rec.puzzles Archive

You have three casks holding 19, 13 and 7 units.

logic/three.casks New

New entry, edited from material collected for the archive and not yet in a released version.

Problem

You have three casks holding 19, 13 and 7 units.  The 13 and 7 are
full and the 19 is empty.  By pouring from one cask to another, with
no other measures, decant the liquid to the required amounts.

[ Editor: problem statement reconstructed from the postings; the
original statement was not collected.  The posted solution ends with
10 units in the 13-cask, which may have been the goal. ]
Show solution
I think I got it

     Contents           Action
19        13        7        taken
==============================
0        13        7        Start
7        13        0        7 to 19
7        6        7        13 to 7
14        6        0        7 to 19
14        0        6        13 to 7
1        13        6        19 to 13
1        12        7        13 to 7
8        12        0        7 to 19
8        5        7        13 to 7
15        5        0        7 to 19
15        0        5        13 to 7
2        13        5        19 to 13
2        11        7        13 to 7
9        11        0        7 to 19
9        4        7        13 to 7
16        4        0        7 to 19
16        0        4        13 to 7
3        13        4        19 to 13
3        10        7        13 to 7

Is this correct?

-- tclancy@rapnet.sanders.lockheed.com

Very nice - hope it was fun.

-- Jay Rothman <jrothman@carbon.denver.colorado.edu>

[ Editor: a later restatement and table for every amount, rec.puzzles,
October 1994. ]

Jay Rothman (jrothman@carbon.denver.colorado.edu) writes:
> This certainly answers the question the way that I asked it. I need to be
> more precise. You are not allowed to use any other vessels - you must end
> up with the 10 qts in one of these 3 casks.

OK, the correct problem statement would be:

  You have three casks of sizes 7, 13 and 19 quarts.  Starting with the
  7 and 13 quart casks full, and without throwing away any of the
  liquid, finish with n quarts in one of the casks, for 1<=n<=19.

Again, number of operations are in parentheses.  I think most of them
are minimal.

 1  (2) 13-19,7-19
 2  (7) 13-19,7-19,7-13,19-7,7-13,19-7,7-13
 3 (13) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7,7-13,19-7,7-13
 4 (12) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7,7-13,19-7
 5  (6) 13-19,7-19,7-13,19-7,7-13,19-7
 6  (3) 7-19,13-7
 7  (0)
 8  (5) 13-19,7-19,7-13,19-7,7-13
 9 (11) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7,7-13
10 (15) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7,7-13,19-7,7-13,
        13-19,19-7
11 (10) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7
12  (4) 13-19,7-19,7-13,19-7
13  (0)
14  (3) 7-19,13-7,7-19
15  (9) 7-19,13-7,7-19,13-7,19-13,13-7,7-19,13-7,7-19
16 (15) 7-19,13-7,7-19,13-7,19-13,13-7,7-19,13-7,7-19,13-7,19-13,13-7,
        7-19,13-7,7-19
17 (14) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19,7-13,19-7,7-13,19-7,7-13,
        13-19
18  (8) 13-19,7-19,7-13,19-7,7-13,19-7,7-13,13-19
19  (2) 13-19,7-19

-- Gareth Rees <gdr11@cl.cam.ac.uk>

I went ahead and checked the solution for decanting 10 quarts, just
because it looked interesting...and it has a small bug.  The first
fourteen steps leave 3 quarts in the 7-quart cask and 17 quarts in the
19-quart cask (the solution for decanting 17 quarts).  To end up with 10
quarts in the 19-quart cask, you have to empty the 3 quarts in the
7-quart cask into the 13-quart cask first, and *then* pour seven quarts
from the 19-quart cask into the 7-quart cask.  So, the actual solution
has 16 steps, with the second to last step being 7-13.

Unless I made a mistake myself, I think the correction is needed.  Has
anybody confirmed this?

-- Michael Wolfe <wolfe@ils.nwu.edu>

blask@gpx01.d39.lilly.com wrote:
: I recall seeing a book about 25 years ago that had an algorithm for
: solving any of these "3 casks" problems.  The algorithm involved
: drawing the path of a billiard ball on a trapezoidal billiard table
: and observing where it bounced.  Does any of this sound familiar to
: anyone?  What's the name of the book?

You have a good memory from 25 years ago.  The pool table solution was
mentioned by an earlier poster.  I was able to find it in Steinhaus,
Mathematical Snapshots.  It probably appears in many other places.

-- Jay Rothman <jrothman@carbon.denver.colorado.edu>