## NEED HELP about SK-LOOP, or something else

Advanced methods and approaches for solving Sudoku puzzles

### NEED HELP about SK-LOOP, or something else

Hi there

I have a problem when i solving a puzzle. a strange structure seems like SK-LOOP but i think it's not; and something like the MSLS.
134+2/9/8 in r6c4689; 234+8/9/2 in r9c4689; 147+9/2/8 in r4578c5; 147+8/9/2 in r4578c7
question:
1.is there any candidate can be eleminated through these grids (with a symbol in front of it) ONLY?
2.is this a SK-loop? or something else? or nothing?

thank you very much!

Code: Select all
`*--------------------------------------------------------------------------------------------*|     24589     24589         1|     24678         3      4679|      2789     56789      2567||       289         3       289|     12678     12789         5|     12789     16789         4||         6     24589         7|      1248     12489       149|         3      1589       125||------------------------------+------------------------------+------------------------------||     13459      4569     34569|     13457     #1479         8|      &147         2      1367||    123458     24568    234568|    123457     #1247      1347|     &1478     13678         9||    123489         7     23489|     *1234         6     *1349|         5      *138       *13||------------------------------+------------------------------+------------------------------||     34579      4569     34569|     13467      #147         2|     &179     13579         8||     23578      2568     23568|         9      #178      1367|      &127         4     12357||    234789         1     23489|     %3478         5      %347|         6      %379      %237|*--------------------------------------------------------------------------------------------*`
borescoper

Posts: 19
Joined: 11 June 2016

### Re: NEED HELP about SK-LOOP, or something else

My solver produces the following:

MSLS Base: 2489
16 cell Truths: c4689 r1369
16 links: 248c4, 49c6, 89c8, 2c9, 67r1, 1r3, 13r6, 37r9, 5b3
20 eliminations: r1c7<>7, r3c5<>1, r6c1<>1, r6c1<>3, r6c3<>3, r9c1<>3, r9c1<>7, r9c3<>3, r2c4<>2, r2c4<>8, r4c4<>4, r5c4<>2, r5c4<>4, r7c4<>4, r5c6<>4, r2c8<>8, r2c8<>9, r5c8<>8, r7c8<>9, r8c9<>2,

No sign of an SK loop
Phil
pjb
2014 Supporter

Posts: 2025
Joined: 11 September 2011
Location: Sydney, Australia

### Re: NEED HELP about SK-LOOP, or something else

Similar solver results :

MSLS 1 : Base 1367; r1369 c4689 : 16 Links; 67r1 1r3 13r6 37r9 ; 248c4 49c6 89c8 2c9 ; 5b3 ;

MSLS 2 : Base 2489; c4689 r1369 : 16 Links; 248c4 49c6 89c8 2c9 ; 67r1 1r3 13r6 37r9 ; 5b3 ;

Multifish 1.1: 18 Truths = { 2489R1 2489R3 2489R6 2489R9 2N13 } 18 Links = { 2c49 4c46 8c48 9c68 1n7 3n5 6n13 9n13 2489b1 }

Multifish 3.1: 18 Truths = { 2489R6 2489R9 2489C5 2489C7 2N13 } 18 Links = { 289r2 1n7 3n5 6n13 9n13 2b59 4b568 8b68 9b59 }

All gave the same 20 eliminations as Phil. No SK loop.

Leren
Leren

Posts: 3366
Joined: 03 June 2012

### Re: NEED HELP about SK-LOOP, or something else

This is an Almost SK Loop which would be complete if r5c4 held 4.
Using the standard SK Loop cover pattern it can be rescued by adding a cover set for r2

MS-NS:(289)r2,(13)r6,(37)r9,(17)c5,(17)c7,(249)b5,(48)b6,(48)b8,(29)b9; 20 DigitInstances/20 Cells
=> 20 Elims:7r1c7, 28r2c4, 89r2c8, 1r3c5, 4r4c4, 24r5c4, 4r5c6, 8r5c8, 13r6c1, 3r6c3, 4r7c4, 9r7c8, 2r8c9, 37r9c1, 3r9c3

Phil, Leren, it also contains (289)JuniorExocet2:r2c13,r3c5,r1c7

borescoper, see the < Domino Loops thread > for a description of the SK Loop and its close relatives, but I'd leave studying the JuniorExocet pattern for later if I were you.

David
David P Bird
2010 Supporter

Posts: 1040
Joined: 16 September 2008
Location: Middle England

### Re: NEED HELP about SK-LOOP, or something else

David P bird wrote : Phil, Leren, it also contains (289)JuniorExocet2:r2c13,r3c5,r1c7

Yes I thought there might be an Exocet on offer because of the Multifish 3.1 pattern, which is often assocated with Exocets.

I was able to prove that 89 was an incompatible pair using the methods you outlined in your Compendium (so 2 is True in the Base), and that 8 was False in the base using some older more complex code.

The Exocet base reduces to 29, for a total of 30 Exocet eliminations, followed by Multifish 1.1: 16 Truths = { 2489R1 2489R3 2489R6 2489R9 } 16 Links = { 2c49 4c46 8c48 9c68 1n7 3n5 6n13 9n13 48b1 } for 13 more eliminations (no doubt a subset of the original 20 eliminations). This reduces the PM to

Code: Select all
`*-----------------------------------------------------------------------*| 458    458    1       | 24678  3      4679    | 29     56789  2567    || 29     3      29      | 167    178    5       | 178    167    4       || 6      458    7       | 1248   29     149     | 3      1589   125     ||-----------------------+-----------------------+-----------------------|| 1345   4569   3456    | 1357   1479   8       | 147    2      1367    || 13458  24568  34568   | 1357   1247   137     | 1478   1367   9       || 2489   7      2489    | 1234   6      1349    | 5      138    13      ||-----------------------+-----------------------+-----------------------|| 3457   4569   3456    | 1367   147    2       | 179    1357   8       || 3578   2568   3568    | 9      178    1367    | 127    4      1357    || 2489   1      2489    | 3478   5      347     | 6      379    237     |*-----------------------------------------------------------------------*`

It's a bit late at night - I might investigate further tomorrow if I have the time.

Leren
Leren

Posts: 3366
Joined: 03 June 2012

### Re: NEED HELP about SK-LOOP, or something else

David P Bird wrote:This is an Almost SK Loop which would be complete if r5c4 held 4.
Using the standard SK Loop cover pattern it can be rescued by adding a cover set for r2

MS-NS:(289)r2,(13)r6,(37)r9,(17)c5,(17)c7,(249)b5,(48)b6,(48)b8,(29)b9; 20 DigitInstances/20 Cells
=> 20 Elims:7r1c7, 28r2c4, 89r2c8, 1r3c5, 4r4c4, 24r5c4, 4r5c6, 8r5c8, 13r6c1, 3r6c3, 4r7c4, 9r7c8, 2r8c9, 37r9c1, 3r9c3

Phil, Leren, it also contains (289)JuniorExocet2:r2c13,r3c5,r1c7

borescoper, see the < Domino Loops thread > for a description of the SK Loop and its close relatives, but I'd leave studying the JuniorExocet pattern for later if I were you.

David

hi David,
would u plz post a puzzle code with candidates, and sign the 20 cells of the MS-NS? i am not very clear on locating them... thanks a lot!!!
Last edited by borescoper on Sun Jun 19, 2016 11:24 am, edited 1 time in total.
borescoper

Posts: 19
Joined: 11 June 2016

### Re: NEED HELP about SK-LOOP, or something else

and i would like to extend my sincere appreciation to you guys:)
borescoper

Posts: 19
Joined: 11 June 2016

### Re: NEED HELP about SK-LOOP, or something else

David P Bird wrote:This is an Almost SK Loop which would be complete if r5c4 held 4.
Using the standard SK Loop cover pattern it can be rescued by adding a cover set for r2

MS-NS:(289)r2,(13)r6,(37)r9,(17)c5,(17)c7,(249)b5,(48)b6,(48)b8,(29)b9; 20 DigitInstances/20 Cells
=> 20 Elims:7r1c7, 28r2c4, 89r2c8, 1r3c5, 4r4c4, 24r5c4, 4r5c6, 8r5c8, 13r6c1, 3r6c3, 4r7c4, 9r7c8, 2r8c9, 37r9c1, 3r9c3

Phil, Leren, it also contains (289)JuniorExocet2:r2c13,r3c5,r1c7

borescoper, see the < Domino Loops thread > for a description of the SK Loop and its close relatives, but I'd leave studying the JuniorExocet pattern for later if I were you.

David

now i understand.. thank you!
borescoper

Posts: 19
Joined: 11 June 2016