2 skyscrapers in sequence not so good idea

Programs which generate, solve, and analyze Sudoku puzzles

2 skyscrapers in sequence not so good idea

Postby Hajime » Thu Apr 30, 2020 2:01 pm

Puzzle 301 of PG26855.txt is:
100000002030040050006201400001706200000000000003102800008309700020060090300000006 56 28 26 m_b_metcalf
After solving 37 singles (and some naked/hidden subsets) this is the result:
1845.3..2.32648.5...62.14388917362452.5984.1..431528.9..8319724427865.9.319427586

Code: Select all
 +-----------+------------+-----------+
 |  1   8  4 |  5  79   3 |  69 67  2 |
 |  79  3  2 |  6   4   8 |  19 5  17 |
 |  579 57 6 |  2  79   1 |  4  3   8 |
 +-----------+------------+-----------+
 |  8   9  1 |  7   3   6 |  2  4   5 |
 |  2  67  5 |  9   8   4 |  36  1 37 |
 |  67  4  3 |  1   5   2 |  8  67  9 |
 +-----------+------------+-----------+
 |  56 56  8 |  3   1   9 |  7   2  4 |
 |  4   2  7 |  8   6   5 |  13  9 13 |
 |  3   1  9 |  4   2   7 |  5   8  6 |
 +-----------+------------+-----------+

Now I think there are 2 Skyscrapers :
(7)r5,32c29 =>r2c1<>7 and after eliminating the 7, the second (7)r6,21c18 =>r2c9<>7
That leaves r1c8 = 7 and no solution for 7's in row 2.

But there is a unique solution to the puzzle:
184573962932648157756291438891736245275984613643152879568319724427865391319427586

What happened...
2 skyscrapers in sequence is not a good idea.
After the first skyscraper you have to look for singles (stte)
and not for another skyscraper.

But in my solver I often search for more candidate eliminations during the same solving method :oops:

Edit: Typo 5->2 in box 7 corrected
Last edited by Hajime on Thu Apr 30, 2020 3:15 pm, edited 1 time in total.
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby yzfwsf » Thu Apr 30, 2020 2:18 pm

There are Two single by '5' in row 8.

BTW:My solver found a bug+1 then stte
Last edited by yzfwsf on Thu Apr 30, 2020 3:37 pm, edited 1 time in total.
yzfwsf
 
Posts: 845
Joined: 16 April 2019

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Thu Apr 30, 2020 3:16 pm

yzfwsf wrote:There are Two single by '5' in row 8.

Sorry. Is now corrected. Point stands. How to proceed during elimination of candidates process during the same solving method?
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby champagne » Thu Apr 30, 2020 5:03 pm

Hajime wrote:
yzfwsf wrote:There are Two single by '5' in row 8.

Sorry. Is now corrected. Point stands. How to proceed during elimination of candidates process during the same solving method?


except for rules involving uniqueness, the solving order has no importance, so generally speaking, applying in once all equivalents (in rating) moves is correct;.

Side question??
How did you get this wrong pm
champagne
2017 Supporter
 
Posts: 7334
Joined: 02 August 2007
Location: France Brittany

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Thu Apr 30, 2020 5:47 pm

I was extending my solver with Kites and Skyscrapers. Like always I programmed the subroutine finding all skyscrapers. And accidentally I tested it on the 26855 puzzles from m_b_metcalf. There the problem shows up...
At 301 I get the message that the puzzle was ruined...
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Fri May 01, 2020 4:02 pm

yzfwsf wrote:BTW:My solver found a bug+1 then stte

But did you search for skyscrapers? Not in the list of SE, but probably with a lower value than Bug+1
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Fri May 01, 2020 4:06 pm

champagne wrote:except for rules involving uniqueness, the solving order has no importance, so generally speaking, applying in once all equivalents (in rating) moves is correct;.

This is in contradiction with the example with a sequence of skyscrapers above. Like you I never would expect this, but shit happens...
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Sun May 03, 2020 8:44 am

After r2c1<>7 there is no second skyscraper. Only if some sloppy administration does not remove the 7, the second is present (illegally).
My apologies to every one for my mistake.
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby champagne » Sun May 03, 2020 11:59 am

Hajime wrote:Now I think there are 2 Skyscrapers :
(7)r5,32c29 =>r2c1<>7 and after eliminating the 7, the second (7)r6,21c18 =>r2c9<>7

Can you tell more about this strange Skyscrapers description
the pm at this point is valid, each final value is in the pm

Code: Select all
     +-----------+------------+-----------+
     |  1   8  4 |  5  79   3 |  69 67  2 |184573962
     |  79  3  2 |  6   4   8 |  19 5  17 |932648157
     |  579 57 6 |  2  79   1 |  4  3   8 |756291438
     +-----------+------------+-----------+
     |  8   9  1 |  7   3   6 |  2  4   5 |891736245
     |  2  67  5 |  9   8   4 |  36  1 37 |275984613
     |  67  4  3 |  1   5   2 |  8  67  9 |643152879
     +-----------+------------+-----------+
     |  56 56  8 |  3   1   9 |  7   2  4 |568319724
     |  4   2  7 |  8   6   5 |  13  9 13 |427865391
     |  3   1  9 |  4   2   7 |  5   8  6 |319427586
     +-----------+------------+-----------+
champagne
2017 Supporter
 
Posts: 7334
Joined: 02 August 2007
Location: France Brittany

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Sun May 03, 2020 1:34 pm

(7)r5,32c29 : skyscraper with base row 5 en column 2 and 9 crossing top rows 3 and 2 respectively. Then candidate 7 at r2c1 can be eliminated.
What is the appropriate description?
Note that there is no problem anymore. The second skyscraper can not occur in this example/puzzle.

Edit: Again typos in red... Terrible sorry
Last edited by Hajime on Sun May 03, 2020 6:40 pm, edited 1 time in total.
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby yzfwsf » Sun May 03, 2020 2:17 pm

Hajime wrote:But did you search for skyscrapers? Not in the list of SE, but probably with a lower value than Bug+1

skyscraper.png
skyscraper.png (65.19 KiB) Viewed 896 times
yzfwsf
 
Posts: 845
Joined: 16 April 2019

Re: 2 skyscrapers in sequence not so good idea

Postby champagne » Sun May 03, 2020 2:37 pm

Hajime wrote:(7)r2,32c29 : skyscraper with base row 2 en column 2 and 9 crossing top rows 3 and 2 respectively. Then candidate 7 at r2c1 can be eliminated.
What is the appropriate description?
Note that there is no problem anymore. The second skyscraper can not occur in this example/puzzle.

I am not specialist of the notations, but I would have written this
skyscraper 7c2r43 + 7 c9r42 => <7r2c1>
champagne
2017 Supporter
 
Posts: 7334
Joined: 02 August 2007
Location: France Brittany

Re: 2 skyscrapers in sequence not so good idea

Postby tarek » Sun May 03, 2020 3:00 pm

Going through SE1.17.5:

3 Skyscrapers available from your starting PM that have different eliminations.
You can still arrive to these eliminations using other "2 strong link variations" (An additional 7) (All found by YZF_Sudoku)

BTW I would prefer using the term Empty rectangle to refer to a grouped strong link in a box. The variations include strong links in boxes but no grouped Strong links and therefore no empty rectangle.

You can eliminate 7 at r3c5 using 3 Strong links:
SE1.17.5 wrote:This technique relies on Conjugate Pairs. The value 7 in row 1, row 6 and column 2 forms 3 (regional) strong links or conjugate pairs with 2 sets of bridge cells in column 8 and block 4. The 3 strong links are joined therefore with these 2 weak links.
Any occurrence of the value 7 can be removed from any cells sharing a row, column or block with both start and end cells r1c5 and r3c2.
This technique is closely related to 2x2x1 Finned Sashimi Swordfish technique logic, Coloring technique or L1-Wing technique. It is a subcategory of X-Chains group of techniques which is in itself a subgroup of Alternating Inference Chains (AIC) group of techniques. These use strong links in Location inferred from 1 value (L1), therfore, L1-Wing can be used to describe this pattern.
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Re: 2 skyscrapers in sequence not so good idea

Postby Hajime » Mon May 04, 2020 4:37 pm

I searched this forum and I think the correct notation for skyscraper is:
Skyscraper (7)r3c2=r5c2-r5c9=r2c9 => (-7) r2c1
In this case the weak link is also strong because there are only 2 candidates 7 in row 5, but that is not mandatory for skyscrapers.
Furthermore I like the (-7) better than <>7, because all the candidates are now between brackets.
User avatar
Hajime
 
Posts: 1348
Joined: 20 April 2018
Location: Fryslân

Re: 2 skyscrapers in sequence not so good idea

Postby rjamil » Mon May 04, 2020 6:23 pm

Hi yzfwsf,

Well, it's off topic, I see two BUG+1 moves in your "All possible steps". The "Bivalue Universal Grave + 1: => r3c1 <> 7" is wrong.

BTW, the full form of BUG is "Bivalue Universal Grave". Either it should be 7r3c1 or r3c1<>59.

R. Jamil
rjamil
 
Posts: 729
Joined: 15 October 2014
Location: Karachi, Pakistan

Next

Return to Software