PROTECTED]> wrote:
On 3/29/07, John Tromp <[EMAIL PROTECTED]> wrote:
> On 3/29/07, Weston Markham <[EMAIL PROTECTED]> wrote:
> > It appears to me that at least 91 is possible:
> Nice! If you use O's instead like
>
> .OO.
On 3/29/07, Weston Markham <[EMAIL PROTECTED]> wrote:
It appears to me that at least 91 is possible:
.xx.x.xx.
xx.xxx.xx
.xx.x.xx.
xx.xxx.xx
.xx.x.xx.
xx.xxx.xx
.xx.x.xx.
xx.xxx.xx
.xxx.xxx.
Nice! If you use O's instead like
.OO.O.OO.
OO.OOO.OO
.OO.O.OO.
OO.OOO.OO
.OO.O.OO.
OO.OOO.OO
.OO.O.OO
On 3/29/07, John Tromp <[EMAIL PROTECTED]> wrote:
Is 88 the maximum number of pseuoliberties a string can have on 9x9?
Make that 89:-)
-John
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/li
Out of curiosity,
Is 88 the maximum number of pseuoliberties a string can have on 9x9?
(it should be safe to use only 6 bits in practice, if you need every last bit:)
regards,
-John
___
computer-go mailing list
computer-go@computer-go.org
http://www.c
On 3/19/07, Don Dailey <[EMAIL PROTECTED]> wrote:
I'm testing a future Anchor player for CGOS. I am calling
it FAT for Future Anchor Test!
It plays fixed depth and I pre-calculated what level to make
it play at 1800 strength. I came pretty close, Fat-25 is
playing at 1836 at the moment and do
dear Don,
Crazy me. I just remembered why my numbers are not matching. I forgot
that what I call the lite play-out version is not random. It's mostly
lite but it favors capture moves.
Yes, I can see how that will shorten the games somewhat...
Is it easy to temporarily turn off that bias?
-
hi Don,
> Are you trying to make a Monte Carlo program?
Guilty:-)
Since about a week and a half, me and my colleague Alvaro Begue are
working on a Go program, which (like many others) wil try to imitate
Mogo's success...
regards,
-John
___
computer-
I've seen the number 107.3... reported earlier
for the average length, without the 2 final passes.
Is this allowing multi stone suicides or not?
And what's the outcome in the other case?
Thanks!
regards,
-John
___
computer-go mailing list
computer-go@c
Many people here have reported on their MC playout speed,
with numbers approaching a most impressive 100K playouts/sec.
I would like to know what impact adding UCT has on this speed.
In the playout you need only spend a small constant amount of
work per move, but choosing a single child node in UC
On 1/5/07, Darren Cook <[EMAIL PROTECTED]> wrote:
The playstation multiprocessing looks very different: you get 1 general
purpose CPU and 6 specialized CPUs. Their key feature is they have 256K
of local memory - this is not cache, it is all the memory they can
access. Not useful for UCT designs
Those of you looking to wring more performance out of your
MonteCarlo Go programs might be interested in this article about
installing Linux on the Sony PlayStation 3 and programming the
6 available SPE coprocessors on its Cell cpu:
http://www-128.ibm.com/developerworks/power/library/pa-linuxps3-1
1 - 100 of 118 matches
Mail list logo
