Science Forums 2

[CraigD]

Sorry to have ignored this thread for the past week

I get an “installer corrupt or incomplete” message when attempting to run the windows installer downloaded from Flowix Games, so haven’t tried the game yet. I found this copy of a game called “Sherlock”, which seems to be the same game, and finally gave it a try. I finished its first (and presumably easiest) game in an rather pathetic 13:58, though I’m sure I’d get quicker with repeated playing.

“Einstein’s puzzle” usually refers to a 5x5 (rather than Sherlock’s 6x6) logic puzzle called the “Zebra Puzzle”, which, legend has it, either Albert Einstein or Lewis Carol claimed only about 2% of people could solve. Usually speed isn’t considered much in solving logic puzzles – an experienced logic puzzle solver typically takes 10 to 30 minutes, using paper and pencil or equivalent, much of the time spent organizing the attributes and clues, which computerized version do for you.

Some logic puzzles termed “Einstein’s puzzle”, such as the “Fish” variation, can be solved with a “single pass”, while the usual Zebra variation needs 2 guesses, each with 2 possibilities, so may need 4 trys to solve the puzzle.

Here’s a small MUMPS program that can solve the Zebra Puzzle, but not the Fish Puzzle – that is, can’t try guesses:

x X(1),X(4) ;X: logic puzzle solver
k B,D,BD s N=5,I="" f  s I=$o(X(2,I)) q:I=""  f J=1:1:N f K=1:1:N s V=$p($p(X(2,I)," ;"),",",J),B(K,I,V)="",BD(I,V)="",D(V)=I ;X(1): init B()
n (X,B,D,BD,N,XWR,C) s (C0,S)="" f  s S=$o(X(3,S)) x X(5),XWR q:S=""&(C=C0)!'C  s:S="" C0=C i S]"" s VV=$p(X(3,S)," ;") x X(4,$p(VV,",")) ;X(4): solve
s V=$p(VV,",",2),I=D(V),K=$p(VV,",",3) zt:'$d(B(K,I,V))  k B(K,I) s B(K,I,V)="" ;X(4,1): is in #
s V1=$p(VV,",",2),I1=D(V1),V2=$p(VV,",",3),I2=D(V2),KI=$p(VV,",",4) f K=1:1:N k:'$d(B(K-KI,I1,V1)) B(K,I2,V2) k:'$d(B(K+KI,I2,V2)) B(K,I1,V1) ;X(4,"="): is/is to left or right of
s V1=$p(VV,",",2),I1=D(V1),V2=$p(VV,",",3),I2=D(V2),KI=$p(VV,",",4) f K=1:1:N k:'$d(B(K-KI,I1,V1))&'$d(B(K+KI,I1,V1)) B(K,I2,V2) k:'$d(B(K-KI,I2,V2))&'$d(B(K+KI,I2,V2)) B(K,I1,V1) ;X(4,"+-="): is next to
n (X,B,N,C) x X(5,1),X(5,2) s C=V]"" x X(5,3) i C s Q="B" f C=0:1 s Q=$q(@Q) q:Q="" ;X(5): check for uniques & clean
s (K,I,V)="" f  s K=$o(B(K)) q:'K  f  s I=$o(B(K,I)) q:I=""  f  s V=$o(B(K,I,V)) q:V=""  s (C1,B1(I,V))=$g(B1(I,V))+1,B1(I,V,K)="",B2(I,V)=K k:C1>1 B2(I,V) ;X(5,1)
s I="",V=1 f K=1:1:N q:V=""  f  s I=$o(X(2,I)) q:I=""  s V=$o(B(K,I,"")) q:V=""  i $o(B(K,I,V))="" k B2(I,V) f KK=1:1:N k:KK-K B(KK,I,V) ;X(5,2)
s (K,V)="" f  s I=$o(B2(I)) q:I=""  f  s V=$o(B2(I,V)) q:V=""  s K=B2(I,V) k B(K,I) s B(K,I,V)="" ;X(5,3)

Here’s the data that defines the Zebra Puzzle:

Blue,Green,Red,White,Yellow ;X(2,"Color")
Brit,Dane,German,Norwegian,Swede ;X(2,"Nationality")
Beer,Coffee,Milk,Tea,Water ;X(2,"Beverage")
Blue Master,Dunhill,Pall Mall,Prince,Blend ;X(2,"Cigar brand")
Cats,Birds,Dogs,Fish,Horses ;X(2,"Pet")

=,Brit,Red ;X(3,1): The Brit lives in a red house.
=,Swede,Dogs ;X(3,2)
=,Dane,Tea ;X(3,3)
=,Green,White,1 ;X(3,4): The green house is on the left of the white, next to it.
=,Green,Coffee ;X(3,5)
=,Pall Mall,Birds ;X(3,6)
=,Yellow,Dunhill ;X(3,7)
1,Milk,3 ;X(3,8)
1,Norwegian,1 ;X(3,9): The Norwegian lives in the first house
+-=,Blend,Cats,1 ;X(3,10): The man who smokes Blends lives next to the one who keeps cats
+-=,Horses,Dunhill,1 ;X(3,11)
=,Blue Master,Beer ;X(3,12)
=,German,Prince ;X(3,13)
+-=,Norwegian,Blue,1  ;X(3,14)
+-=,Blend,Water,1 ;X(3,15)

This program can only process 5 types of rules, using 3 subroutines, but it only need 4 of them for the Zebra Puzzle. Sherlock would need a few more kinds of rule processors. Many logic puzzles would need even more. I’m unaware of any comprehensive review and catalog of these rule types, or an estimate or precise proof of how many there are.