What are the chances of a puzzle being minimal...

Everything about Sudoku that doesn't fit in one of the other sections

What are the chances of a puzzle being minimal...

Postby 999_Springs » Thu Jan 01, 2009 4:03 am

...if you are given the number of clues it has? Clearly it will decrease from 100% for a 17-clue puzzle to 0% for a 38-or-so-clue puzzle, but what happens in between?
Hidden Text: Show
Once upon a time I was a teenager who was active on here 2007-2011
Posts: 367
Joined: 27 January 2007
Location: In the toilet, flushing down springs, one by one.

Postby StrmCkr » Thu Jan 01, 2009 6:59 am

the curve is formed based on cover set logic.

Code: Select all
{18-39 could be minimal}
100% at 17 clues

       100% at 39 clues. {chance to find this is extremly slim}.
0% -81 clues


the perentage has variation depending
on the arrangment of all cover set logic.
the more sets covering the same logical forcing the same assertion = non-minimal.
Last edited by StrmCkr on Fri Jan 02, 2009 5:48 am, edited 3 times in total.
Some do, some teach, the rest look it up.
User avatar
Posts: 647
Joined: 05 September 2006

Postby coloin » Fri Jan 02, 2009 4:22 am

A lot of work was done here by JPF,Ocean and Red Ed



From this I take it that if you reduce a solution grid to

43 clues ~ 50 % will be valid, [none will be minimal]

Generating minimal puzzles from a full grid - is relativly easy to get 22-28 clue minimal puzzles. Its difficult to produce a minimal puzzle with >28 clues !

Generating puzzles from other minimal puzzles with the same number of clues will give both minimal and non-minimal puzzles.

from making 20s-19s-18s-17s

10% of 20s are non minimal
5% of 19s are non-minimal
1% of 18s are non-minimal

taking a 24 clue minimal puzzle and making many others - 75% of these were non-minimal

30 clues - 99.9 % non-minimal

StrmCkr wrote:100% at 39 clues. {chance to find this is slim}.

"slim" should read "extremely slim" [ approx 1 in 10^30 perhaps]

Posts: 1637
Joined: 05 May 2005

Return to General