The New Sudoku Players' Forum

by **mith** » Wed Aug 12, 2020 8:00 pm

This one came up on github, so figured I'd make it my first post in this subforum.

Code: Select all: +-------+-------+-------+ | . . . | . . . | . . . | | . . 9 | 8 . . | . . 7 | | . 8 . | . 6 . | . 5 . | +-------+-------+-------+ | . 5 . | . 4 . | . 3 . | | . . 7 | 9 . . | . . 2 | | . . . | . . . | . . . | +-------+-------+-------+ | . . 2 | 7 . . | . . 9 | | . 4 . | . 5 . | . 6 . | | 3 . . | . . 6 | 2 . . | +-------+-------+-------+ ...........98....7.8..6..5..5..4..3...79....2...........27....9.4..5..6.3....62..

[edit]And a companion piece, too similar to warrant a separate thread:

Code: Select all: +-------+-------+-------+ | . . . | . . . | . . . | | 1 . . | . . 2 | 3 . . | | . 4 . | . 5 . | . 6 . | +-------+-------+-------+ | . 6 . | . 7 . | . 1 . | | 2 . . | . . 3 | 8 . . | | . . . | . . . | . . 7 | +-------+-------+-------+ | . . 9 | 5 . . | . . . | | . 5 . | . 6 . | . 7 . | | 3 . . | . . 8 | 2 . . | +-------+-------+-------+ .........1....23...4..5..6..6..7..1.2....38..........7..95......5..6..7.3....82..

by **SpAce** » Wed Aug 12, 2020 11:54 pm

mith wrote:This one came up on github, so figured I'd make it my first post in this subforum.

A real treat for a Sith Lord 8-)

Thank you!

Code: Select all: \136 \18 \12 \138 \148 \18 .-------------------------.---------------------------.---------------------------. | 124567 27-136 3456-1 | 345-12 279-13 1234579 | 134689 29-148 346-18 | | 1456-2 *1236 9 | 8 *123 1345-2 | 1346 *124 7 | \2 | 127-4 8 *134 | *1234 6 1279-34 | 19-34 5 *134 | \34 :-------------------------+---------------------------+---------------------------: | 1289-6 5 *168 | *126 4 1278 | 1789-6 3 *168 | \6 | 1468 *136 7 | 9 *138 1358 | 14568 *148 2 | | 124689 29-136 346-18 | 356-12 27-138 123578 | 1456789 79-148 456-18 | :-------------------------+---------------------------+---------------------------: | 1568 *16 2 | 7 *138 1348 | 13458 *148 9 | | 79 4 *18 | *123 5 1289-3 | 178-3 6 *138 | \3 | 3 *79 *158 | *14 *189 6 | 2 *178-4 *1458 | \4579 '-------------------------'---------------------------'---------------------------'

Step 1. MSLS (24 cells) (found manually)

24x24 {2579N258 3489N349 \ 2r2 34r3 6r4 3r8 4579r9 136c2 18c3 12c4 138c5 148c8 18c9} => 40 elims

eliminations: Show

Code: Select all: .------------------.-------------------.-----------------. | 1456 27 3456 | 345 79 1345 | 8 29 346 | | 1456 *136 9 | 8 *123 145-3 | *346 124 7 | | 27 8 134 | 1234 6 79 | 19 5 134 | :------------------+-------------------+-----------------: | 29 5 168 | 16 4 27 | 79 3 18 | | 1468 *136 7 | 9 *138 158-3 | 456 18 2 | | 18 29 346 | 356 27 18 | 456 79 456 | :------------------+-------------------+-----------------: | 1568 16 2 | 7 *138 148-3 | *345 148 9 | | 79 4 18 | 123 5 29 | 17 6 138 | | 3 79 158 | 14 189 6 | 2 178 1458 | '------------------'-------------------'-----------------'

Step 2. Lightsaberfish: (3)c257\r257 => -3 r257c6; stte

Website · by **denis_berthier** » Thu Aug 13, 2020 3:40 am

mith wrote:This one came up on github, so figured I'd make it my first post in this subforum.

Code: Select all
+-------+-------+-------+ | . . . | . . . | . . . | | . . 9 | 8 . . | . . 7 | | . 8 . | . 6 . | . 5 . | +-------+-------+-------+ | . 5 . | . 4 . | . 3 . | | . . 7 | 9 . . | . . 2 | | . . . | . . . | . . . | +-------+-------+-------+ | . . 2 | 7 . . | . . 9 | | . 4 . | . 5 . | . 6 . | | 3 . . | . . 6 | 2 . . | +-------+-------+-------+ ...........98....7.8..6..5..5..4..3...79....2...........27....9.4..5..6.3....62..

A solution for normal people, using only elementary Subsets. The number of Subsets present in the puzzle is impressive.

Hidden Text: Show

Could you be more precise about the origin?

by **eleven** » Thu Aug 13, 2020 7:04 am

Impressing puzzle.
It can be solved with 7 swordfish (!) in different colors (5792681), some pairs and triples and an x-wing - before you get a single number !
And it looks very nice.

by **pjb** » Thu Aug 13, 2020 7:07 am

Repeated fish (7 SF, 1 finned SF, 1 finned XW, 1 XW) solve it.
I found 3 consecutive different MSLSs:

15 cell Truths: r129 c258 +r5c2 r7c2 r5c5 r7c5 r5c8 r7c8
15 links: 279r1, 2r2, 79r9, 136c2, 138c5, 148c8
18 eliminations
several basic moves

15 cell Truths: r3489 c1349
15 links: 34r3, 6r4, 3r8, 45r9, 279c1, 18c3, 12c4, 18c9
13 eliminations
several basic moves

16 cell Truths: r2579 c2578 +r6c7
16 links: 23r2, 3r5, 3r7, 79r9, 16c2, 18c5, 456c7, 148c8
3 eliminations => stte

Phil

by **SpAce** » Thu Aug 13, 2020 12:11 pm

no longer relevant for the most part: Show

denis_berthier wrote:A solution for normal people, using only elementary Subsets.

I'm curious. How do normal people justify classifying Finned X-Wings as 'elementary subsets'?

I'm wondering, because when they're not viewed as finned fishes (i.e. not subsets), they're either ALSs or AHSs which are almost-subsets (i.e. not subsets). Furthermore, they're Rank 1 patterns, which makes them fundamentally different and more complex than basic subsets and fishes that are Rank 0 (just like singles and intersection moves, which are the simplest kinds of subsets or fishes). For that reason I wouldn't count them as 'elementary' either. Definitely not 'subsets'.

Besides, somehow I would think that most normal people don't count even non-finned basic fishes as subsets (or elementary). Personally I'm somewhat ok with that classification because I'm fluent in any solving space and the related pattern transformations, but obviously I'm not normal. Even so, I don't call a fish a subset when it's in the fish form. That's just confusing, because the term has no obvious relation to the fish pattern. (It's more logical to see a subset as a generalized fish, and I do.) It's also misleading, because a fish is almost always harder to spot than the corresponding naked and hidden subsets, and far less likely to be spotted or even understood by casual solvers at all. Thus, not exactly elementary.

PS. As already mentioned by eleven, the puzzle can be solved with 'elementary subsets' (if non-finned basic fishes are somewhat illogically included in that category). If you turn off Finned X-Wings, yours would probably qualify too (added: not quite; see * below). It would shorten and simplify the solution anyway.

PPS. In general, it seems like a totally arbitrary decision to include Finned X-Wings as named patterns in your solve paths and exclude others of similar difficulty. Seems like a misfit to me, considering the whole. It's not nearly as weird as reporting intersections as whips, though.

PPPS. (*) Added. I almost missed this:

biv-chain[2]: c5n2{r6 r2} - r3n2{c4 c1} ==> r6c1 ≠ 2

How is that Kite an 'elementary subset'? Furthermore, it's related to my previous point. Why include Finned X-Wings (aka Grouped Skyscrapers) as named patterns but not 2-String Kites, for example? Both are single-digit patterns of the same size and Rank 1. As chains, they're both length 2. As fishes, they're both Finned (or 2x3) X-Wings. The only significant difference is that as fishes, one is basic and the other mutant. While the latter can be considered more complex, it's compensated by the fact that as a chain both of its strong links are simple (unless it's a Grouped Kite) while one of those of a Finned X-Wing is always grouped (unless it's Sashimi, i.e. a basic Skyscraper). That evens out any significant difference in spotting difficulty. Thus, they're much more similar than different.

--
Edit. Hid the text.

Website · by **denis_berthier** » Thu Aug 13, 2020 1:16 pm

I had missed the finned-fish and the bivalue-chain. Here's a pure Subsets solution. Still more of them.

Code: Select all: *********************************************************************************************** *** SudoRules 20.1.s based on CSP-Rules 2.1.s, config = S *** Using CLIPS 6.32-r770 *********************************************************************************************** 252 candidates, 1927 csp-links and 1927 links. Density = 6.09% hidden-pairs-in-a-block: b7{r8c1 r9c2}{n7 n9} ==> r9c2 ≠ 1, r8c1 ≠ 8, r8c1 ≠ 1 swordfish-in-columns: n5{c3 c4 c9}{r9 r1 r6} ==> r6c7 ≠ 5, r6c6 ≠ 5, r1c6 ≠ 5, r1c1 ≠ 5 swordfish-in-columns: n7{c2 c5 c8}{r9 r1 r6} ==> r6c7 ≠ 7, r6c6 ≠ 7, r1c6 ≠ 7, r1c1 ≠ 7 swordfish-in-columns: n9{c2 c5 c8}{r6 r9 r1} ==> r6c7 ≠ 9, r6c1 ≠ 9, r1c7 ≠ 9, r1c6 ≠ 9 hidden-pairs-in-a-block: b2{r1c5 r3c6}{n7 n9} ==> r3c6 ≠ 4, r3c6 ≠ 3, r3c6 ≠ 2, r3c6 ≠ 1, r1c5 ≠ 3, r1c5 ≠ 2, r1c5 ≠ 1 hidden-pairs-in-a-block: b6{r4c7 r6c8}{n7 n9} ==> r6c8 ≠ 8, r6c8 ≠ 4, r6c8 ≠ 1, r4c7 ≠ 8, r4c7 ≠ 6, r4c7 ≠ 1 swordfish-in-rows: n2{r3 r4 r8}{c4 c1 c6} ==> r6c6 ≠ 2, r6c4 ≠ 2, r6c1 ≠ 2, r2c6 ≠ 2, r2c1 ≠ 2, r1c6 ≠ 2, r1c4 ≠ 2, r1c1 ≠ 2 hidden-pairs-in-a-block: b4{r4c1 r6c2}{n2 n9} ==> r6c2 ≠ 6, r6c2 ≠ 3, r6c2 ≠ 1, r4c1 ≠ 8, r4c1 ≠ 6, r4c1 ≠ 1 hidden-triplets-in-a-column: c1{n2 n7 n9}{r4 r3 r8} ==> r3c1 ≠ 4, r3c1 ≠ 1 hidden-triplets-in-a-row: r1{n2 n7 n9}{c8 c2 c5} ==> r1c8 ≠ 8, r1c8 ≠ 4, r1c8 ≠ 1, r1c2 ≠ 6, r1c2 ≠ 3, r1c2 ≠ 1 naked-pairs-in-a-block: b1{r1c2 r3c1}{n2 n7} ==> r2c2 ≠ 2 hidden-triplets-in-a-column: c6{n2 n7 n9}{r8 r4 r3} ==> r8c6 ≠ 8, r8c6 ≠ 3, r8c6 ≠ 1, r4c6 ≠ 8, r4c6 ≠ 1 naked-triplets-in-a-row: r4{c1 c6 c7}{n9 n2 n7} ==> r4c4 ≠ 2 hidden-pairs-in-a-block: b5{r4c6 r6c5}{n2 n7} ==> r6c5 ≠ 8, r6c5 ≠ 3, r6c5 ≠ 1 swordfish-in-rows: n8{r1 r4 r8}{c7 c9 c3} ==> r9c9 ≠ 8, r9c3 ≠ 8, r7c7 ≠ 8, r6c9 ≠ 8, r6c7 ≠ 8, r6c3 ≠ 8, r5c7 ≠ 8 naked-triplets-in-a-row: r9{c3 c4 c9}{n5 n1 n4} ==> r9c8 ≠ 4, r9c8 ≠ 1, r9c5 ≠ 1 swordfish-in-columns: n1{c2 c5 c8}{r2 r5 r7} ==> r7c7 ≠ 1, r7c6 ≠ 1, r7c1 ≠ 1, r5c7 ≠ 1, r5c6 ≠ 1, r5c1 ≠ 1, r2c7 ≠ 1, r2c6 ≠ 1, r2c1 ≠ 1 x-wing-in-columns: n1{c1 c6}{r1 r6} ==> r6c9 ≠ 1, r6c7 ≠ 1, r6c4 ≠ 1, r6c3 ≠ 1, r1c9 ≠ 1, r1c7 ≠ 1, r1c4 ≠ 1, r1c3 ≠ 1 hidden-pairs-in-a-row: r6{n1 n8}{c1 c6} ==> r6c6 ≠ 3, r6c1 ≠ 6, r6c1 ≠ 4 hidden-pairs-in-a-block: b6{r4c9 r5c8}{n1 n8} ==> r5c8 ≠ 4, r4c9 ≠ 6 hidden-triplets-in-a-column: c7{n1 n7 n9}{r3 r8 r4} ==> r8c7 ≠ 8, r8c7 ≠ 3, r3c7 ≠ 4, r3c7 ≠ 3 hidden-single-in-a-column ==> r1c7 = 8 swordfish-in-columns: n3{c2 c5 c7}{r2 r5 r7} ==> r7c6 ≠ 3, r5c6 ≠ 3, r2c6 ≠ 3 stte 124573896 569814327 783269154 251647938 437985612 896321475 612738549 948152763 375496281

by **rjamil** » Thu Aug 13, 2020 1:19 pm

Here's my solver's solution path:

1) HP 79 @ r8c1 r9c2 in Box 7;
2) Row wise Sword Fish: 2 @ r348c146;
3) Row wise Sword Fish: 5 @ r257c167;
4) Row wise Sword Fish: 6 @ r257c127;
5) Row wise Sword Fish: 7 @ r348c167;
6) Row wise Sword Fish: 9 @ r348c167;
7) HT 279 @ r1c258 in Row 1;
8) HT 279 @ r6c258 in Row 6;
9) HT 279 @ r348c1 in Column 1;
10) HT 279 @ r169c2 in Column 2;
11) HT 279 @ r348c6 in Column 6;
12) HP 79 @ r1c5 r3c6 in Box 2;
13) NP 27 @ r4c6 r6c5 in Box 5;
14) HT 279 @ r4c167 in Row 4;
15) Row wise Sword Fish: 8 @ r148c379;
16) HT 789 @ r9c258 in Row 9; and
17) Row wise Jelly Fish: 1 @ r3489c3479 (compromised); STTE

R. Jamil

by **tarek** » Thu Aug 13, 2020 3:51 pm

Analysing how SE solves this ...

Conquering the puzzle into singles ends with the Naked single r6c7=4

To get there it requires are 6 Swordfishes in that solution path and 1 X-wing ...

Impressive!!!!

The debate about how easy it is to spot finned X-wings compared to Swordfishes is tricky because clearly when you are only dealing with conjugate pairs in the simplest 2 strong links techniques with rank 1 logic is different than having the grouped strong links with rank 1 logic v a swordfish with rank 0 logic ...

Also from a different perspective, some of these swordfishes like the franken swordfish used in a recent puzzle is a monster that I will always find difficulty in catching because the base sectors are not easy to spot conjugate pairs or grouped strong links that can be easily linked. I would certainly consider these grouped situations and swordfishes fairly advanced ...

Currently SE ranks a skyscraper 4.0 and grouped 2 strong linked techniques at 4.3 while the swordfish is 3.8 :idea:

(the basic swordfish technique is has a fixed rating even if it is a 3x3x3 fish) SE currently will catch fish larger than X-wing only if they are basic fish or if they are formed of conjugate pairs or grouped strong links!

The SE121 would have caught non basic fish larger than x-wing formed of conjugate pairs only as part of the Forced X-chain technique. Current Sukaku explainer has downgraded that rating as these fish can be caught now with the corresponding number of strong links algorithm in addition to the ones formed from grouped strong links.

tarek

by **mith** » Thu Aug 13, 2020 9:17 pm

denis_berthier wrote:Could you be more precise about the origin?

Oh, it's an original puzzle. As tarek alluded to, I posted this on github as an example as we were discussing SE's new ordering of techniques - SE 1.2.1 gives this a 4.0 (for the hidden triplets), while SE 1.17.7.1 with the new order/rating enabled gives this a 4.3 because it catches all the turbot fish before it applies the swordfish, despite swordfish having a 4.0 rating.

As mentioned in the 19 clue thread, I've been playing around a lot with this type of pattern in the givens, with a pair of offset 3x3s, which are prone to have swordfish just by the geometry. This one I found doing neighborhood searches starting from a pool of 19 and 20 clue puzzles, filtering based on the number of swordfish found by gsf's solver. (I've found a few with 9 swordfish available, though of course not all are required. I'll post one of them at some point.)

by **enxio27** » Thu Aug 13, 2020 10:11 pm

Hmmm. . . I may actually give this one a try. Swordfish is the most difficult technique I've learned, but I hadn't yet found a puzzle in which I found one that actually eliminated any candidates. This may give me a good workout, if I can get through it.

Website · by **denis_berthier** » Fri Aug 14, 2020 4:12 am

mith wrote:
denis_berthier wrote:Could you be more precise about the origin?

Oh, it's an original puzzle. As tarek alluded to, I posted this on github as an example as we were discussing SE's new ordering of techniques - SE 1.2.1 gives this a 4.0 (for the hidden triplets), while SE 1.17.7.1 with the new order/rating enabled gives this a 4.3 because it catches all the turbot fish before it applies the swordfish, despite swordfish having a 4.0 rating.

Great!
If you don't mind, I'd like to add it to the SudoRules standard examples (with your name, of course).
Do you have anything similar involving quads/jellyfish?

by **mith** » Fri Aug 14, 2020 4:41 am

Sure, I don't mind at all. And you might take a look at the one I posted today.

(I have a number of puzzles with quads and jellyfish, though I'm still working my way through them to check which ones are actually interesting.)

Website · by **denis_berthier** » Fri Aug 14, 2020 5:06 am

mith wrote:Sure, I don't mind at all. And you might take a look at the one I posted today. (I have a number of puzzles with quads and jellyfish, though I'm still working my way through them to check which ones are actually interesting.)

Great example also! Thanks

by **tarek** » Fri Aug 14, 2020 6:57 am

I'm sure all of you remember the Pure Swordfish Collection http://forum.enjoysudoku.com/a-pure-swordfish-collection-t5775.html

The New Sudoku Players' Forum

Tatooine Sunset / Tatooine Sunrise

Tatooine Sunset / Tatooine Sunrise

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset

Re: Tatooine Sunset