2010/4/8 Andrés Domínguez :
> Migos only solved up to 5x5 AFAIK.
Actually, for empty boards its up to 5x6.
It also analyzed several positions on larger boards (from solutions
claimed by humans and some published problems). So far Migos has not
been able to disprove the main lines of the human sol
2010/4/7 Erik van der Werf :
> 2010/4/7 Andrés Domínguez :
>> I think the big komi is an anomaly with very small boards, where white
>> player can't even live.:
>
> That's one thing, the other is that on even size boards the lack of a
> center makes it hard to defeat mirror Go.
So I said later abo
2010/4/7 John Tromp :
>>> 2x2 : Draw, both with super ko rule or simple ko.
>
> B+1 under area rules and superko.
I can't see it, I think is B+0 with territory and area
scoring, also with superko.
>>> 4x4 : Dificult, but I think is a draw with japanese rules
>
> B+2 under area rules and superko.
2010/4/7 Andrés Domínguez :
> I think the big komi is an anomaly with very small boards, where white
> player can't even live.:
That's one thing, the other is that on even size boards the lack of a
center makes it hard to defeat mirror Go.
> 1x1 : Draw, because no legal move (or losing points al
>> 2x2 : Draw, both with super ko rule or simple ko.
B+1 under area rules and superko.
>> 4x4 : Dificult, but I think is a draw with japanese rules
B+2 under area rules and superko.
>> 6x6 : Probably white can live and black wins but less than 24 points
B+4 under area rules and superko.
>> Bi
what may likely go to zero is (komi / number of points on board),
since absolute first move advantage should go down.
right?
s.
On Wed, Apr 7, 2010 at 3:35 PM, Don Dailey wrote:
>
>
> 2010/4/7 Andrés Domínguez
>>
>> 2010/4/7 Erik van der Werf :
>> > On Wed, Apr 7, 2010 at 3:43 PM, Brian Sheppa
2010/4/7 Andrés Domínguez
> 2010/4/7 Erik van der Werf :
> > On Wed, Apr 7, 2010 at 3:43 PM, Brian Sheppard
> wrote:
> >>>decreases with board size
> >>
> >> Since the game-theoretic value is a small positive integer, I don't
> think it
> >> can decrease with
> >>
> >> increasing board size.
> >
2010/4/7 Erik van der Werf :
> On Wed, Apr 7, 2010 at 3:43 PM, Brian Sheppard wrote:
>>>decreases with board size
>>
>> Since the game-theoretic value is a small positive integer, I don't think it
>> can decrease with
>>
>> increasing board size.
>
> What I meant is that it tends to start with ful
I suspect that it's not zero or one - just a hunch.My theory is that it
asymptotically approaches some small value as the boards get larger and that
value is probably attained pretty quickly, perhaps much less than 100x100
board size. Maybe even close to what we play on now.
It seems to me th
On Wed, Apr 7, 2010 at 3:43 PM, Brian Sheppard wrote:
>>decreases with board size
>
>
>
> Since the game-theoretic value is a small positive integer, I don't think it
> can decrease with
>
> increasing board size.
What I meant is that it tends to start with full-board wins on tiny
boards (up to
Note that with small boards, the komi is really high, such as 25 for 5x5.
With larger board it seems to get smaller.So it seems likely to me that
it should get smaller with increasing board size.Of course this is not
any kind of proof - it's just a trend that we can observe at only the
smal
it could flatten out at 1 or 0 (is there a reason why it cannot be
zero?). further, it could bounce around near two small values
depending upon the parity of the boardsize or other arithmetic
properties of the boardsize.
s.
On Wed, Apr 7, 2010 at 9:43 AM, Brian Sheppard wrote:
>>decreases with
it might not be too hard to find out in practice -- just let players
bid for komi on boardsizes that have been solved. it might take them
a few thousand games to get used to the teeny-tiny board sizes, but...
;)
s.
On Wed, Apr 7, 2010 at 8:55 AM, Don Dailey wrote:
>>
>>
>> Perfect komi is most
>decreases with board size
Since the game-theoretic value is a small positive integer, I don't think it
can decrease with
increasing board size.
___
Computer-go mailing list
Computer-go@dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/c
On Apr 7, 2010, at 1:52 AM, Erik van der Werf wrote:
> I suspect that if one would do a study to find the optimal statistical
> komi (for balancing the probability of winning with equal strength
> players) we would find that it decreases with board size and increases
> with playing strength.
It'
>
>
>
> Perfect komi is most likely to be 5, 7 or 9 points but nobody has a
> proof for the true value of the empty 19x19 board (but even if it was
> known it might not be the right value for balancing fallible players).
>
Indeed this seems likely to me. In other games where the believed game
the
On Wed, Apr 7, 2010 at 1:04 PM, Chase Albert wrote:
> Nitpick: it doesn't necessarily give the players an equal chance of winning,
> it just makes perfect players tie. Equal probability of winning is likely
> not achievable in go (though it would be a poor game if it weren't at least
> close).
Ni
--
From: "Nick Wedd"
Sent: Wednesday, April 07, 2010 6:11 PM
To:
Subject: Re: [Computer-go] Komi and the value of the first move
In message <85bf1f3d394747e0b5b6383a2f34b...@homepc1d6062ef>, Aja
writes
I couldn't find the article again where I
Nitpick: it doesn't necessarily give the players an equal chance of winning,
it just makes perfect players tie. Equal probability of winning is likely
not achievable in go (though it would be a poor game if it weren't at least
close).
On Wed, Apr 7, 2010 at 06:11, Nick Wedd wrote:
> In message <
In message <85bf1f3d394747e0b5b6383a2f34b...@homepc1d6062ef>, Aja
writes
I couldn't find the article again where I read that 20 - 30 points
was an accepted value for the first move.
If you mean 19x19 and the first move is played in the corner, then its
value is around 10 points.
I think it
> But if the first move is played in the center or the edge, then its
> value is lower than 10 points. So, actually the value of the first
> move depends on the location it is played.
To get this out of the way, we're not talking about the value of any old
starting move.
We're talking about the bi
On Wed, Apr 07, 2010 at 11:13:24AM +0800, Aja wrote:
> >I couldn't find the article again where I read that 20 - 30 points
> >was an accepted value for the first move.
>
> If you mean 19x19 and the first move is played in the corner, then
> its value is around 10 points. That's why we generally re
hi all,
Is there some standard “case benchmark” for testing the playing level of Go
playing programs? for example, some boards with known good moves that could
test program's playing level? As I know, there exist some good boards to
test life and death problems, however, is there any boards on oth
23 matches
Mail list logo
