Where to find extreme(+) puzzles?

Everything about Sudoku that doesn't fit in one of the other sections

Re: Where to find extreme(+) puzzles?

Postby SpAce » Tue Sep 26, 2017 4:24 am

I said above that I like to take the path of least resistance, but I guess it's a relative term. On second thought, I may need to make some adjustments even if they're not absolutely necessary for solving success. For example:

Last weekend I solved another Nightmare (Feb 3, 2008) which turned out to be the most difficult (7.9) of them I've encountered so far: 040000030200503004007000600030409050000010000090205040008000400500706008070000090. Still, I guess it wasn't nearly as difficult as my solve path would suggest.

My solution used 18 chains and nets, some of which were pretty complicated. I did enjoy hunting them, but it's hard to justify the complexity if I look at the Hoduko solution which used only 9 chains, none of which were very complicated.

I suspect my biggest leak was that I skipped fishing as always. Seems that I need to start paying attention to that if I want to simplify my solve paths. There are other leaks, of course, but that seems the most obvious and perhaps easiest to fix first.
SpAce
 
Posts: 69
Joined: 22 May 2017

Re: Where to find extreme(+) puzzles?

Postby StrmCkr » Tue Sep 26, 2017 5:07 am

What are those? I can find information about M-wings and L-wings you also mentioned but not those two.


in essence A.I.C / discontinuous/continuous Nice loops of 3/4 strong=weak links between 1-2 digits are classed and defined and finally named:

this post contains direct links for the other types i mentioned]

S-wing

the H-wing is used by a few people on here an a few other different forums and first appeared and was joking dubbed as a "purple cow" for us long timer :) {which is still displayed as such in xsudoku! }

    Hybrid wings. {pattern defined but no developed thread with examples as it has way to many to display them all. }

    H2-Wing: (X=Y)a - (Y)b = (Y-Z)c = (Z)d "a" and "d" in same unit; a<>Z, d<>X
    H3-Wing: (X=Y)a - (Y=Z)b - (Z)c = (Z)d "a" and "d" in same unit; a<>Z, d<>X


Code: Select all
.-----------------.-------------------.---------------------.
| 1689  4    1569 | 1689  26789  1278 | 125789  3     12579 |
| 2     168  169  | 5     6789   3    | 1789    178   4     |
| 3     158  7    | 189   2489   1248 | 6       128   1259  |
:-----------------+-------------------+---------------------:
| 1678  3    126  | 4     678    9    | 1278    5     1267  |
| 4678  568  2456 | 368   1      78   | 2789    2678  23679 |
| 1678  9    16   | 2     3678   5    | 178     4     1367  |
:-----------------+-------------------+---------------------:
| 169   126  8    | 139   2359   12   | 4       1267  12567 |
| 5     12   49   | 7     49     6    | 3       12    8     |
| 146   7    3    | 18    2458   1248 | 125     9     1256  |
'-----------------'-------------------'---------------------'

basics to this point then:
Almost Locked Set XY-Wing: A=r3c24568 {124589}, B=r8c5 {49}, C=r1268c3 {14569}, X,Y=4,5, Z=9 => r12c5<>9
W-Wing: 1/2 in r7c6,r8c8 connected by 2 in r78c2 => r7c89<>1
Finned Swordfish: 8 r146 c157 fr1c4 fr1c6 => r23c5<>8
Locked Candidates Type 1 (Pointing): 6 in b1 => r2c5<>6
followed by easy steps brings it to here:
Code: Select all
.----------------.-----------------.-------------------.
| 189   4    159 | 6    28    128  | 12579  3    12579 |
| 2     168  169 | 5    7     3    | 19     18   4     |
| 3     158  7   | 189  249   1248 | 6      128  125   |
:----------------+-----------------+-------------------:
| 1678  3    126 | 4    68    9    | 1278   5    127   |
| 48    58   245 | 38   1     7    | 289    6    239   |
| 1678  9    16  | 2    368   5    | 178    4    137   |
:----------------+-----------------+-------------------:
| 169   126  8   | 139  2359  12   | 4      7    256   |
| 5     12   49  | 7    49    6    | 3      12   8     |
| 146   7    3   | 18   2458  1248 | 125    9    1256  |
'----------------'-----------------'-------------------'

Almost Locked Set XY-Wing: A=r1c156 {1289}, B=r23c8,r3c9 {1258}, C=r7c1269 {12569}, X,Y=5,9, Z=1,2 => r1c79<>1, r1c79,r3c56<>2
Almost Locked Set XY-Wing: A=r2c238 {1689}, B=r126c3 {1569}, C=r1c79,r2c7 {1579}, X,Y=1,5, Z=9 => r1c1,r8c3<>9
more basic steps:
Code: Select all
.---------------.-------------.--------------.
| 18    4   59  | 6   2    18 | 579   3  579 |
| 2     16  169 | 5   7    3  | 19    8  4   |
| 3     58  7   | 9   4    18 | 6     2  15  |
:---------------+-------------+--------------:
| 1678  3   126 | 4   68   9  | 1278  5  127 |
| 4     58  25  | 38  1    7  | 289   6  239 |
| 1678  9   16  | 2   368  5  | 178   4  137 |
:---------------+-------------+--------------:
| 9     16  8   | 13  35   2  | 4     7  56  |
| 5     2   4   | 7   9    6  | 3     1  8   |
| 16    7   3   | 18  58   4  | 25    9  256 |
'---------------'-------------'--------------'

Almost Locked Set XZ-Rule: A=r1246c3 {12569}, B=r12456c7 {125789}, X=5, Z=2 => r4c9<>2
Almost Locked Set XZ-Rule: A=r15c3 {259}, B=r134679c9 {1235679}, X=9, Z=2 => r5c9<>2

singles to the end.
Some do, some teach, the rest look it up.
User avatar
StrmCkr
 
Posts: 663
Joined: 05 September 2006

Re: Where to find extreme(+) puzzles?

Postby ghfick » Tue Sep 26, 2017 12:23 pm

champagne wrote:
Here, we have an heavily downgraded form of potential exocet for the base r12c9 target r6c7 r7c8

1 in the base => 1r7c8
5 in the base => 5r6c7

if you can prove that 4 in the base => 4 (r6c7,r7c8) you have an exocet.

Junior exocet in the most common pattern to prove it and is there in most cases.
In other cases "classical AICs" can establish it


Using the [hidden] UR : r16c79 with {4,5} we see that r6c7 <> 4. So to have an exocet here, we would need to prove that 4 in the base => 4r7c8. Alas, there does not appear to be such a proof.
Can you suggest [or point to] a tough puzzle that contains a 'real' exocet that is not Junior?
David P Bird's compendium contains a section on the 'Senior Exocet'. You are suggesting a different type of extension where the base cells and target cells are in the same band but the 'S' cell requirement is not met.
The use of AICs in this situation suggests some commonalities with Krakens.
ghfick
 
Posts: 23
Joined: 06 April 2016

Re: Where to find extreme(+) puzzles?

Postby David P Bird » Tue Sep 26, 2017 4:45 pm

GF the 'Senior Exocet' variation described in the compendium allows target cells to be outside the base cell band. It can just about be classed as a recognisable pattern as it consists of a defninite number of elements that can be checked without needing to track chains. But Champagne's base cells and targets are all in the same band so it does not apply.

As Champagne is willing to accept inference tracking using chains or nets as a means of proving that a pattern exists, he has greater freedom to confirm the pattern is an Exocet.

Champagne, I hope that this contribution makes your life easier as it should save you from having to either read for the first time or revise the defintion of a Senior Exocet!

David
.
David P Bird
2010 Supporter
 
Posts: 960
Joined: 16 September 2008
Location: Middle England

Re: Where to find extreme(+) puzzles?

Postby SpAce » Mon Oct 02, 2017 10:34 pm

Thanks for the info, StrmCkr!
SpAce
 
Posts: 69
Joined: 22 May 2017

Previous

Return to General