Re: Minesweeper solver

Article: 7794 of alt.hackers
From: bpheintz@bu.edu (Brad Heintz)
Newsgroups: alt.hackers
Subject: Re: Minesweeper solver
Date: 10 May 1995 22:09:09 GMT
Organization: Information Technology, Boston University, Boston, MA, USA
Lines: 50
Approved: Shyah-hah!  As if!
Message-ID: 3ordi5$qla@news.bu.edu
NNTP-Posting-Host: sunflower.bu.edu
X-Newsreader: TIN [version 1.2 PL0]
Status: RO

Peter Seebach (seebs@solutions.solon.com) wrote:
: In article <3odgpn$jch@erinews.ericsson.se>,
<eeittn@eei.ericsson.se> wrote:
: >There already exists a cheat program for minesweeper.
: [snip]
: >choice remained.  Was very good and could do all the
: >levels down to about 2 or 3 remaining without guessing.

: I don't believe it.  I think it's safe to say that there is no way
: to even start a minesweeper level without guessing.

Actually, *any* square you choose will be a free one.  Try it.  When you
think about it, it'd be sort of a no-brainer to implement.  If you don't
get a "0" square the first try, though, the second one will be
guesswork.  What I'd be interested to know, though, is whether or not
the solver mentioned earlier in the thread had a semi-smart
probabalistic guessing mechanism, or if it just guessed randomly when it
was stuck, i.e., did it ask "Based on the number of mines and the number
of untouched squares, am I better off choosing randomly, or going with a
square bordering a known square?"

ObHack: Back in my undergrad days, I took a course in Computation
Statistcial Physics.  For one assignment, we were to compute the net
field of a given piece of magnetic material (in 2-D), and given certain
external conditions (temperature, time-dependent external field, et al)
model the changes the net field of the area of interest over time.
Well, everyone else did it by probabalistically determining the behavior
of individual domains over time, then summing the microscopic fields to
get the macroscopic field.

After goofing with it a while, I figured out that for a region of the
size we were looking at, it was just as accurate and far easier simply
to determine the number of domains pointing in each direction, and
probabalistically determine how that number would change with time.
Everyone else had pages of code, and the professor turned red when I
turned up with half a page of source - in VAX BASIC, no less!.

Quote from the prof:  "You've completely trivialized my problem."

It felt good, I must confess.

- Brad

--
*-------------------------*------------------------------------------*
|Reach Brad at:           |          "I'm growing older but never up,|
|bpheintz@sunflower.bu.edu|         My metabolism's comfortably stuck|
|Home Page coming         | Let the winds of change blow over my head|
|Real Soon Now!           |     I'd rather die while living than live|
|                         |          while I'm dead."  - Jimmy Buffet|
*-------------------------*------------------------------------------*

Child

Back to index