Re: [computer-go] Protocol B

["Eric Boesch" <ericboesch@xxxxxxxxxxx>]

My comment about "Japanese style" was misunderstood (and, as Peter says, not 
central anyhow).    It appears that SlugGo plays AREA RULES *with a 
territory-rules style*, just as Japanese players usually do when they play 
by Chinese rules -- that is, SlugGo would rather pass than play an 
unnecessary move inside its own territory.  That is all I meant by "Japanese 
style".  People play perfectly correct and smart games under area rules in a 
Japanese style all the time.  I brought it up because David Doshay clearly 
prefers this style -- and because Protocol B does not support it in an 
elegant way.  If players refuse to capture dead stones except when faced 
with a genmove_cleanup, then an iniitial disagreement under Protocol B will 
lead to many useless passes and agreement phases.

The issue with Protocol B isn't so much "catering to" lazy or exploitative 
programs as "watching out for" them, because their behavior affects their 
opponent as much as themselves.  No side is penalized more than the other 
when the game gets dragged out.  Moreover, Protocol B expands the definition 
of lazy or exploitative.  Passing whenever you can isn't so lazy or 
exploitative under Protocol A.  Playing Tromp-Taylor is fine by Protocol A.  
Letting people be lazy is good.

> From: "Peter McKenzie" <peter_mckenzie@xxxxxxxxxxxxxxxxx>
[...]
> As someone who is potentially the author of type 1 (smart) program, would I 
> allow my program to be bombarded to a huge number of extra commands (all of 
> which take non-trivial time for my program to process)?  No way!
Yes, you would.  You're free to set about capturing all your opponent's dead 
stones, but if he wants a hundred or more agreement phases, then that's what 
you get, because reasonable precautions won't stop it.

Let's suppose your smartbot does almost everything right -- as you said, you 
don't make things worse by attempting extra agreement phases of your own, 
for example.  Suppose you each end up with 120 points apiece of territory 
after 100 moves -- that's a lot of territory, but it does happen, and I'm 
trying to look at upper bounds.  Your opponent, annoybot, is black.  So:

100 moves to start

Opponent fills your eyespace: about 50 times, result b move, w move (to 
capture b), b pass, w pass, disagree = 200 moves + 50 disputes

Opponent fills own eyespace: 118 times, result w pass, b move, w pass, b 
pass, disagree = 472 moves + 118 disputes

(If suicide is allowed, these numbers are too low.)

Total game duration under Protocol B:  772 moves + 168 disputes, without any 
unusual situations (molasses ko, et cetera)

Total game duration under Protocol A: 438 moves + 1 dispute

This is what a defensive program has to be prepared for:  games that can be 
nearly double as long under Protocol B.  The disputes will often last 5/2 as 
long *whether or not they are induced maliciously*, unless BOTH players take 
pains to play nicely under this protocol.

Now, with perfect play, white could shorten the game slightly, but that's 
about it, and here "perfect play" means here filling your own territory even 
when there are no prisoners left to capture, and I don't think many people 
will expend much effort in that direction.

> Overall I could live with either protocol but I favour protocol A as I 
> think it gives the best balance between features and simplicity.  It 
> happily handles the best case scenario (where both programs agree on 
> score/dead stones) just as well as protocol B, but doesn't have the nasty 
> worst case scenario of protocol B.  Protocol A doesn't give programs a 
> second chance to agree on score/dead stones but in practice I doubt that 
> this is important - if they don't agree first time then I suspect it is 
> likely they won't ever agree.  And lastly, Protocol A involves fewer 
> complications for the program authors to worry about.
Agreed.  Protocol C is interesting, but I prefer Protocol A's simplicity.  
Protocol B's unnecessary agreement shenanigans, I don't like.  There's a 
reason that NZ and AGA rules, which were intended for humans, don't have 
multiple agreement either.

> Anyways, I've probably said more than enough on this issue so I'll try to 
> make this one my last post on it!!
I will too, I hope.

Eric

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/