GONNECT
João Neto
>>"João Neto" wrote:
>> GONNECT
>...
>> GOAL - A player wins if he can create a chain of connected
>> stones between two opposite board sides (west-east or north-south)
>> or if his adversary has no valid move.
>
>I conjecture that with best play this is a draw. Have you
>looked at the game Hex, which has a similar goal and even
>simpler rules? When the board gets large enough it's quite
>challenging.
yes, I know Hex, Y and Havannah. Gonnect is also
a connection game, but the main difference is the
capturing option (and in this way, territory has
a double importance, since you must worry about
connections and captures).
You should think about the strategic global connection
on the long term, and in tactical attacks/defenses
around local connections on the short term.
The rule that forbids two hole structures is for
avoiding draw oriented strategies, in that way,
all pieces are capturable.
Since Gonnect inherits part of the depth of Go,
perfect play would be very hard (my humble guess...).
And in that sense, Gonnect has more depth than
Hex (with similar boards). I suppose that is natural,
since the capturing rule adds more complexity to
game strategies (excluding that rule, Gonnect would
be like Hex on a square board)
The board size is very important, as you noticed,
like in Hex. I suppose the 19x19 Go board
seems sufficiently complex for it (eventually it
is too great, perhaps the 15x15 Renju board
would be better...)
Joao Neto
Jim Gillogly
> The rule that forbids two hole structures is for
> avoiding draw oriented strategies, in that way,
> all pieces are capturable.
Naively, I think it might be difficult to see whether a group
is capturable. For example, it seems likely that a group
might not have two eyes, but it still could not be captured
because it would give the opponent's surrounding group two
eyes. Would the player with an otherwise capturable group be
prevented from making a formation that would result in two
eyes for the opponent if captured, since that makes the group
in effect uncapturable? That is, the opponent can't take the
group without making his own two-eyed live group, so wouldn't
the first player have made a live Gonnect group without having
made a live Go group?
This seems like a rather recursive rule set... I can try
to construct such a position if this seems unclear. Or is
there some theorem that says life and death will be obvious?
--
Jim Gillogly
Highday, 20 Afterlithe S.R. 2000, 00:39
12.19.7.6.14, 8 Ix 17 Tzec, Eighth Lord of Night
David Rush
> "João Neto" wrote:
> > The rule that forbids two hole structures is for
> > avoiding draw oriented strategies, in that way,
> > all pieces are capturable.
>
> Naively, I think it might be difficult to see whether a group
> is capturable.
I was also doubted the viability of this rule.
> Or is there some theorem that says life and death will be obvious?
The opposite actually. Life and death problems are to Go what endgame
problems are to Chess (well, more or less since Go has different
victory conditions).
david rush
--
From the start...the flute has been associated with pure (some might
say impure) energy. Its sound releases something naturally untamed, as
if a squirrel were let loose in a church." --Seamus Heaney
Richard Rognlie
>
> Matti Siivola wrote
>
> >If players are not allowed to pass, the after the teeritiories are
> >settled and dame filled they must continue filling their own territory
> >and eventually filling eyes. The player who has more territory apart
> >from group tax and some move gaining tesujis will survive since the
> >opponent must fill eyes of some group first.
>
> While this is undoubtedly the case almost always (that the player first
> forced to fill down to one eye and one liberty will lose) I regard it as an
> unsolved problem in the general case. This was discussed in a thread not so
> long ago, but inconclusively.
>
> It can be explained like this: imagine the situation after a (large) group
> with one eye and one liberty has just been captured. There's a big hole on
> the board, edged by stones of the other colour. Playing first in such a
> hole is worth something - a certain "komi". You have to prove a theorem
> about this komi.
>
> In all normal situations where White has just had such a group captured,
> White by playing back in can't do very much that is useful. However in
> extreme case that ceases to be true.
You need some form of no suicide rule, else the game will stagnate.
e.g. "You can not place a stone such that it dies alone". This would
prevent
me from sacrificing a stone to a known live enemy group in an effort to
not be
forced to fill in my own eyes.
But even that statement is not strong enough.. e.g. if my opponent had a
group
with 2 eyes, one of them 2 spots big... I could place two stones there
which
then die together ("not alone") But this has the affect of letting me
"pass"
which the rules specifically forbid.
--
/ \__ |
\__/ \ | Richard Rognlie / Sr. Technical Consultant / Sendmail, Inc.
/ \__/ | http://www.gamerz.net/rrognlie/ <rrog...@gamerz.net>
\__/ |
Eric Osman
I think we've seen enough discussion now that it is useful to have a
prototype executable of the
game so we can actually try it. Someone want to whip one up in java ?
I like this explanation:
There's no need to enforce any rule saying it's illegal to
make an uncapturable
group. Merely enforce the rule that passes and suicides are
illegal.
With the above rule, a two-eyed group would eventually be capturable
since the person who
made it would run out of places to play and be forced to fill their own
eye.
Let's look at a sample 5x5 game:
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
Aha ! After drawing a blank board, a question comes up. Do we
immediately establish that
white's goal is to make the horizontal connection and black's goal is to
make the vertical
connection ?
Or is this left up to the players ? I'll assume it's up to the players
for now.
. . . . .
. . . . .
. . X . .
. . . . .
. . . . .
. . . . .
. . . . .
. O X . .
. . . . .
. . . . .
. . . . .
. X . . .
. O X . .
. . . . .
. . . . .
. . . . .
. X . . .
. O X . .
. . O . .
. . . . .
. . . . .
. X . . .
. O X . .
. X O . .
. . . . .
. . . . .
. X . . .
. O X O .
. X O . .
. . . . .
. . . . .
. X X . .
. O X O .
. X O . .
. . . . .
. . . . .
. X X O .
. O X O .
. X O . .
. . . . .
. . . . .
. X X O .
. O X O .
. X O X .
. . . . .
. . . . .
. X X O .
. O X O .
. X O X O
. . . . .
. . . . .
. X X O .
. O X O .
. X . X O
. . X . .
. . . O .
. X X O .
. O X O .
. X . X O
. . X . .
. X . O .
. X X O .
. O X O .
. X . X O
. . X . .
Did X win yet ? Or does X need to fill to truly make a solid line of
stones ?
. X . O .
. X X O .
. O X O .
. X . X O
. . X . O
. X . O .
. X X O .
. O X O .
. X X X O
. . X . O
Now X has certainly won.
Jim Gillogly
> I like this explanation:
>
> There's no need to enforce any rule saying it's illegal to
> make an uncapturable
> group. Merely enforce the rule that passes and suicides are
> illegal.
...
> Let's look at a sample 5x5 game:
...
> . . . . .
> . . . . .
> . . X . .
> . . . . .
> . . . . .
>
> . . . . .
> . . . . .
> . O X . .
> . . . . .
> . . . . .
I think O's best move here is to change the X to an O (allowable on
the first move only). This forces X to pick something that's not
dominant, but rather tries to make the position after O's best move
as even as possible. Cute rule! I divide the portions, you pick.
--
Jim Gillogly
Sterday, 21 Afterlithe S.R. 2000, 05:59
12.19.7.6.15, 9 Men 18 Tzec, Ninth Lord of Night
João Neto
Eric Osman escreveu na mensagem <396E8C63...@mediaone.net>...
>
>I think we've seen enough discussion now that it is useful to have a
>prototype executable of the
>game so we can actually try it. Someone want to whip one up in java ?
I'm playing a game in Richard's PbEM server with himself using the 9x9
GO game facilities, and another with a 7x7 board. I will post them
here when they end.
If anyone wants to make a Java applet, I would be very grateful for it!!!
>I like this explanation:
>
> There's no need to enforce any rule saying it's illegal to
>make an uncapturable
> group. Merely enforce the rule that passes and suicides are
>illegal.
I like it too. Make things a lot more elegant
>Aha ! After drawing a blank board, a question comes up. Do we
>immediately establish that
>white's goal is to make the horizontal connection and black's goal is to
>make the vertical connection ?
it's up to the players, i.e., it is not like Hex!
> . . . . .
> . . . . .
> . . X . .
> . . . . .
> . . . . .
This move would make White call for the PIE rule, ie, exchange
places :-)
Joao
João Neto
>...
>> . . . . .
>> . . X . .
>> . . . . .
>> . . . . .
>>
>> . . . . .
>> . . . . .
>> . . . . .
>> . . . . .
>
>the first move only). This forces X to pick something that's not
>dominant, but rather tries to make the position after O's best move
>as even as possible. Cute rule! I divide the portions, you pick.
Precisely! The PIE rule is a very nice way to balance unbalaced games,
it was sugested to me by Stephen Tavener.
Joao