Alright, figuring that "lists of things" are popular, I thought it'd be a good idea
to create a list of the most challenging logic/number puzzles in existence

Obviously one (or two?) of them will be a Calcudoku puzzle, but what will be
the others?

I'm looking for specific puzzles, so not "Kakuro is pretty hard", but a pointer
to a specific puzzle.

I'll probably first create a list and then ask for feedback, but of course suggestions
are welcome :)

Patrick

Candidates I have so far:

puzzle "b4m10" from the "Calcudoku, 101 Advanced Puzzles" book
(everyone seems to agree that that is the hardest puzzle in the book)

the Eternity II puzzle
(would you believe it, I have a copy, just no time to work on it )

This sudoku puzzle by Arto Inkala is a bona fide toughie!

http://www.telegraph.co.uk/science/scie ... ck-it.html

Right, I remember that one, thanks for the pointer

I'm now doing some stats on the site's solved puzzles,
to find the most difficult "Tuesday 9x9", and the most difficult Killer Sudoku

I managed this Thursday's 12x12 but found it exceptionally hard – harder than the higher-rated 15x15s. Of course I may have missed a key trick somewhere

The (rather crude) stat that I'm using is the percentage of solved puzzles when compared to
the number of solved 4x4 "easy" puzzles.

I'll check the same stat for some of the other large puzzles as well (10x10, 12x12, 8x8 with mod, etc.)

There are plenty of logic/number puzzles out there, but surprisingly little info
about rating the difficulty of one.

For example, for "minesweeper" (a very well known puzzle game) it turns out
to be hard to find (1) puzzles that require only logic to solve, and (2) rated puzzles (?)

(only thing I found was this "number of clicks required"
http://www.minesweeper.info/wiki/3BV)

I am glad for Bram's comment on this Thursday's 12x12. I had imagined that the unusual amount of difficulty I was having with it was just me. It is impressive that 86 of us have solved it thus far. I like Patrick's idea of dividing numbers of solvers of the hardest and easiest puzzles. I'm curious what the findings will be.

Here are the numbers (date and percentage when compared to the number
of solved 4x4 easy puzzles) for the top 25 12x12 puzzles, and yes, Thursday's 12x12 is up there...

(the highest percentages are 30-35%, for example, 2013-02-10: 33%, 2013-02-17: 32%)

The top one should be discarded because then there were 2400+ 4x4 easy solvers (this was
at the time of the "browser vs numerical intelligence" article).
The 3rd should also be discarded because it was a subscribers-only puzzle.

I'm not sure if the 2011-03-27 one is really very difficult, or if there is another reason why so few people solved it.

(Edit: table removed and recomputed, see later post.)