The New Sudoku Players' Forum

[fathom]

Hi
Sorry I arrive late at this game.
However, could someone describe the technique I am using.
Seems failproof, no guess work even on the difficult ones (does take a lot a fun out of solving theses).

Here's an attempt at explaining what I am doing:

(1)I divide each little square into 9 smaller square- similar to your phone pad...
(2)If a number , eg 5, is present, I check all the 5's in that column, row and 3X3 area.
(3)Do that for all numbers given.

Most of the time, it's then just fill in the blank, and it's done.

On the harder ones,
(4)I look at each tiny column/ row I created on the phone pad, and usually there is one number in the phone pad that is obvious, because all the number in that tiny row or column are already checked, and repeat the procedures (1) to (2) all over again.
(5) Fill in the blank on the phone pad and solve each phone pad row/ column individually.

Eg:

I can see one thing you guess about and that is the cell you chose.

You normally (or atleast I) look at a specific number, row, column or box, but never on one particular cell.

I can't believe I'm reading something like this. This really looks like a denial phase.

I'm not sure on this yet, but I'm begining to believe that for each 'Trial and Error'-case there is a 'Foreched chain'-case.

This isn't so. Up to now, I've compiled about 18,000 problems. About 12,000 can be solved using common techniques. About 2,000 can be solved using advanced coloring (which includes swordfish and foreched chains). About 4,000 my program still cannot solve without guessing (9 times in one case).

Note: I haven't yet implemented naked/hidden triples in my program so this changes a bit the figures above.

Should anyone be interested in the 9-guesses problem, here it is; I don't expect it to be solvable by logic.

Code: Select all

..1.9.8.5
.326.....
4........
.....5...
..3.1.6..
...4.....
........4
.....712.
8.9.6.3..


Once you have done step (1) and (2),
then r6,c6 can only be =6,
fill in slowly the phone pad for r6c6 and see what other number becomes obvious.
etc...

Or I'm I off somewhere?

[stuartn]

This looks like a nice graphic way of solving by exclusion - but I suspect it'll meet it's match with grids that require more advanced logic! - have a look at some on this site that need x-wings and swordfish to solve - and do let us know how it works out! (and after the 6 @R6C6 - what next?)

[Nick70]

Should anyone be interested in the 9-guesses problem, here it is; I don't expect it to be solvable by logic.

Code: Select all

..1.9.8.5
.326.....
4........
.....5...
..3.1.6..
...4.....
........4
.....712.
8.9.6.3..


That was a long time ago. Now I know it can be solved with forcing chains.

[fathom]

Thanks,
will look into some other one and report back.

[fathom]

By the way, could you give me a new example of an extreme one to solve, ideally having to guess, to see whether my technique still works.
Thanks

[fathom]

I was given this link before the site crash.
http://www.paulspages.co.uk/sudoku/sudokugallery2.php?r=outlaw
I could not do the first one in its entirety without more involved techniques.
However, for everything else published (ie not outlawed) so far that I tried, it seems to work fine.
Warning, very tedious, mindless and not enjoyable.

[fathom]

I found someone who uses the same technique (he explains it a lot better)
http://www.sudoku.org.uk/discus/messages/2/71.html?1121515983

The only difference, is where he is stuck and has to start guessing.
I just go from each row/ column/ 3x3, in each dial pad from 1 to 9, see if that is the only one not marked, if so, then that's the number. And so on. No guessing (except for outlawed as discussed earlier). No fun, I can push paper at work...

I'm back to doing them without marking.
Thanks for all your feedbacks.

[Cec]

fathom wrote: ".....Here's an attempt at explaining what I am doing:

(1) I divide each little square into 9 smaller square - similar to your phone pad..."

Hi fathom - I would appreciate more help The puzzle grid contains nine 3X3 boxes with each box containing nine cells. In your above explanation could you please explain what a "little square" means?

Regards Bonsai Cec

[fathom]

123
456
789

would be one cell of the 3x3
I took the shortcut of putting an x instead of numbers.
eg, a '6' would look like this:

XXX
XX
XXX

on an easy puzzle, once you marked all the given numbers, usually there is at least one cell looking like above, and you just have to fill in the blank with a number.

Alternatively in this row, column 2 has to be '4':

_X_ |X_ X|_ XX||X_ _|XX_|_ XX||XXX|_X X |_ _X
X _X|_X_ |X_ X||XXX |X_ _|X_ X||X_ X|X_ _ | XXX
_X_ |X_ X|XXX ||_ _ _|_XX|XX_ ||_ X_ | _ _X|XXX

Hope that helps

[stuartn]

Blimey - all looks a bit latin to me.

[Karyobin]

I'm obviously misunderstanding, but why don't you just fill in all the possibilities and see which ones, etc...

What's all this 'x' business?

[fathom]

I guess it would come to the same thing.
When you eliminate a possibility down the road, you cross it out, presumably with an x, so I just start with x and leave the possibility blank.