In message <5212e61a0911231302j6d838d2dnae1cbc875af0...@mail.gmail.com>,
Don Dailey <dailey....@gmail.com> writes
On Mon, Nov 23, 2009 at 3:23 PM, Nick Wedd <n...@maproom.co.uk> wrote:
In message <
5212e61a0911231136t1e83ce37i9375a033fe3e0...@mail.gmail.com>, Don
Dailey <dailey....@gmail.com> writes
On Mon, Nov 23, 2009 at 12:01 PM, Robert Jasiek <jas...@snafu.de>
wrote:
Don Dailey wrote:
In win game mode [God] will play ANY move randomly that is
"good
enough."
If God is set to play any randomly chosen winning move, yes.
Since it is omnicient there is no point in talking about
risk,
or chances
in any context.
For a simple definition of God applied to a single game, yes.
For an
entity in strength between God and Devil (who knows also the
opponent's strategy in hindsight), possibly no. For God without
hindsight during a tournament, no. For Devil in a single game or
Devil with tournament hindsight, yes.
> In a lost game it would play a move at random.
Why random?
I don't understand the question. If all moves lose, how would
YOU
select?
Did you get the point that I'm defining 2 separate strategies?
One
is to maximize the points on the board and the other is to not
make any
distinction whatsoever between moves except whether they win or
lose
the game.
And I'm trying to make the point that maximizing the points on
the
board is a superior strategy because it is a super-set of the
strategy
to be only concerned with winning.
"Superior" in what sense? Your bang neki strategy is superior if
you are playing bang neki, and inferior if you are playing Go. The
Go strategy employed by MC-UCT programs is superior for playing Go
and inferior for bang neki.
I don't know what bang neki means but it's superior against fallible
opponent, but against perfect opponent it's not inferior. You can
argue that it's not superior agianst perfect opponents and I would
agree, but it's not inferior either. In other words it's greater
than or equal to playing for the win if you are god.
If you are NOT god, then it's "easier" to play for the win because
there is simply less to think about. You only distiguish between wins
and losses and not the magnitude of them, which is a simplification for
us mere mortals. In chess, it is said that WHITE has an
advantage, but that is probably only true for falliable players, since
it's probably a draw from gods point of view. But for us it's easy to
win when we have the white pieces.
Let's call this strategy A and strategy B. Strategy A is to
maximize
the points on the board and strategy B is to only distinguish
winning
moves. If you play strategy A, then a strategy B player would
see
those moves as a perfectly valid B strategy. But a strategy A
player would frown on many of the moves a strategy B player would
play.
Are you assuming that the players can examine the entire game tree?
Yes. I'm saying that strategy A is >= strategy B if you are God (it
will not help against another God but it won't hurt either. But it
will help you win against a mortal, therefore I say A >= B)
If you are NOT god, then strategy B apparently is better. It's better
for computers for sure.
Ok, so you were talking about infallible players.
If you are, I do not understand your paragraph above, the whole
issue seems undefined. But if you assume that they (like me) cannot
read everything out with certainty, I disagree with your conclusion.
I think you actually agree with me. Strategy A is >= B if you are
God, otherwise strategy B appears to be best.
Indeed, I often see players using strategy A in a poor position, by
trying to play out the yose well, when I can see that their only
hope of winning the game is to start a messy fight (which none of us
knows who will win). I do not consider their strategy A as
"perfectly valid".
If you are fallible, I agree that strategy B is best. Strategy A is
just as good for winning as strategy B but only if you are God.
However, if you ARE god, then strategy A is better than strategy B
against fallible players and against another God player it doesn't
matter.
I think this has been confusing because there are too many frames of
reference here and I probably didn't explain it very well. I
really only set out to explain why I think the Hahn tournaments may be
harder to play after all because it seems to require strategy A, which
is harder for fallible players to do as well. (Humans and computers
have not mastered either strategy of course but I think strategy B is
easier to handle.)
As for the question of which strategy is harder - I am thinking of golf
as an analogy. Or rather, a simplified one-stroke golf. Strategy A is
like "hit the ball so as to minimise its expected distance to the
green". Strategy B is like "hit the ball so as to maximise its
probability of landing near to the green than your opponent's did (or
will do)". I have no opinion about which is harder. I do think that
the latter would be more fun, but that's not relevant.
Bang neki is a form of Go, played in Korea for significant sums of
money. The loser (or his backer) pays a sum dependent on the winning
margin.
Nick
In maximize score mode it would choose the move that
maximizes
the total
points taken on the board. It would be the perfect Hahn
system
player
> for instance.
Wrong, since Hahn system has an upper score rewarding boundary.
(The
thing that punishes me for having taken a "too great" risk when
killing 70 stones groups.)
> What I cannot decide is if it is really more
challenging - I just know it's more challenging to do it
perfectly.
More challenging for whom? For God, it is equally boring.
--
robert jasiek
_______________________________________________
computer-go mailing list
computer...@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
--
Nick Wedd n...@maproom.co.uk
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
--
Nick Wedd n...@maproom.co.uk
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/