The New Sudoku Players' Forum

[ArkieTech]

Code: Select all

 *-----------*
 |.5.|...|.8.|
 |..2|47.|...|
 |...|..1|..6|
 |---+---+---|
 |..1|...|.9.|
 |4..|792|..3|
 |.7.|...|54.|
 |---+---+---|
 |8..|3..|...|
 |...|.24|8..|
 |.3.|...|.7.|
 *-----------*


Play/Print this puzzle online

[pjb]

Code: Select all

 1379    5       3479   | 2      36    b369    | 13479  8      1479   
 1369    169     2      | 4      7      8      | 139    15     159   
 379     49      8      |a59     35     1      | 3479   2      6     
------------------------+----------------------+---------------------
 236     26      1      | 68-5   4     d356    | 27     9      278   
 4       8       5      | 7      9      2      | 16     16     3     
 2369    7       369    | 168    1368   36     | 5      4      28     
------------------------+----------------------+---------------------
 8       12469   469    | 3      156    7      | 12469  156    12459 
 15679   169     679    | 1569   2      4      | 8      3      159   
 12569   3       469    | 15689  1568  c569    | 12469  7      12459 


(5=9)r3c4 - r1c6 = (9-5)r9c6 = r4c6 => -5 r4c4;lclste

Phil

[bat999]

Code: Select all

.----------------------.---------------------.---------------------.
|  1379    5      3479 |  2      36    c369  |  13479  8     1479  |
|  1369    169    2    |  4      7      8    |  139    15    159   |
|  379     49     8    | b59     35     1    |  3479   2     6     |
:----------------------+---------------------+---------------------:
| a236    a26     1    | a568    4      56-3 | a27     9    a278   |
|  4       8      5    |  7      9      2    |  16     16    3     |
|  2369    7      369  |  168    1368  c36   |  5      4     28    |
:----------------------+---------------------+---------------------:
|  8       12469  469  |  3      156    7    |  12469  156   12459 |
|  15679   169    679  |  1569   2      4    |  8      3     159   |
|  12569   3      469  |  15689  1568   569  |  12469  7     12459 |
'----------------------'---------------------'---------------------'

(3=5)r4c12479 - (5=9)r3c4 - (9=3)r16c6 => -3 r4c6; lclste

[Ngisa]

Code: Select all

+------------------+----------------+-----------------+
| b1379  5     b3479 | 2     a36   a369 | 13479 8   1479  |
| c1369  169   2    | 4     7    8   | 139   15  159   |
| c379   49    8    | 59    35   1   | 3479  2   6     |
+------------------+----------------+-----------------+
| d236   26    1    | 568   4    e356 | 27    9   278   |
| 4     8     5    | 7     9    2   | 16    16  3     |
| 2369  7     369  | 168   1368 36  | 5     4   28    |
+------------------+----------------+-----------------+
| 8     12469 469  | 3     156  7   | 12469 156 12459 |
| 15679 169   679  | 1569  2    4   | 8     3   159   |
| 12569 3     469  | 15689 1568 f56-9 | 12469 7   12459 |
+------------------+----------------+-----------------+


(9=3)r1c56 - r1c13 = r2c3 - r4c1 = (3-5)r4c6 = (5)r9c6 => -9 r9c6; stte

Clement

[Sudtyro2]

(9=3)r1c56 - r1c13 = r2c3 - r4c1 = (3-5)r4c6 = (5)r9c6 => -9 r9c6; stte

Hi Clement,
I think your elimination is a good one, but it may be an "lclste" rather than "stte" solution.
The A/S Solver, at least, needs one "Pointing Pairs" step beyond "Singles" to complete the puzzle.

SteveC

[Ngisa]

(9=3)r1c56 - r1c13 = r23c1 - r4c1 = (3-5)r4c6 = (5)r9c6 => -9 r9c6; stte

Hi Clement,
I think your elimination is a good one, but it may be an "lclste" rather than "stte" solution.
The A/S Solver, at least, needs one "Pointing Pairs" step beyond "Singles" to complete the puzzle.

SteveC

I have checked it again, it is stte. Try it also.
Step three in my chain should be r23c1.

[bat999]

I have checked it again...

Check it again.
I agree with Sudtyro2,

[pjb]

I agree that the solution is lclste. BTW I wrote a simple routine that checks if the elimination of a candidate can lead to a stte solution, and lists all such candidates. In this case there is not a single one, so any single digit elimination cannot be stte. Of course, a solution that eliminates multiple digits such as a loop can produce a stte solution in this case.
Phil

[SteveG48]

I agree that the solution is lclste. BTW I wrote a simple routine that checks if the elimination of a candidate can lead to a stte solution, and lists all such candidates. In this case there is not a single one, so any single digit elimination cannot be stte. Of course, a solution that eliminates multiple digits such as a loop can produce a stte solution in this case.
Phil

Care to make that routine available?

[Leren]

SteveG48 wrote : Care to make that routine available?

I was also intrigued by Phil's statement and started wondering how he did it - remember he says it is simple.

One simple way to do this is to remove each candidate (one at a time) from the puzzle end-of-basics status.

You can determine the end-of-basics status when you need to apply any non-basic move to make a further elimination, or if the puzzle solves with stte from the point where the last basic non-single move is used.

If there is a failure (non-solution) or any non-single move is used, the candidate is not added to the stte list, put back into the puzzle status and the next candidate is tried.

If an stte solution is reached the candidate is added to the list. Is it that simple Phil ?

Leren

[pjb]

Steve: This is what I have. It is javascript. Maybe not so simple as it uses a few functions along the way. Leren, more or less spot on.

Code: Select all

s="";
for(i=1; i<10; i++) {
    for(w=0; w<9; w++) {
   for(x=0; x<9; x++) {
       if(array0[w][x][i] == 1 && length1(w, x) > 1) {     // if cell contains i and has more than one candidate
      hit = false;
      remove(i, w, x);  //removes i from w.x
      do {
          size1 = getSize();   // total number of digits in grid
          Results1("second");  // runs naked singles
          Results2("second");  // runs hidden singles
          size2 = getSize();
      } while(size2 < size1);
      hit = solved3();  // checks that grid is reduced to 81 and is valid solution
      fix_grid(temp);   // restores grid to starting state
      if(hit) {
          sx = sx+i+" at r"+(w+1)+"c"+(x+1)+";" ;  // lists results
      }
       }
   }
    }
}


Phil

[Leren]

I didn't mention it in my previous post, but if you remove a candidate and the solution is not valid, then the candidate is True in the solution (because we always trust the puzzle makers on this site ! )

If the cell contains more than 2 candidates then you could use the True candidate as the start and end of a discontinuous loop, solving the cell and possibly finding another stte solution.

It's not clear from your Javascript if you include this possibility. In any event, what you are doing is a survey, not some amazing stte detection function - that would really be amazing.

BTW I actually tried assuming all the True candidates in multi-candidate cells to be False in this puzzle to see if there was an stte solution in this area but had no luck.

Leren

[SteveG48]

Thanks, Phil. Thanks, Leren.