exocet pattern in hardest puzzles

Advanced methods and approaches for solving Sudoku puzzles

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 9:14 am

David P Bird wrote:champage Those results are fantastic!

They may be biased though because they are all from the 'hardest' category where a lot of the same design tricks have often been used. I would expect lower figures for the general population.

DPB



At the start, the population of "potential hardest" was surely heavily biased.
To-day, I am intimately convinced it's not any more. A lot of work has been done in that field and a recent game has shown that the vicinity search can find all the solutions for a specific pattern.

The only true bias in that population is that they all are "chains resisting" and "chain nets" resisting.

David P Bird wrote:champage Those results are fantastic!

As I see it, the 5 puzzles match your bi-bi definition, but don't match the Exocet definition. If your code made the Exocet eliminations as soon as the pattern was found there should be a saving in execution time. How big that might be will depend on how often you need to continue with a brute force template analysis for the bi-bi digits anyway.

DPB


They all strictly match the exocet definition. It's just that the way to establish the elementary "base to target" links are not complying with your limited definition.

I'll prepare an example to show that

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby David P Bird » Mon Apr 09, 2012 10:06 am

champagne wrote:They all strictly match the exocet definition. It's just that the way to establish the elementary "base to target" links are not complying with your limited definition.

I'll prepare an example to show that

OK, OK. I can't accept that your Exocet definition is complete, but as you are its 'owner' it's your decision.

We therefore have two types of Exocet: those that just comply with your base and target cell conditions and the subset that also comply with the extended conditions I have proposed for manual solvers. How do you want to name them?

DPB
David P Bird
2010 Supporter
 
Posts: 957
Joined: 16 September 2008
Location: Middle England

Re: bi bi pattern in hardest puzzles

Postby ronk » Mon Apr 09, 2012 10:25 am

champagne wrote:They all strictly match the exocet definition. It's just that the way to establish the elementary "base to target" links are not complying with your limited definition.

Now would be a good time to re-post your definition. I say "re-post", even though I'm not actually aware of a thorough definition anywhere. Examples everywhere, but complete definition nowhere AFAIK.

David P Bird's recent offering is close, I suspect, but it doesn't fully handle optional candidates and degenerate cases.

[edit: add the following]
champagne, you might have two definitions, the first being ideal and complete, the second being "as implemented" in your program. If so, it would be helpful to know the differences between the two.

Also, it would be nice if someone else would/could confirm champagne's findings. Has anyone else programmed an identical or similar pattern search?
Last edited by ronk on Mon Apr 09, 2012 11:43 am, edited 1 time in total.
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 11:28 am

ronk wrote:
champagne wrote:They all strictly match the exocet definition. It's just that the way to establish the elementary "base to target" links are not complying with your limited definition.

Now would be a good time to re-post your definition. I say "re-post", even though I'm not actually aware of a thorough definition anywhere. Examples everywhere, but complete definition nowhere AFAIK.

David P Bird's recent offering is close, I suspect, but it doesn't fully handle optional candidates and degenerate cases.


not so old

exocet definition

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 11:34 am

David P Bird wrote:
champagne wrote:They all strictly match the exocet definition. It's just that the way to establish the elementary "base to target" links are not complying with your limited definition.

I'll prepare an example to show that

OK, OK. I can't accept that your Exocet definition is complete, but as you are its 'owner' it's your decision.

We therefore have two types of Exocet: those that just comply with your base and target cell conditions and the subset that also comply with the extended conditions I have proposed for manual solvers. How do you want to name them?

DPB


my definition as I posted it recently includes already some limitations.
The extended definition is wider, but my experience is that id does not brings more.

In the chapter dedicated to the extended definition, i'll describe more "partial exocet".

Regarding the name, we have some past examples BUG type 1; BUG type 2 ...
why not exocet type 1
Another name would create some confusion.

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby ronk » Mon Apr 09, 2012 12:43 pm

champagne wrote:The extended definition is wider, but my experience is that id does not brings more.

I can't imagine a workable definition wider than what you have now (quoted below). BTW I actually agree with this definition.

champagne wrote:The reduced definition uses

A base of 2 unassigned cells in the same region (row, column, box)
A target of 2 unassigned cells in other regions

having the following property ...

if for any digit solution of the base
one at least of the target is occupied by the same digit

then the target can not contain any other digit than the base.
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Re: bi bi pattern in hardest puzzles

Postby David P Bird » Mon Apr 09, 2012 12:52 pm

champagne wrote:Regarding the name, we have some past examples BUG type 1; BUG type 2 ...
why not exocet type 1
Another name would create some confusion.

I hate resorting to type 1, type 2 ... categories as it is only storekeepers and those that use them regularly who can remember which way round they go. It would be a waste of time for me to suggest any alternative qualifiers though, as it's an unwritten rule here that they'll never be accepted.

Now a word about the required conditions in a pattern definition for manual solvers. For me the important thing is that they make up yes/no tests that can easily be applied to the candidates in the grid. If all these tests are passed there is no need to do anything else, any eliminations the pattern provides can be made immediately. You have already accepted that your definitions don't do that.

Those that are fed up with me will be pleased to learn that I'm taking a break, so won't be contributing for a while.
David P Bird
2010 Supporter
 
Posts: 957
Joined: 16 September 2008
Location: Middle England

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 1:20 pm

David P Bird wrote:
champagne wrote:Regarding the name, we have some past examples BUG type 1; BUG type 2 ...
why not exocet type 1
Another name would create some confusion.

I hate resorting to type 1, type 2 ... categories as it is only storekeepers and those that use them regularly who can remember which way round they go. It would be a waste of time for me to suggest any alternative qualifiers though, as it's an unwritten rule here that they'll never be accepted.

Now a word about the required conditions in a pattern definition for manual solvers. For me the important thing is that they make up yes/no tests that can easily be applied to the candidates in the grid. If all these tests are passed there is no need to do anything else, any eliminations the pattern provides can be made immediately. You have already accepted that your definitions don't do that.

Those that are fed up with me will be pleased to learn that I'm taking a break, so won't be contributing for a while.


David,

If it's so important to you, take another name, it will not disturb me except that, as i always work on the general concept, I'll continue to use the word exocet sticking to the general definition.

On the other side, I never said that no easy test can be done here to prove the exocet. It requires a PM per digit, but it is never hard to prove.

As i wrote, 4 of these puzzles have a double exocet.

If you are suspecting such a pattern, it is worth to prove it.
The solution is then usually so quick that you get paid back.

It is the case for the puzzle 14808 I analysed.

I'll also have a break in few days, I'll try to put it in a nice form on my website before.

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 1:27 pm

ronk wrote:
champagne wrote:The extended definition is wider, but my experience is that id does not brings more.

I can't imagine a workable definition wider than what you have now (quoted below). BTW I actually agree with this definition.



I studied exocet based on 2 cells of a box not in a mini row
I studied also AAHS with three or four free digits,

either it came to redundancy with the "mini row" form
or it did not give visible solving potential.

But this has been done 2 years ago, I don't remember exactly the results.

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby ronk » Mon Apr 09, 2012 4:15 pm

champagne wrote:As i wrote, 4 of these puzzles have a double exocet.

Interesting! It makes sense to you that a puzzle with two exocets would not be on a list with one exocet? Is this true for all of your double exocet puzzles, or just this handful?

My suspicion is that your revised program is now missing another degenerate case.
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 5:13 pm

ronk wrote:
champagne wrote:As i wrote, 4 of these puzzles have a double exocet.

Interesting! It makes sense to you that a puzzle with two exocets would not be on a list with one exocet? Is this true for all of your double exocet puzzles, or just this handful?

My suspicion is that your revised program is now missing another degenerate case.


my file of puzzles with exocets includes all puzzles with double exocet

I like very much suspicions as soon as I have one example. It is a very fast way to improve the process.

BTW I checked carefully the first 2 puzzles with a double exocet.

Once you have made all possible eliminations based on the double exocet logic
(with here a special property due to the fact that both exocets have one of the target cells sharing the same row/cloumn)


the puzzle is solved using not more that the basic set of rules.

I think (if you did not yet do it yourself) that you'll be interested in seeing how the exocet can be easily established in that situation.

This is a good stuff to build the XSUDO diagram, but you are better than me in that exercise.

I"ll put it in a separate post

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby ronk » Mon Apr 09, 2012 6:05 pm

champagne wrote:
ronk wrote:My suspicion is that your revised program is now missing another degenerate case.
I like very much suspicions as soon as I have one example. It is a very fast way to improve the process.

Digit 9 [edit: in column 2 becomes a hidden single after asserting r9c89=9.]

____Image
Code: Select all
980700600500000000004080050309000004060800200000030000010006000000100026000078100
     12 Truths = {3459C2 345C4 345C7 9N89}
     15 Links = {34r2 3r3 5r4 45r6 3459r9 8n2 7n4 345b9}
[edit3: added puzzle and truth/link logic set]

champagne wrote:Once you have made all possible eliminations based on the double exocet logic
(with here a special property due to the fact that both exocets have one of the target cells sharing the same column)
the puzzle is solved using not more that the basic set of rules.

Yes, for this GP-kz1, which you called 14808, I saw only four strong links (or less) needed in all the remaining moves.

BTW isn't that 14808 number merely an index number, which might change in future publications of the hardest?
Last edited by ronk on Fri May 18, 2012 12:08 pm, edited 3 times in total.
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Re: bi bi pattern in hardest puzzles

Postby champagne » Mon Apr 09, 2012 6:44 pm

ronk wrote:
BTW isn't that 14808 number merely an index number, which might change in future publications of the hardest?


I'll study that post to morrow morning, just a quick answer to that point:

Puzzles are automatically numbered by ACCESS in my data base.

So the sequence ID must not change over time, but reversely, some numbers disappear if I make a wrong entry.
So the highest sequence number is not the number of puzzles in the data base.

this gives a short "nickname" to any puzzle. It is also the short for the puzzle in the "attached tables of the data base".

champagne
champagne
2017 Supporter
 
Posts: 5612
Joined: 02 August 2007
Location: France Brittany

Re: bi bi pattern in hardest puzzles

Postby daj95376 » Mon Apr 09, 2012 10:30 pm

ronk wrote:
champagne wrote:
ronk wrote:My suspicion is that your revised program is now missing another degenerate case.
I like very much suspicions as soon as I have one example. It is a very fast way to improve the process.

Digit 9 is required only in column 2.

Yes, for this GP-kz1, which you called 14808, I saw only four strong links (or less) needed in all the remaining moves.

Hmmm!!! Cells r9c89 are "blue boxes". These cells aren't needed for <9>?
daj95376
2014 Supporter
 
Posts: 2624
Joined: 15 May 2006

Re: bi bi pattern in hardest puzzles

Postby ronk » Mon Apr 09, 2012 11:05 pm

daj95376 wrote:
ronk wrote:Digit 9 is required only in column 2.
Hmmm!!! Cells r9c89 are "blue boxes". These cells aren't needed for <9>?

My too terse statement assumes r9c89=9, just as you did in your single-digit PMs. IOW while there are swordfish for the other digits, the digit 9 case degenerates to a hidden single.
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

PreviousNext

Return to Advanced solving techniques