The New Sudoku Players' Forum

[mith]

Code: Select all

+-------+-------+-------+
| . 9 . | . . 1 | 8 6 . |
| . . 8 | . 3 . | 1 . . |
| . . . | 9 . . | . . 3 |
+-------+-------+-------+
| . 3 1 | . 9 . | . . 6 |
| 2 . . | . . 7 | . . . |
| . . . | . . . | 5 . . |
+-------+-------+-------+
| 3 . . | . . . | . 4 . |
| . 6 2 | . 1 . | . . 8 |
| . . . | 2 . . | . . . |
+-------+-------+-------+
.9...186...8.3.1.....9....3.31.9...62....7.........5..3......4..62.1...8...2.....

(This is the first in what may be a series of puzzles which are rated 11+ by SE but fall to a single application of some advanced technique not implemented by SE.)

[Cenoman]

Code: Select all

 +--------------------------+-----------------------------+--------------------------+
 |  457      9       3      |  457     T457-2    1        |  8      6        2457    |
 |  457-6   T457-2   8      |  4567     3        2456     |  1      2579     24579   |
 |  14567    12457   4567   |  9        245678   24568    | B247   B257      3       |
 +--------------------------+-----------------------------+--------------------------+
 |  4578     3       1      |  458      9        2458     |  247    278      6       |
 |  2        458     4569   |  134568   4568     7        |  349    1389     149     | 45
 |  46789    478     4679   |  13468    2468     23468    |  5      123789   12479   |24 7
 +--------------------------+-----------------------------+--------------------------+
 |  3        1578    579    |  5678     5678     5689     |  2679   4        12579   |2 57
 |  4579     6       2      |  3457     1        3459     |  379    3579     8       |
 |  145789   14578   4579   |  2        45678    345689   |  3679   13579    1579    | 457
 +--------------------------+-----------------------------+--------------------------+
             CL2     CL-extra           CL1                                  CLB

JE2 (2457)r3c78, r1c5, r2c2, with extra cross-line c3

The typical criteria, as defined by David P. Bird in his JExocet Compendium are not met: three cover houses for digits 4, 5, 7 in cross-lines CLB, CL1, CL2. As per extension of the pattern proposed by yzfwsf here (first found by qiuyanzhe), the fundamental exocet property is met if an extra cross-line can be found, that contains no base digit in its two cells in sight of target cells: here, column 3 does the job.

Eliminations from the pattern: -2 r1c5, r2c2 (base digit false in opposite mirror nodes), -6 r2c1 (non base digit false in mirror node, as 8r2c3 locked in mirror node)
Finish with basics: lclste.

[mith]

Yep (and this puzzle comes from checking puzzles against yzf's solver, no surprise there )

SE rates this one 11.6.

[SpAce]

Nice pattern. I'm really surprised to hear it's a recent invention, as it seems like a very obvious JE extension. Instead of a Swordfishy pattern it has a Jellyfishy pattern. Nothing else changes. (Logically the S-cells should be called J-cells in such a pattern. As I recall, the 'S' originally meant Swordfish.)

The use of fish terms to describe the pattern of 'S' cells has been completely dropped to avoid unnecessary controversy.

I wonder what the controversy was. To me it's much easier to see the whole thing as a sort of fish pattern (c\r). In this case, if you have two of (4,5,7) in the base cells, there has to be one of each in the targets, or otherwise you have at least one headless and finless Jellyfish with those digits (i.e. an illegal fish; 4 bases\3 covers => Rank -1).

[SpAce]

Hi Cenoman,

Eliminations from the pattern: -2 r1c5, r2c2 (base digit false in opposite mirror nodes)

That works nicely, giving the critical eliminations directly. You could eliminate 2r3c78, too, because they just became impossible base candidates. I got the same four eliminations the other way around. I saw 2 as an impossible base digit, because it would occupy both targets.

-6 r2c1 (non base digit false in mirror node, as 8r2c3 locked in mirror node)

That's needed either way.

[Cenoman]

I'm really surprised to hear it's a recent invention, ....

My own awareness is recent, but I read nothing about it before yzfwsf posted solutions using it to solve "Weekly Unsolvable" at sudokuwiki.org (end 2019 or early 2020) I suggested him to post some information about it. I may have missed previous written information...

...as it seems like a very obvious JE extension.

Agreed, it's obvious, and easy to spot manually.

Hi Cenoman,

Eliminations from the pattern: -2 r1c5, r2c2 (base digit false in opposite mirror nodes)

That works nicely, giving the critical eliminations directly. You could eliminate 2r3c78, too, because they just became impossible base candidates. I got the same four eliminations the other way around. I saw 2 as an impossible base digit, because it would occupy both targets.
Agreed. False in both mirror nodes is enough to conclude that a base digit is false.

-6 r2c1 (non base digit false in mirror node, as 8r2c3 locked in mirror node)

That's needed either way.

Agreed too.

The concern with Exocet pattern is: how much detailed information to post ? (notably, should elimination criteria in David's Compendium be referred to ?)

[SpAce]

My own awareness is recent, but I read nothing about it before ysfwsf posted solutions using it to solve "Weekly Unsolvable" at sudokuwiki.org (end 2019 or early 2020) I suggested him to post some information about it.

Be as it may, I'm glad it's in the spotlight. Perhaps no one thought about that possibility before. One reason might be that the JE pattern, as defined in the Compendium, is a bit overspecified and underexplained. It might make it harder to see the big picture and possible generalizations.

(Make no mistake, I think the Compendium is awesome as a whole. Just not perfect, but what is?)

Agreed, it's obvious, and easy to spot manually.

Personally I can't say that any JE is exactly easy to spot manually, but I don't think this variant is much more difficult than a normal JE.

(Btw, if it wasn't clear, I used your finding in this case. I didn't even try to look for it manually. My JE spotting skills are very poor, so any tips you might have would be appreciated. So far I've been satisfied with understanding the pattern and its ramifications.)

The concern with Exocet pattern is: how much detailed information to post ? (notably, should elimination criteria in David's Compendium be referred to ?)

I think it's good to refer to the Compendium rules for any derived eliminations, like you did. It made your solution very easy to understand for someone who's familiar with the documentation.

I count as derived eliminations anything but the very basic "non-base digits in targets". All of the other rules, even the simple mirror node inferences etc, are easy to forget if one doesn't use the JE a lot. For example, I don't think this JE solution presents nearly enough information.

[RSW]

I don't think this JE solution presents nearly enough information.

Fixed.

[SpAce]

I don't think this JE solution presents nearly enough information.

Fixed.

Perfect! You could also mention that only one or the other JE is actually needed, if you want to make your solution a two-stepper. Btw, my solution was based on your findings, so thanks for that.