The New Sudoku Players' Forum

[ronk]

in that list, puzzle 7417 is said to have nearly no chance to give a multi-fish pattern.

In fact, my solver finds no floor group of size 4 or 5 leading to eliminations (danny should have the same results in template search)
So my solver did not try to build a multi-fish in that case

It will be very very interesting to see ronk's findings for that puzzle.

To all, I apologize for wasting your time. My so-called #7417 with a multi-fish is not the same as that in the "03 NN ..." file. I must have copied the wrong puzzle or title.

champagne, I thought your solver finding k-templates that lead to at least one exclusion was one of the criteria for listing the puzzle in "03 NN ..." file in the first place. Had I known it wasn't, I wouldn't have spent so much time on some of these puzzles.

daj95376, if you have the time and interest, would you please report your template findings on #7417

Code: Select all

1..4..7....7.8...6.9.......2..3......4...71....5.4..2..3.9..4......6...8..2....5. ;7417;elev;H193;10.80;10.80;10.80;592

[champagne]

champagne, I thought your solver finding k-templates that lead to at least one exclusion was one of the criteria for listing the puzzle in "03 NN ..." file in the first place. Had I known it wasn't, I wouldn't have spent so much time on some of these puzzles.

As explained with more details in the new hardest thread, that file is a short list established long ago of puzzles that could claim to be the hardest after elimination of puzzles having an exotic pattern.

The list published in the hardest thread is sorted in decreasing order of chances to find a new pattern.

It was clearly indicated that in my opinion, the top part of the list should show no exotic pattern.

[ronk]

The list published in the hardest thread is sorted in decreasing order of chances to find a new pattern.

It was clearly indicated that in my opinion, the top part of the list should show no exotic pattern.

Of the 27 puzzles, we found 11 0-rank multi-fish and 5 exocets or almost sk-loops (1 with no elims)***, leaving only 11 without an exotic pattern. That makes me feel a lot better.

What criteria did you use to rate the "chances to find a new pattern", as you phrased it?

[edit: *** 1 of the 5 posted here.]

[daj95376]

daj95376, if you have the time and interest, would you please report your template findings on #7417

Code: Select all

1..4..7....7.8...6.9.......2..3......4...71....5.4..2..3.9..4......6...8..2....5. ;7417;elev;H193;10.80;10.80;10.80;592

My template solver used several 5-templates to solve the puzzle. There weren't any significant number of eliminations in any of them.

My template analyzer only searches for 2/3/4-template eliminations. Here's its results.

Hidden Text: Show

Code: Select all

1..4..7....7.8...6.9.......2..3......4...71....5.4..2..3.9..4......6...8..2....5.

 # 7417;elev;H193;10.80;10.80;10.80;592

 +-----------------------------------------------------------------------+
 |  1      2568   368    |  4      2359   23569  |  7      389    2359   |
 |  345    25     7      |  125    8      12359  |  2359   1349   6      |
 |  34568  9      3468   |  12567  12357  12356  |  2358   1348   12345  |
 |-----------------------+-----------------------+-----------------------|
 |  2      1678   1689   |  3      159    15689  |  5689   46789  4579   |
 |  3689   4      3689   |  2568   259    7      |  1      3689   359    |
 |  36789  1678   5      |  168    4      1689   |  3689   2      379    |
 |-----------------------+-----------------------+-----------------------|
 |  5678   3      168    |  9      1257   1258   |  4      167    127    |
 |  4579   157    149    |  1257   6      12345  |  239    1379   8      |
 |  46789  1678   2      |  178    137    1348   |  369    5      1379   |
 +-----------------------------------------------------------------------+
 # 169 eliminations remain

 N:          1  2  3 4  5  6  7  8  9
 Templates: 56 14 56 5 50 32 25 44 50

 <2568>        accepted = 146 template combinations   -- no   UR detection
 <2568>        accepted = 142 template combinations   -- with UR detection

 <2568>   <>2  r2c6
 <2568>   <>6  r4c3,r6c1

 <2568>        r12c2,r5c4   locked for candidates


Only "accepted template combinations" changed between with/without UR detection. I would have marked any other differences.

[daj95376]

Crossing my fingers!

Code: Select all

.......8....1.9..66...2...4.7...8.9.5..........4.3...73.2.....5..5.6.3...1.....7.

 #  3218;elev;997;11.20;1.20;1.20;876


Hidden Text: Show

Code: Select all

 +--------------------------------------------------------------------------------+
 |  1249    23459   1379    |  34567   457     34567   |  12579   8       1239    |
 |  248     23458   378     |  1       4578    9       |  257     235     6       |
 |  6       3589    13789   |  3578    2       357     |  1579    135     4       |
 |--------------------------+--------------------------+--------------------------|
 |  12      7       136     |  2456    145     8       |  12456   9       123     |
 |  5       23689   13689   |  24679   1479    12467   |  12468   12346   1238    |
 |  1289    2689    4       |  2569    3       1256    |  12568   1256    7       |
 |--------------------------+--------------------------+--------------------------|
 |  3       4689    2       |  4789    14789   147     |  14689   146     5       |
 |  7       489     5       |  2489    6       124     |  3       124     1289    |
 |  489     1       689     |  234589  4589    2345    |  24689   7       289     |
 +--------------------------------------------------------------------------------+
 # 173 eliminations remain

 1 R35678 \ c68b6 + r3c37,r5c35,r6c1,r7c57,r8c9   (2 in cells)
 7 R35 7  \ c46   + r3c37,r5c35,r6c1,r7c57,r8c9   (1 in cells)
 8 R35678 \ c24b6 + r3c37,r5c35,r6c1,r7c57,r8c9   (2 in cells)
 9 R35678 \ c24   + r3c37,r5c35,r6c1,r7c57,r8c9   (3 in cells)


[ronk]

Crossing my fingers!

Looks good to me, congratulations! While the #3218 puzzle is fresh in your mind, try for the version with the base sets in columns too.

[daj95376]

While the #3218 puzzle is fresh in your mind, try for the version with the base sets in columns too.

Very similar results.

Hidden Text: Show

Code: Select all

 1 C13579 \ r14b6 + r3c37,r5c35,r6c1,r7c57,r8c9   (2)
 7 C 357  \ r12   + r3c37,r5c35,r6c1,r7c57,r8c9   (1)
 8 C13579 \ r29b6 + r3c37,r5c35,r6c1,r7c57,r8c9   (2)
 9 C13579 \ r19   + r3c37,r5c35,r6c1,r7c57,r8c9   (3)


[David P Bird]

I've now worked through all these puzzles and revised my results <here>.

It would be interesting to see how the potential elimination counts for the 13 Almost Sharks I found compare with template results.

Puzzle 889 would look to be the most fruitful and 3411 the least fruitful.

[champagne]

It would be interesting to see how the potential elimination counts for the 13 Almost Sharks I found compare with template results.

David,

What are the eliminations you deduct from your Almost sharks (I did not see them)

It is easy to cross check with the potential of the corresponding floors

[David P Bird]

David, What are the eliminations you deduct from your Almost sharks (I did not see them)

It is easy to cross check with the potential of the corresponding floors

I'm sorry I haven't recorded them. I thought the counts would be enough and that it would be relatively easy to get the same PEs by using truth and link sets.

Here's an example using
1.......9..67...2..8....5.......83.....2...64..7.4..1...462....5..8......9...3...;889;elev;L59;11.30;11.30;9.50;607

Code: Select all

                   v         v       v                         v       v
 *-------------------------*-------------------------*-------------------------*
 | 1       247-35  -       | -       -       246-5   | 467-8   -       -       |
>| -       -       6       | 7       13589   -       | -       2       138     |<
 | 247-39  8       -       | -       -       246-19  | 5       -       -       |
 *-------------------------*-------------------------*-------------------------*
 | 246-9   246-15  -       | -       -       8       | 3       -       -       |
>| -       -       13589   | 2       1359-7  -       | -       6       4       |<
>| -       -       7       | 359     4       -       | -       1       58-2    |<
 *-------------------------*-------------------------*-------------------------*
>| -       -       4       | 6       2       -       | -       3589-7  1358-7  |<
 | 5       267-13  -       | -       -       47-19   | 2467-19 -       -       |
 | 267-8   9       -       | -       -       3       | 2467-18 -       -       |
 *-------------------------*-------------------------*-------------------------*
                   ^         ^       ^                         ^       ^

(2467)Shark:r2567,c34589 Potential Eliminations 24 Candidates 16 Cells
Home Min Truths = 11, Away Max Truths = 12
r5c5,r6c9,r7c89 <> 2674
r1c26,r3c16,r4c12,r8c267,r9c17 <> 13589

Briefly, as I don't think you have been able to follow my explanation:
In the solution grid the Home cells (covered twice) must hold the same number of instances of each digit as the Away cells (completely uncovered).
Making all the eliminations shown would reduce the Home cells to 11 truths for the digit set and force the Away Cells to hold 12 truths. So just one of these eliminations must be invalid.

I hope this helps you to see the relationship between the two approaches. I admit that I'm not sure how you would get the same PE list from a XSudo approach when there is one less truth than there should be, but it must be possible.

DPB

[daj95376]

It would be interesting to see how the potential elimination counts for the 13 Almost Sharks I found compare with template results.

Here's what I have for template results.

Hidden Text: Show

Code: Select all

Almost Sharks:

   35* (1234)AShark:r4689,c25689,b 18 Candidates 13 Cells

 <1234>   <>1  r2c2,r8c8
 <1234>   <>2  r6c2,r9c9
 <1234>   <>3  r9c2
 <1234>   <>4  r2c5
 <1234>   <>6  r1c7
 <1234>   <>7  r7c12
 <1234>   <>8  r2c4,r7c1

  243* (2345)AShark:r1356,c24678   18 Candidates 14 Cells

 <2345>   <>2  r6c7
 <2345>   <>3  r1c8
 <2345>   <>4  r5c2

  620* (1248)AShark:r4589,c23479   20 Candidates 15 Cells

 <1248>   <>4  r89c9
 <1248>   <>5  r2c1
 <1248>   <>7  r2c1

  889* (2674)AShark:r2567,c34589   24 Candidates 16 Cells

*** exceeds a 4-template

 1102* (1567)AShark:r12579,c1268   23 Candidates 15 Cells

 <1567>   <>7  r2c6

 1362* (2478)AShark:r3578,c1678    21 Candidates 14 Cells

*** exceeds a 4-template

 2051* (1346)AShark:r2379,c1278,b5 16 Candidates 12 Cells

 <1346>   <>1  r4c8
 <1346>   <>4  r7c2
 <1346>   <>5  r8c6
 <1346>   <>6  r9c1
 <1346>   <>7  r8c6
 <1346>   <>8  r4c9
 <1346>   <>9  r6c3

 2780* (2457)AShark:r23568,c1579   20 Candidates 15 Cells

*** exceeds a 4-template

 4024* (3459)AShark:r1256,c2346,b9 13 Candidates 10 Cells

 <3459>   <>4  r5c3,r6c2
 <3459>   <>8  r3c8

 4993* (1578)AShark:r23467,c1368   22 Candidates 14 Cells

 <1578>   <>1  r2c16
 <1578>   <>3  r1c57,r6c5
 <1578>   <>4  r6c25,r8c27,r9c9
 <1578>   <>5  r2c1,r3c3
 <1578>   <>7  r6c6,r7c7
 <1578>   <>8  r4c7,r256c8,r4c9
 <1578>   <>9  r6c2,r8c27

10216* (45789)AShark:r1247,c1247   18 Candidates 14 Cells

*** exceeds a 4-template

12120* (1236)AShark:r25689,c4589   19 Candidates 14 Cells

 <1236>   <>1  r2c28,r8c8
 <1236>   <>2  r2c24,r5c9,r7c3
 <1236>   <>3  r2c28,r6c8
 <1236>   <>4  r1c7,r2c7,r3c12
 <1236>   <>5  r12c7
 <1236>   <>6  r9c9
 <1236>   <>7  r4c2

14596* (1236)AShark:r25689,c4589   15 Candidates 10 Cells

 <1236>   <>1  r2c28,r3c8,r7c1
 <1236>   <>2  r2c24,r7c3
 <1236>   <>3  r2c29
 <1236>   <>4  r1c7,r2c7,r3c12
 <1236>   <>5  r12c7
 <1236>   <>7  r4c2


[Edit: corrected entry 1210 -> 12120 and added template counts.]

[David P Bird]

daj Interesting, your results seem to prove the opposite to what I was expecting – the fewer the PEs from an Almost Shark, the more fruitful a template analysis is likely to be!

[ronk]

David P Bird, as before, I believe this shark discussion belongs in your shark thread, so don't be surprised to find it moved.

[champagne]

my computer is busy on another test I launched this morning, so I have to wait for more experimentation, but for me danny's findings are in line with the list I published.

Danny, there is a typo in david's list you should read 12120 as puzzle number instead of 1210.

David, I have a question after Danny's confirmation for puzzle 889.

The "template" and the "permutation" process show no valid elimination for floors 2467. (I have to confirm that on my side, but the recorded result says it is anyway less than 5 eliminations)

You find if I am right 24 eliminations
If this correct (and logic), it is a true breakthrough.

so do you really state that the almost shark logic says we can clear 24 candidates

[David P Bird]

David P Bird, as before, I believe this shark discussion belongs in your shark thread, so don't be surprised to find it moved.

Your prejudices are showing again. The topic is exotic patterns not Xsudo truth and link sets or Sharks. How come you can coach daj <here> in this thread, and object to my post – which is not a discussion!
What I'm attempting to do is to see if there's anything else we can ring out of the results whatever tool is used.

As champagne is the owner of the thread, I will comply with his wishes if he asks me to delete that post.

Ironically, by pressing this issue you now added more off-topic posts to this thread than before.