rrect reply was discovered, the move has
considerable "inertia" and it takes time to recover.
Has anyone tried only counting the most recent games through a move
(say, the last 100), rather than the entire history?
Peter Drake
Assistant Professor of Computer Science
Lewis &
On Oct 30, 2006, at 7:51 PM, Don Dailey wrote:
On Mon, 2006-10-30 at 19:34 -0800, Peter Drake wrote:
I'm running into a problem where my Monte Carlo program is very slow
to acknowledge that its favorite move has a strong counter.
Part of the problem is that the value of a move is based o
To those of you with Monte Carlo programs:Consider this board configuration:...ww..wBBw..wBBw..wB I believe black's best move is to play in the center and escape via the broken ladder.1) Is this true? (This is a Go-playing question.)2) Can
ogn(p) / r)
Is this correct? Are there any things to watch out for?
Also, are people pruning any move for which (average + uncertainty)
is less than the (average - uncertainty) of some other move?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://ww
available for something else.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
doing that.
Since I use transpositions, it is not really possible to remove a
subtree, anyways.
Yes, another thing I have to (re-)implement...
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
Consider this position:...wBww.wB...wBBBBw..wBw...w..wBw.w.w.wwBw...This is black to play, no komi.As I read it (correct me if I'm wrong!), the white groups at upper left and lower right can't be killed (assuming white defends them). Black can kill either of the
On Nov 7, 2006, at 11:10 AM, Magnus Persson wrote:
Quoting Peter Drake <[EMAIL PROTECTED]>:
The probability of winning by starting at b2 is greater than the
probability starting elsewhere, but shouldn't it approach 1.0,
since b2 is a winning move? Do others get this sa
mmand like "name" from a
front end is different from doing it manually?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
ht
n and return you to your regularly
scheduled programming.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Nov 14, 2006, at 9:19 AM, Peter Drake wrote:
Orego supports GTP, but for some reason today I can't get it to
work wi
Orego speaks GTP, as does gnugo. I'd like to run a bunch of games
(say, 50) between them to see how many Orego wins. Does anyone have a
handy script (ideally bash or Python) for this?
Thanks,
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.l
One could use the heuristic that the outside is larger than the
inside, but perhaps a situation can be created where this is not
true. This might even be something beginning Go players could see/
understand.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
If anyone's interested in digging through our C++ code, here it is:
http://www.lclark.edu/~drake/go/
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
computer-go mailing list
co
Here it is:
https://webdisk.lclark.edu/xythoswfs/webui/_xy-2115826_1-t_OX34gnaB
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
t the difference between "server mode" and
the default "client mode" is -- maybe just-in-time compilation? In
any case, I found that it nearly doubled the speed of a small program.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
just an improvement but a very large
improvement -- 95% statistical confidence.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Nov 29, 2006, at 3:22 AM, Chrilly wrote:
I am confused. In your paper you write "Orego is
This is something we hope to do once we have Orego multithreaded:
give each version the same amount of time, so the time costs of
adding a heuristic are automatically taken into account.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~d
o use a big array of raw ints. Yes, this will be just as ugly
as doing it in C/C++, but at least I'll be able to confine the
ugliness to that part of the program.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
__
A note: we're working on converting Orego back from C++ to Java, and
we're getting 5,000 (totally random at this point) simulated games
per second. We'll probably continue in this direction.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
ht
moves have been tried from a given move, UCT takes
over.)
In any case, my profiling data indicates that choosing the random
move per se is not the expensive part; playing the move is.
Thanks for the suggestion,
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark Col
a "new" one. Similarly,
instead of
Foo x = y.clone();
do something like
x.copyDataFrom(y);
where of course you have to write copyDataFrom().
3) Algorithmic improvements are always more important that fine-tuning.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clar
d that it can be overwritten, we
must leave it alone lest it mess up another part of the tree. At this
point, the parent's forced leaf turn number is set to the child's
turn number; all moves from that parent will be random until a later
point in the game when it is safe to overwrite
Only 28 bytes? What's in your nodes? That doesn't seem like enough
room for 82 pointers to children. If you don't have pointers to
children, how do you find the children's statistics for UCT? Play
each move and examine the resulting hash code?
Peter Drake
Assistant P
9x9, 2 GHz Intel Core Duo iMac.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 6, 2006, at 2:59 PM, Don Dailey wrote:
On Sun, 2006-12-03 at 19:53 -0800, Peter Drake wrote:
A note: we're working on converting Orego
I'm getting about 98.2 for purely random games, but Orego has a turn
limit that prevents a game from taking more than 162 moves on 9x9.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 6, 2006, at 3:48 PM, Don Dai
on a 2 GHz machine, even in C++.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
eaded my program and ... it's
slower than the single-thread version! I'm pretty sure that this
happened because I threaded it at too fine a scale; I'm incurring the
overhead of thread creation for each MC run. I'll have to do some
thinking about how to reorganize this.
Pet
Those of you with multithreaded UCT programs -- how do you do it?
Doesn't UCT pretty much require updating a common data structure
after each MC run?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at 9:17
eady doing this and I'm curious how.
Remi? Sylvain?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at 9:41 AM, steve uurtamo wrote:
Those of you with multithreaded UCT programs -- how
do you do it?
Doesn
ot of mutexes (mutices?).)
It would have to be set up so that any number of threads can read
from the tree at once, but only one can write to it, and nobody can
read while someone is writing.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.e
Aha! Now I get it. You only have to look at the tree during the
opening part of the run. Once you've fallen off and are making purely
random moves, you can let someone else use the tree.
Thanks!
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.l
Got it -- now I'm getting just under 10,000 games per second! Whee!
(FWIW, I actually don't have the UCT part in there yet -- these are
purely random games.)
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7,
time and space. For example, that latter
suggests that the first child / next sibling tree representation
mentioned here recently is a much better idea than the array-of-
children representation I'm currently using.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark Col
Precisely: I'm getting almost optimal use of my dual-core processor.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at 10:47 AM, Don Dailey wrote:
On Thu, 2006-12-07 at 10:24 -0800, Peter Drake wrote:
Got
On Dec 7, 2006, at 11:08 AM, Don Dailey wrote:
On Thu, 2006-12-07 at 09:17 -0800, Peter Drake wrote:
I do have the undo ability, but I think it's done in (I think) a
very
efficient way. For example, when I want to undo a bunch of move at
once (e.g., after a MC run) I just reduce a
ve is associated with which child node. If a
node might have more than one parent, the node can't store its last
move.
Any clever solutions?
If not, any opinions (or better yet, evidence) as to whether the
space savings or the DAG transposition table is more valuable?
Peter Drake
Assistant Pr
) In performing UCT, I need to traverse my children, find the value
and confidence bound for each one, and then choose the move leading
to the "best" one. This requires knowing which move leads to which
child node.
2) In testing, I like to be able to print out the tree.
Peter Drake
I see. I only recently had this realization that the within-tree and
purely-random parts of the search were deeply different. I'll take a
shot at this.
Thanks,
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006
ose of you keeping track, that's about 12,000 pure MC runs per
second with the dual-core processor.
I seem to have reached the point of diminishing returns here.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at
ght really require several nodes, this trick won't
work. Oh, well. I guess I was trying to kill too many birds with the
same stone.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at 3:46 PM, Anders Kierulf
is hugely faster and uses an order of magnitude less memory
than it did 24 hours ago. Thanks, everyone!
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 7, 2006, at 10:07 PM, Anders Kierulf wrote:
I was also hoping to use m
this and took a big
performance hit because the games run longer now. I'm down to 3400
games / sec with one thread, roughly 6000 with two. I think it's
worth it, though, because the random games will be more accurate.
Peter Drake
Assistant Professor of Computer Science
Lewis &
s negative."
Yikes!
So, I'll have to check the sign of my result and try again in those
rare (but not ignorably rare) instances where it is negative.
I hope this saves some time for the next person to encounter this
problem.
Peter Drake
Assistant Professor of Computer Scie
Ah. Orego will have the "ponder" feature soon.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 20, 2006, at 3:11 PM, Don Dailey wrote:
ponder means to "use the opponents time to think"
- Don
On Wed, 2
ondon time the same as GMT/UCT
only part of the year?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
On Dec 23, 2006, at 10:58 AM, Nick Wedd wrote:
I have written up the week's Slow KGS bot tournament. My report,
less for the opening moves
than for middle / endgame moves? Is there a smooth curve, or is there
a relatively abrupt shift from joseki to analysis?
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~d
ot;?
(Lest Orego's honor be besmirched, I should clarify that I was only
allowing Orego one second per move while testing out the protocol.
Hopefully it won't get wiped off the board by GNU Go level 1 if I
give Orego more time.)
Peter Drake
Assistant Professor of Computer Science
Le
maybe 6x6.
There are some games, such as Hex, for which we know who wins from
the starting position given optimal play, but we don't know how to
figure out the best move.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
White passes
White plays D8
Black plays c8
White plays D9
Black plays d5
Later, the script sent my program "genmove black" twice in a row,
which of course confused it.
What's going on here? Has anyone else run into these problems?
Peter Drake
Assistant Professor of Computer S
bring their own machines. (Since there will be prize
money involved, we can't reasonably allow remote connections.)
I'm also trying to get some funding from industry, but I'm not
holding my breath on that one.
Peter Drake
http://www.lclark.edu/~drake/
_
Can you expand on that? I'm not familiar with VMWare.
I suspect many programmers here would be concerned about the
performance hit of working on a virtual machine.
Peter Drake
http://www.lclark.edu/~drake/
On Mar 18, 2008, at 7:46 PM, Michael Williams wrote:
If they install V
Thanks to everyone for all the comments.
Another question: Should the tournament be 9x9, 19x19, or both?
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo
should adjust our programs just yet, but we may be
getting closer to an international standard. Of course, it's
irrelevant until Japan, China, and Korea get on board.
You can find the rules in question here:
http://www.cs.cmu.edu/~wjh/go/rules/AGA.html
e code for this. If you plan to work in Java,
the Orego code is intended to be relatively easy to read, so you
might start there.
BTW, Orego has moved to:
http://www.lclark.edu/~drake/Orego.html
Peter Drake
http://www.lclark.edu/~drake/
___
c
GO),
although I've gone to somewhat greater length to explain them. If you
find C++ easier to read than Java, by all means use Lew's code.
Orego is here (you'll have to download and unpack the latest .jar):
http://www.lclark.edu/~drake/Orego.html
Peter Drake
http://www.lclark.ed
r you need to check for single-stone suicide or simple
ko violation. See playNonPassUnoccupied().
4) Tell whether a point is surrounded when counting the score at the
end of a playout. See playoutScore().
Lukasz could say more -- he spent a lot of time optimizing the heck
out of this stuf
d like the tournament to run as
smoothly as possible.
Peter Drake
http://www.lclark.edu/~drake/
Tournament Director
Peter Drake (AND CO-DIRECTOR?)
Description
This 19x19 tournament is for computer programs only. While there have
been notable breakthroughs in recent years, computer Go remains
programmers an
opportunity to network in a way that is not possible over email.
3) To create an event on which the media can report.
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/m
he current schedule has dinner from 5:30-7:00.
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
We'd like to get an estimate of numbers. Who's planning to enter the
US Go Congress Computer Go Tournament?
Here is the latest version of the info on the tournament:
Computer Go Tournament
Tournament Director
Peter Drake (AND CO-DIRECTOR?)
Description
This 19x19 tournament is fo
(please explain)
Thanks,
Peter Drake
http://www.lclark.edu/~drake/
On Jun 9, 2008, at 10:25 AM, Peter Drake wrote:
We'd like to get an estimate of numbers. Who's planning to enter
the US Go Congress Computer Go Tournament?
Here is the latest version of the info on the
In Sunday's tournament, Orego lost a game embarrassingly by playing
out a lost ladder. I know how to write a ladder checker in general,
but I'm not sure how to incorporate such a thing into UCT. What are
other people doing?
Peter Drake
http://www.lclark.
Mark:
Can you say more? Do you mean ALWAYS play such moves first if they
are available, do so with some probability, or what?
Peter Drake
http://www.lclark.edu/~drake/
On Jun 16, 2008, at 4:09 PM, Mark Boon wrote:
play ladder-capturing and ladder-escaping moves during playout
Without the 10% random moves, would every playout from a given leaf
be identical?
Peter Drake
http://www.lclark.edu/~drake/
On Jun 17, 2008, at 3:14 AM, Magnus Persson wrote:
Valkyria plays uniformily the highest ranked move. Ladders as a
response to the last move are almost always
x27;t yet heard from MoGo, CrazyStone, or any other programs known
to be very strong. You might do better than you think!
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/m
Orego Peter Drake Same as above
FirstGo Edward de Grijs Needs operator, will borrow hardware
ManyFaces David Fotland
Argus Sam Gross
HouseBotJason House Needs operator, will borrow hardware
Any others? None of the very strong UCT programs are here, s
Terry -- thanks for the offer! We'll likely take you up on it.
We're looking at starting after lunch on Sunday and finishing up by
dinner on Tuesday. Hopefully things will be automated enough that, if
necessary, games can continue overnight.
Peter Drake
http://www.lclark.edu/~dr
It looks like the server is running, but the standings page is not
being updated:
http://cgos.boardspace.net/9x9/standings.html
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer
lately. Because the tree is growing, the
values of the moves are moving targets.
Has anyone done any work dealing with this phenomenon, e.g., somehow
giving more weight to more recent playouts?
Peter Drake
http://www.lclark.edu/~drake
siderable improvements since the results reported
here. It's not as strong as UCT yet, but one of our genetic player
(Orego-genetic2) has a rating of 967 on the 9x9 CGOS. Expect more in
the coming weeks...
Peter Drake
http://www.lcl
On Jun 26, 2008, at 1:35 AM, Magnus Persson wrote:
Quoting Peter Drake <[EMAIL PROTECTED]>:
UCB (and hence UCT) would treat the following sequences of wins
(1) and
losses (0) the same:
01010101010101010101010101010101
00
Can anyone point me to a thread, or at least some buzzwords?
I'm having little luck googling for words like "recent" and "forget".
Thanks,
Peter Drake
http://www.lclark.edu/~drake/
On Jun 26, 2008, at 11:40 AM, Ivan Dubois wrote:
This same topic already occured
Just what I was looking for -- thanks!
On Jun 26, 2008, at 12:04 PM, Rémi Coulom wrote:
Peter Drake wrote:
Can anyone point me to a thread, or at least some buzzwords?
I'm having little luck googling for words like "recent" and "forget".
Thanks,
Peter Drake
htt
Since a few people have asked, the stats on the machines we have to
loan for the tournament are given below.
Peter Drake
http://www.lclark.edu/~drake/
Begin forwarded message:
You bet. The boxes we have are currently running Linux, but
can be made to dual-boot Windows if we need that for
quickly resolved.
Sent from my iPhone
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
Peter Drake
http://www.lclark.edu/~drake/
___
comput
. The abortted playout trick would cause the
bot to cycle uselessly until the game takes another path.
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
Ah, I forgot that you had a transposition table. (Orego currently does
not.)
I, too, will be interested to hear the solution to this problem.
Peter Drake
http://www.lclark.edu/~drake/
On Jul 11, 2008, at 10:49 AM, Jason House wrote:
On Jul 11, 2008, at 12:43 PM, Peter Drake <[EM
ist
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
ss.
So it cuts off the infinite loop.
On Fri, Jul 11, 2008 at 12:08 PM, Peter Drake <[EMAIL PROTECTED]>
wrote:
My sense is that most programs ignore superko except for checking
right before a "real" move (as opposed to a playout move) is played.
The way out of the infini
I believe all of the programs that need operators (FirstGo, HouseBot,
and Leela) all run on Windows. I hope someone will take Terry up on
his generous offer!
Peter Drake
http://www.lclark.edu/~drake/
On Jul 11, 2008, at 12:30 PM, terry mcintyre wrote:
I will be attending the US Congress
processor cores. The game will be broadcast on KGS.
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
19x19. There will be a handicap. We're currently planning on five
blitz games to adjust the handicap, then one "real" game.
On Jul 21, 2008, at 2:07 PM, Eric Pettersen wrote:
9x9 or 19x19?
On Jul 21, 2008, at 2:04 PM, Peter Drake wrote:
(This is from the US Go Congres
Pacific time.
We'll do this in the Computer Go room. We'll announce the usernames
when the time comes.
On Jul 21, 2008, at 2:28 PM, Jason House wrote:
1pm in which timezone? Which room & user name(s) will be used on KGS?
Sent from my iPhone
On Jul 21, 2008, at 5:04 P
difficult.
I think the interesting algorithmic area is somehow localizing the
search. My team is working on it...
The community is quite good. I wonder if a 13x13 CGOS would help,
because many of us are doing well at 9x9, but 19x19 is MUCH harder.
Peter Drake
http://www.lclark.edu/~drake/
On
some attention.
Second, participants. Here's my current list:
Program Primary Author Notes
SlugGo David Doshay As the author is involved in organizing the
tournament,
this program will not be eligible for
prize money
Orego P
Indeed! That's part of the motivation of organizing the tournament at
the US Go Congress.
Perhaps we (or the subset of us within a given country) could just
pick an existing conference (something on machine learning or games)
and all go there...
Peter Drake
http://www.lclark.edu/~
it's 50-50 whether
the attacker or defender will play on the defender's last liberty.
Thus, the ladder reader doesn't give any pressure to stop running when
caught in a ladder.
What am I missing?
Peter Drake
http://www.lclark.edu/~drake/
__
On Jul 31, 2008, at 4:24 PM, Mark Boon wrote:
On 31-jul-08, at 19:50, Peter Drake wrote:
I know we had this conversation recently, but I just can't seem to
get my head around writing a ladder reader. What, exactly, does the
ladder reader do?
Our approach was to read out la
nough liberties?
If we only consider the last stone and its neighbors, moves elsewhere
on the board will look disproportionately bad because they "disable"
the ladder searcher.
Peter Drake
http://www.lclark.edu/~drake/
On Jul 31, 2008, at 5:06 PM, terry mcintyre wrote:
Fro
make the vital
capture at b.
It seems too expensive to search every point on the board for ladders.
What to do?
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
f you have any special needs regarding libraries, etc.,
please let them know ASAP. The address is:
[EMAIL PROTECTED]
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailma
I prefer consistent 13x13 on the third server, since I use CGOS to
determine if some change to the program is an improvement.
Peter Drake
http://www.lclark.edu/~drake/___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org
m commits to a particular move too early, builds up a lot
of playouts through that move, and keeps playing the move even though
it has read a sequence of follow-ups that lead to a bad result.
Peter Drake
http://www.lclark.edu/~drake/
On Aug 2, 2008, at 7:06 AM, Álvaro Begué wrote:
On Sa
For your information.
Peter Drake
http://www.lclark.edu/~drake/
Begin forwarded message:
From: Aaron Fellin <[EMAIL PROTECTED]>
Date: August 2, 2008 1:04:26 PM PDT
To: Barton C Massey <[EMAIL PROTECTED]>, Peter Drake <[EMAIL PROTECTED]>
Cc: [EMAIL PROTECTED]
Subject: R
The Linux lab is in the Fourth Avenue Building, room 81-03. Leave
some time to find it; the building is rather labyrinthine.
I'll be there by 8:30 AM Monday, possibly a bit earlier, so hopefully
people can set up and then go play in the US Open.
Peter Drake
http://www.lclark.edu/~
To anyone who's here at the Congress:
The American Go Association (most likely the board thereof) is having a meeting
about "Computer / Internet Development" in Smith 229 at 4:00 PM. Since
decisions are made by people who happen to be in the room at the time, it might
be a good idea for us to be t
(Please propagate this information widely. I will make some posters and put them
up around Smith.)
Thursday, August 7, 1 PM PDT, Kim MyungWan 8p will play against MoGo, the
world's strongest computer Go program. MoGo will be connecting remotely,
running on a European supercomputer using hundreds i
I would be in favor of 7.5 everywhere.
We used 7.5 at the US Go Congress computer tournament and got 23/42
wins for white. Of course, every game but one ended in resignation...
Peter Drake
http://www.lclark.edu/~drake/
On Aug 6, 2008, at 5:57 AM, Don Dailey wrote:
I'm wondering
velop that instinct (or something analogous) over the
course of the game. If we could only decide what to store (and how to
store it) between games, we could get good blitz performance and
superhuman long game performance.
Peter Drake
http://www.lclark.edu/~drake/
On Aug 7, 2008, at 10:2
ience were surprised that MoGo was capable of
resigning.
Peter Drake
http://www.lclark.edu/~drake/
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
1 - 100 of 310 matches
Mail list logo
