Fully symmetrical invalid patterns

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

Re: Fully symmetrical invalid patterns

Postby Serg » Fri Feb 24, 2017 1:51 pm

Hi, blue!
Since we're talking about patterns maximality ...
F1 pattern has no valid puzzles. It is, at least, symmetrically maximal (I checked this property). But is it maximal in the strict sense? My old result - "Man" pattern has no valid puzzles - proves, that F1 pattern is just symmetrically maximal. Unfortunately there is no independent crosscheck of my result. Could you check, does "Man" pattern really have no valid puzzles?
Code: Select all
  "Man" pattern

x . . . . . . . x
. x . . . . . x .
. . x . x . x . .
. . . x x x . . .
. . . x x x . . .
. . . x x x . . .
. . x . . . x . .
. x . . . . . x .
x . . . . . . . x

Serg
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Fri Feb 24, 2017 5:10 pm

I can confirm that "Man pattern" has no valid puzzles. I had to check about 1.571e9 ED puzzles which took ~3 hours.
Afmob
 
Posts: 130
Joined: 28 June 2011

Re: Fully symmetrical invalid patterns

Postby blue » Fri Feb 24, 2017 6:55 pm

I should have mentioned in my earlier post, that I had verified the "Man" pattern too -- timing similar to Afmob's.

Cheers.
blue
 
Posts: 572
Joined: 11 March 2013

Re: Fully symmetrical invalid patterns

Postby Serg » Fri Feb 24, 2017 8:00 pm

Hi, Afmob and blue!
Thank you for your help!

Serg
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Serg » Thu Mar 02, 2017 3:48 pm

Hi, people!
This is yet another fully symmetrical pattern having no valid puzzles. (Exhaustive search was done in 36 hours.)
Code: Select all
        F6

x x . . . . . x x
x . . . . . . . x
. . x . . . x . .
. . . x . x . . .
. . . . x . . . .
. . . x . x . . .
. . x . . . x . .
x . . . . . . . x
x x . . . . . x x

This pattern is symmetrically maximal (checked), but looks like not maximal in the strict sense (fast random search found no valid puzzles for some 1-clue extentions).

Serg
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Sat Mar 04, 2017 5:22 pm

I can confirm that F6 has no valid puzzles. I had to check 5.34e9 ED puzzles which took 10 hours.
Afmob
 
Posts: 130
Joined: 28 June 2011

Re: Fully symmetrical invalid patterns

Postby Serg » Sun Mar 05, 2017 6:09 am

Hi, Afmob!
Afmob wrote:I can confirm that F6 has no valid puzzles. I had to check 5.34e9 ED puzzles which took 10 hours.

Thank you for confirmation! Your searching program generates and checks 150000 puzzles per second. Impressive result!

This is another fully symmetrical invalid pattern. Exhaustive search took 33 hours.
Code: Select all
        F7

x . . . x . . . x
. x . . x . . x .
. . . . . . . . .
. . . x x x . . .
x x . x x x . x x
. . . x x x . . .
. . . . . . . . .
. x . . x . . x .
x . . . x . . . x

Serg

[Edited. I changed exhaustive search duration, because previous reported duration (2.7 hours) was wrong (partial but not exhaustive search was done because of a bug in searching algoritm).]
Last edited by Serg on Sat Mar 11, 2017 9:54 am, edited 1 time in total.
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Sun Mar 05, 2017 7:04 am

The pattern has 64 automorphisms and I would need to check about 2.95e9 ED puzzles, so I guess it would take me about 6 hours to check your result. I'll later edit this post once the computation is done.

Edit: I can confirm that F7 has no valid puzzles. I had to check 2.95e9 ED puzzles (estimate was spot on!) which took 5 hours.
Last edited by Afmob on Sun Mar 05, 2017 11:58 am, edited 1 time in total.
Afmob
 
Posts: 130
Joined: 28 June 2011

Re: Fully symmetrical invalid patterns

Postby Serg » Sun Mar 05, 2017 11:24 am

Hi!
This is the last fully symmetrical invalid pattern with 9 clues in the central box. Exhaustive search took 2.4 hours.
Code: Select all
        F8

. . . . x . . . .
. x . . . . . x .
. . x . . . x . .
. . . x x x . . .
x . . x x x . . x
. . . x x x . . .
. . x . . . x . .
. x . . . . . x .
. . . . x . . . .

Serg
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Sun Mar 05, 2017 12:31 pm

F8 has also no valid puzzles. I had to check 1.49e8 ED puzzles which took about 25 minutes.
Afmob
 
Posts: 130
Joined: 28 June 2011

Re: Fully symmetrical invalid patterns

Postby Serg » Mon Mar 06, 2017 7:06 am

Hi, Afmob!
Afmob wrote:F8 has also no valid puzzles. I had to check 1.49e8 ED puzzles which took about 25 minutes.

Thank you for confirmation!

Here are new potentially invalid patterns.
Code: Select all
        F9                   F10

. . x . x . x . .     x . . . x . . . x
. . . . x . . . .     . . . . x . . . .
x . . . . . . . x     . . x . . . x . .
. . . x . x . . .     . . . x . x . . .
x x . . x . . x x     x x . . x . . x x
. . . x . x . . .     . . . x . x . . .
x . . . . . . . x     . . x . . . x . .
. . . . x . . . .     . . . . x . . . .
. . x . x . x . .     x . . . x . . . x

Serg

[Edited. I changed "1" by "x" in the patterns F9 and F10 definitions to get more common "human" view.]
[Edited2. Comment: both patterns - F9 and F10 are subsets of F11 invalid pattern. So, it was a waste of time checking both patterns... I should be more careful.]
[Edited3. My searching algorithm for F9 and F10 patterns was not exhaustive, so I withdraw my results for these patterns.]
Last edited by Serg on Sat Mar 11, 2017 9:58 am, edited 2 times in total.
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Serg » Mon Mar 06, 2017 11:51 pm

Hi!
This is the next candidate invalid pattern with 5 clues in the central box.
Code: Select all
       F11

x . . . x . . . x
. x . . x . . x .
. . . . x . . . .
. . . x . x . . .
x x x . x . x x x
. . . x . x . . .
. . . . x . . . .
. x . . x . . x .
x . . . x . . . x

Serg

[Edited. I've done ehaustive search for this pattern - no valid puzzles.]
Last edited by Serg on Wed Mar 15, 2017 8:39 pm, edited 2 times in total.
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Tue Mar 07, 2017 4:27 pm

That took longer than I thought: F9 and F10 have no valid puzzles. I had to check 1.41e10 ED puzzles for F9 and 1.42e10 ED puzzles for F10 which took about 10 hours for each pattern. Note that the computation was done using a different, slower CPU. I think F11 will take at least 24 hours using my approach even though there are a lot of invalid puzzles (e.g. having no solution). Sadly, I don't have my estimation programm on this PC, so it's hard to guess how long it would actually take.
Afmob
 
Posts: 130
Joined: 28 June 2011

Re: Fully symmetrical invalid patterns

Postby Serg » Tue Mar 07, 2017 9:38 pm

Hi!
This is the next fully symmetrical invalid pattern with 5 clues in the central box. Exhaustive search took 13.8 hours.
Code: Select all
       F12

x x . . x . . x x
x . . . . . . . x
. . . . . . . . .
. . . x . x . . .
x . . . x . . . x
. . . x . x . . .
. . . . . . . . .
x . . . . . . . x
x x . . x . . x x

Serg
Serg
2017 Supporter
 
Posts: 499
Joined: 01 June 2010
Location: Russia

Re: Fully symmetrical invalid patterns

Postby Afmob » Wed Mar 08, 2017 5:42 am

I can confirm that F11 has no valid puzzles. I had to check 4.2e9 ED puzzles which took about 10 hours though I added some lines to my otherwise general code for this pattern to filter out invalid puzzles, otherwise it would have taken too long.

Serg, do you have an estimate on how many patterns you have to analyze for this study?
Afmob
 
Posts: 130
Joined: 28 June 2011

PreviousNext

Return to General