Advanced methods and approaches for solving Sudoku puzzles

logic only proof

Postby Guest » Sat Apr 09, 2005 1:47 am

The 'proof' that they can all be solved by logic only, is simply that they (the pappacom/times puzzles) are generated step by step by a program that does not use any trial and error methods. The program ensures that each step is a logical deduction from the previous clues only, and also does a "reasonable" job of grading them.

It also makes certain that they are symetrical and have only one soloution, both requirements of form for the traditional Japanese puzzle.

Hope that answers some questions, more is available fromt the times online pages,


logic vs. trialling

Postby Guest » Sun Apr 24, 2005 11:58 pm

sorry, i'm going to chime in on the argument thread nearer the top of this forum.

I have a theory that the "route" you take to solving each grid actually makes a difference in whether you manage to solve the bloody thing, or how fast it takes to do it.

I do the torygraph puzzles, and one day i bought my paper in the morning, did half a moderately rated puzzle before getting hopelessly stuck, then visited my dad who also gets the telegraph. i started on his copy afresh, and found that by taking another route at random, i not only completed the grid but quite quickly as well.

i have immersed myself in these things since they arrived opposite the obituaries, so i can assure you quite honestly that any memory of which number went where had nothing to do with it - all entries were logical.

perhaps then, there are parallels with chess stalemates? and also, perhaps all grids have at least one purely logical solution, with others requiring trialling methods.

ps. all sudokus have the unexpected side effect of making the cryptic crosswords far less scary.

Postby Hammerite » Mon Jun 20, 2005 10:48 pm

lianne, I'm fairly sure your theory must be wrong... As adding a symbol to the grid only adds to the information available in the grid, the information available with which to solve the puzzle must increase with "symbols added". All deductions usable at some time to find the value in a given cell are therefore still available at any later time. Suppose that there exists an "easy" set of deductions allowing solution of the Su Doku, and a "hard" set of deductions that eventually runs dry, or allows solution of the Su Doku but is too complicated for me to think of. Suppose that I start off along the "hard" route and then get stuck. That doesn't prevent me from starting along the "easy" route (indeed, doing so at any time while on the "hard" route) and finishing the puzzle easily, since all the information needed for the "easy" route (and possibly more) is still there.
Posts: 44
Joined: 20 June 2005


Return to Advanced solving techniques