The New Sudoku Players' Forum

[Pat]

this is not new

yes, in his earlier Topic i already gave the pointer to our old discussion

[Glyn]

Just a comment on Pat's post

Sudoku Explainer only works on proper SuDoku puzzles -- puzzles having exactly one answer.
if no answer, no problem, it will tell you.
but if more than one answer, big problem, unpredictable results!

Actually if you are running it in graphical mode as a helper and the message "Sudoku has multiple solutions" appears it is still possible to get hints by clicking on "Get next hint". This is how I was able to rate the ill-posed grid. This is not a feature which the casual user should need. SE will only give up without trying if there are less than 17 givens.

[Pat]

Sudoku Explainer only works on proper SuDoku puzzles -- puzzles having exactly one answer.
if no answer, no problem, it will tell you.
but if more than one answer, big problem, unpredictable results!

Actually if you are running it in graphical mode as a helper and the message "Sudoku has multiple solutions" appears it is still possible to get hints by clicking on "Get next hint".

as far as i know, it assumes the puzzle has a single answer -- and uses Uniqueness-Of-Answer in solving the puzzle -- hence my warning about unpredictable results

the Uniqueness-Of-Answer assumption
will sometimes lead it to report an answer
(unaware of the existence of more);
on other occasions,
will lead it to report "no answer"

in the present example,
none of those disasters occurred -- just plain luck

~ Pat

[daveisme]

Thanks, all.
I guess, Soduku Explainer is not ready to handle "multiple solution" directly yet.
Grading them or solving them.
I will stay away from subundle tough-tough puzzles then.

BTW, "find something new" does not mean "that is new".
Multiple-solution grids, as well as no-solution grids are probably older than sudoku grids.
I can only imaging that the first ever sudoku grid was created in weeks or months, not days.
How the game was presented 29 years ago on Dell was new.
How the subundle is presented now, is new, at least to me.
I like that, and simply hope i can find people to share info/help with.
As for the SE, maybe I will try it seriously later.
Again, thanks for all the helps.

[Carcul]

Code: Select all

 *------------------------------------------------------------------*
 | 2689   69     1689   | 3      248    5      | 7      19     46   |
 | 4      3      59     | 67     67     1      | 2      59     8    |
 | 25678  567    1568   | 9      248    28     | 15     3      46   |
 |----------------------+----------------------+--------------------|
 | 579    8      3      | 4      279    27     | 6      1257   1579 |
 | 1      4      69     | 2678   5      2678   | 3      278    79   |
 | 5679   5679   2      | 1      6789   3      | 4      578    579  |
 |----------------------+----------------------+--------------------|
 | 5689   1      5689   | 5678   3      678    | 59     4      2    |
 | 589    2      7      | 58     18     4      | 159    6      3    |
 | 3      56     4      | 256    126    9      | 8      157    157  |
 *------------------------------------------------------------------*


Very nice puzzle for finding nice loops:

1) [r7c3]-9-[r7c7]-5-[r3c7]=5=[r2c8]=9=[r2c3]-9-[r7c3], => r7c3<>9.

2) [r2c8]=9=[r2c3]-9-[r1c2]-6-[r9c2]-5-[r9c89]=5=[r78c7]-5-[r3c7]=5=
=[r2c8], => r1c3<>9; r3c2, r6c2<>6; r9c4<>5.

3) [r6c5]=6=[r6c1]-6-[r5c3]-9-[r2c3]=9=[r2c8]=5=[r3c7]=1=[r8c7]-1-
-[r8c5]-8-[r6c5], => r6c5<>8.

4) [r9c8]=7=[r9c9]-7-[r5c9]-9-[r5c3]=9=[r2c3]=5=[r2c8]-5-[r9c8], => r9c8<>5.

5) [r4c9]=1=[r9c9]=5=[r9c2]=6=[r1c2]=9=[r2c3]-9-[r5c3]=9=[r5c9]-9-
-[r4c9], => r4c9<>9.

6) [r9c2]=6=[r9c4]-6-[r2c4]=6=[r2c5]-6-[r6c5]=6=[r6c1]-6-[r5c3]-9-
-[r2c3]=9=[r1c2]=6=[r9c2], => r9c2=6 and the puzzle is solved.