home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!spool.mu.edu!wupost!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!rutgers!hsdndev!husc-news.harvard.edu!husc.harvard.edu!kleber
- From: kleber@husc11.harvard.edu (Gwydden)
- Newsgroups: rec.games.abstract
- Subject: Re: defects in abstract games
- Message-ID: <kleber.725596933@husc.harvard.edu>
- Date: 29 Dec 92 02:42:13 GMT
- References: <1992Dec22.143005.11060@ll.mit.edu> <kleber.725061503@husc.harvard.edu>
- <1hie51INNnvd@usenet.INS.CWRU.Edu> <1992Dec28.224523.28467@news.arc.nasa.gov>
- Lines: 51
- Nntp-Posting-Host: husc11.harvard.edu
-
-
- I don't *think* I was the one who started the "defect-free games"
- thred-- maybe just helped it along a bit. In any case, Melfin
- Nicholson gets into the action and proposes some criteria for defect-free:
-
- > 1: state should be limited to position of pieces on the board and whose
- > turn it is, with no reference to 'previous board positions' and other
- > phenominon which must be tracked
-
- So you would consider a "no repeated positions" rule a defect? I'm
- not sure whether I do or not-- but then again, these are very aesthetic
- questions; no surprise that we don't all agree. Hmm-- what happens
- if you eliminate such a rule, anyway? You end up im places where
- the players might loop forever... if you just say that's a loss for
- both of them, does that take care of it? No, I suppose not, because
- there's no good way to decide which one should break the loop.
- Pity.
-
- > 2: the rules should not admit special cases and be symmetric to all
- > peices and positions
-
- Sometimes it's hard to say what's a special case. The starting
- position in [othello/reversi] always seemed somewhat arbitrary to
- me; one of the beauties of go was the fact that the board starts
- empty, and only the strategy inherent in the rules determines
- the first move.
-
- > 3: the rules should be stable in a concise way, and not too numerous
-
- (statable, I assume you mean.) Hmm (evil grin)-- how about Mornington
- Crescent? From one point of view, the rules are not too numerous,
- but not statable at all; from another, there are arbitrarily many
- rules. And yet, it's an incredibly aesthetic game... ah, but I suppose
- there are those that would argue that it doesn't qualify as an abstract
- game, given the definitions I've seen posted here. (imho, if MC isn't
- abstract, what *is*, but that's personal...)
-
- > While I agree that defect free status isn't all that wonderful of a
- > thing (as someone pointed out, tic-tac-toe is quite 'defect free' in
- > this since, but in general is a lousy game), I do sympathise with the
- > aesthetic quality the original author was trying to put his finger on.
-
- I think flawlessness needs some kind of not-too-easy characteristic...
- I think that perfect play being known is definitely a flaw, for example.
- Yes, I realize how arbitrary "known" is... maybe "practical to play"?
- One of the appeals of "mental jujitsu" is that it's state-tree isn't
- computable, because players don't alternate-- but I think that's
- *too* much to ask from a poor innocent little game...
-
- --Michael Kleber I don't have an overactive imagination...
- kleber@husc.harvard.edu I have an underactive reality... --EG
-