home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!pavo.csi.cam.ac.uk!camcus!gjm11
- From: gjm11@cus.cam.ac.uk (G.J. McCaughan)
- Newsgroups: sci.math
- Subject: Re: Chess Problem
- Message-ID: <1992Sep13.003856.13264@infodev.cam.ac.uk>
- Date: 13 Sep 92 00:38:56 GMT
- References: <BuFpLp.9nI@ecf.toronto.edu> <1992Sep12.174545.23214@ima.isc.com>
- Sender: news@infodev.cam.ac.uk (USENET news)
- Organization: U of Cambridge, England
- Lines: 45
- Nntp-Posting-Host: apus.cus.cam.ac.uk
-
- In article <1992Sep12.174545.23214@ima.isc.com>, karl@ima.isc.com (Karl Heuer) writes:
-
- > Here's a more advanced chessboard question.
- >
- > A white king, a black king, and a white pawn are randomly placed on an NxN
- > chessboard. (Independent, uniform distributions.) What is the limit, as
- > N -> infinity, of the probability that the position is a win for white?
- >
- > (Note that it doesn't matter what you do about collisions or illegal
- > positions, since they occur with probability zero in the limit.)
-
- Eek. Let's think... (This is much better than the last one, to which the
- answer is of course 8!/(64 choose 8) = 560/61474519, or about 9.109e-6)
-
- If the pawn is a "rook's pawn", funny things happen; but that's only
- probability 2/N, so we can forget that.
- If the black king is behind the white pawn [but not next to it, but
- that's only 8/N^2] then White wins, always.
- If the black king is in front of the white pawn, but "too far from" it
- (i.e. can't reach the promoting square before the pawn does) then White
- wins by rushing the pawn forwards.
- If the black king is in front of the pawn, and in range, then it all depends
- on how the two kings are placed. If the white king is further away from the
- pawn than the black one, then Black clearly draws. If the white king is closer
- to the pawn, then White wins (except in O(1/N) of the cases, where it depends
- on parity considerations) by getting in front of the pawn and keeping the
- opposition.
-
- [thinks: if I haven't made a mistake, all the interesting chess is in O(1/N)
- of the cases...]
-
- Right. So we want the probability that the black king is (i) within the pawn's
- "light cone" and (ii) closer to the pawn than the white king. A moment ago I
- thought this was really easy, but it's not quite as nice as I thought...
- I think it might be preferable to place the kings first, so to speak,
- and then we want the probability that the pawn is within the "reverse light
- cone" of the black king and closer to it than to the white king. Ugh.
-
- I don't think I have the patience for this right now, at least not without
- pencil and paper to hand. Still, I've reduced the problem to the point at
- which it's just a matter of summing some fairly easy (ha!) series.
-
- --
- Gareth McCaughan Dept. of Pure Mathematics & Mathematical Statistics,
- gjm11@cus.cam.ac.uk Cambridge University, England. [Research student]
-