The New Sudoku Players' Forum

by **champagne** » Tue May 29, 2012 2:16 pm

..4.....9.1...3.7.6.....2.......8....5..3..8....1.5.....2.....4.7.3...1.9...7.6.. tarx0037,tarek 7.0 SK *BB r46c5 r2c4 r8c6

After SK Loop eliminations:

Code: Select all: *----------------------*----------------------*----------------------* | 3578 238 4 | 25678 158 1267 | 1358 356 9 | >| 258 1 589 | 2469 2469 3 | 458 7 568 |< | 6 389 3578 | 45789 158 1479 | 2 345 1358 | *----------------------*----------------------*----------------------* | 12347 2469 13679 | 24679 2469 # 8 | 13459 2469 123567 | >| 1247 5 1679 | 24679 3 24679 | 149 8 1267 |< | 23478 2469 36789 | 1 2469 # 5 | 349 2469 2367 | *----------------------*----------------------*----------------------* | 1358 368 2 | 5689 158 169 | 7 359 4 | >| 458 7 568 | 3 2469 2469 | 589 1 258 |< | 9 348 1358 | 2458 7 124 | 6 235 358 | *----------------------*----------------------*----------------------* 24 - 69 49 - 26

Compatibility checks show that the base cells for (2469)Jexocet:r46c5,r2c4,r8c6 must hold either (29) or (46)

[/quote]

interesting example where we have the maximum possible effect of the ABI loop.

This proves in once the property I have always seen in the SK loop, one digit from each of the 2 belts in the AALS
The job is not finished, but as you say, we are quickly at the key point

by **champagne** » Tue May 29, 2012 2:23 pm

ronk wrote:
RI <3>r12c7 3r9c7 3r7c1 6r7c3 1r7c8
<13+>r7c1 (3r7c1 3r7c1 6r7c1 3r7c1)
<3+>r4c2 (1r4c2 2|6r4c2 6r4c2 6r4c2)

BTW cut & paste from your link produces two line breaks where there should be one.

typo corrections in progress

UR <3>r12c7 3r9c7 3r7c1 6r7c3 1r7c8
<1>r7c1 (3r7c1 3r7c1 6r7c1 3r7c1)
<3>r4c2 (1r4c2 2|6r4c2 6r4c2 6r4c2)

RI is the French equivalence for UR

regarding the blank lines, this comes from the soft i am using, I don't know so far how to correct it

by **champagne** » Tue May 29, 2012 3:28 pm

a sample file on lot 3

the 2 first cells show the base, the 2 next the 2 target cross line position
eg: first puzzle
base r12c3
target in the rows 4 and 7 (1+2 or 2+2)

just keeping in memory that no check has been done on what should be an AHS

Hidden Text: Show

by **champagne** » Tue May 29, 2012 3:51 pm

a sample file on lot 1

Hidden Text: Show

by **champagne** » Tue May 29, 2012 4:17 pm

a sample file on lot 0

Hidden Text: Show

by **David P Bird** » Tue May 29, 2012 10:32 pm

champagne Firstly let me warn you that the time I'll have available for Sudoku will be limited for the next 2 weeks or so, but I'll be following your posts.

Responding to your other posts on page 7 of this thread:

I'm not clear about what you mean here:

a) In my head, "partial exocet" is a precise concept and here is the wording I have in the text in preparation to come in the "definitions" post

Partial Exocets
We have a partial exocet if the exocet digit per digit rule is verified for 2 or more digits.
In a partial exocet, we have no direct elimination, but the scenario study can be applied to the pairs of digits verifying the rule.
I think I'll add one example as that one from the post "what to do after an exocet has been seen"

This is the first time I've seen the "Exocet digit per digit" rule mentioned.

My best guess is that a base pair and two targets cells have been found but there are no eliminations to make in the target cells. But that doesn't matter, the pattern is still an Exocet and the inference that the target cells will hold the same digits as the base pair still stands. There is therefore no need to call this case a "partial Exocet".

If that's right then we only need to use "partial" to describe if a compatibility check (ABI loops or otherwise) will only eliminate some of the possible options for the base digits. This would avoid the potential confusion between a partial Exocet and a partially successful compatibility check.

I wrote: Reducible = Candidate distributions allow ABI loops to be used
Twin = Target and companion cells contain a hidden pair so may be in either order

I accept my categories are not quite right, but the best time to improve them will be when we analyse the puzzles in the different lots.

champagne wrote:I have no evidence of r8c6 <> 8 so I suspect this is a typo for <9> r9c8.

Not a typo. The base digits will occur in mini-rows r7c123, r8c456, and r9c789.
As r8c4 contains another given, r8c56 must be confined to Exocet digits.

by **ronk** » Wed May 30, 2012 4:03 am

David P Bird wrote:
champagne wrote:Partial Exocets
We have a partial exocet if the exocet digit per digit rule is verified for 2 or more digits.
In a partial exocet, we have no direct elimination, but the scenario study can be applied to the pairs of digits verifying the rule.
This is the first time I've seen the "Exocet digit per digit" rule mentioned. ... My best guess is that a base pair and two targets cells have been found but there are no eliminations to make in the target cells.

Quoting champagne: On my side I am amazed that simple facts can not be understood.

But seriously, by partial exocet he means that at least one digit pair assigned to the base cells causes the same digit pair to be assigned to the target cells. Other digit pairs assigned to the base are permitted to cause assignments in the target different from the base. From a practical POV though, I'm guessing one of the digits in the latter case will always be the same.

by **champagne** » Wed May 30, 2012 5:12 am

David P Bird wrote:champagne Firstly let me warn you that the time I'll have available for Sudoku will be limited for the next 2 weeks or so, but I'll be following your posts.

I'm not clear about what you mean here:
a) In my head, "partial exocet" is a precise concept and here is the wording I have in the text in preparation to come in the "definitions" post

Partial Exocets
We have a partial exocet if the exocet digit per digit rule is verified for 2 or more digits.
In a partial exocet, we have no direct elimination, but the scenario study can be applied to the pairs of digits verifying the rule.
I think I'll add one example as that one from the post "what to do after an exocet has been seen"

This is the first time I've seen the "Exocet digit per digit" rule mentioned.

My best guess is that a base pair and two targets cells have been found but there are no eliminations to make in the target cells. But that doesn't matter, the pattern is still an Exocet and the inference that the target cells will hold the same digits as the base pair still stands. There is therefore no need to call this case a "partial Exocet".
.

here is the generic definition of an exocet (page 1 of that thread)

The reduced definition uses

Code: Select all: 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 (whatever is the process used to prove it.)

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.

The logic is trivial and is just requiring that the puzzle has at least one solution.

Code: Select all: The solution of the base require 2 different digits The target must contain the 2 digit and this forces each cell of the target to have one of them So there is no room for other digit.

you can see here the digit per digit condition in red.

the condition can be valid for 2 of the 3 digits in a 3 digits base and 2 or 3 digits in a 4 digits base;

The corresponding pairs can be processed as pairs of a valid exocet (one or 3 pairs).
The pattern is not an exocet, I name it a "partial exocet"

by **David P Bird** » Wed May 30, 2012 8:00 am

champagne Once again it seems that our differences arise from our different interpretations of what a "pattern" is. I also appreciate you are working in a second language.

Using the Exocet/JExocet compromise has allowed us to separate the discovery processes but discuss them together when considering how the inferences they provide can be used.

If we want to continue discussing them together, then the terms we use should be appropriate to both.

Your definition of what an Exocet is doesn't make any mention of how many options there may be for which two of the base digits can be true. The difference between an Exocet and a partial Exocet therefore appears to be an afterthought (and I would need case by case examples to understand precisely what you mean).

Whatever you could decide about that would not carry over to the JExocet pattern which simply would be true if the pattern requirements were met. For example, ask yourself what a partial Swordfish would be.

Now the question in my mind is how does using "partial" help our discussions to any great extent? At the moment I can't see an overwhelming need for it, and it just leads to potential confusion.

So to use your expression :-)

how strongly are you glued to wanting to talk about "partial Exocets" ?

by **champagne** » Wed May 30, 2012 8:36 am

David it's clear that you are working on patterns that a player can easily detect.
No problem , but I am on a wider scope.

let me just comment one or two points

David P Bird wrote:
Your definition of what an Exocet is doesn't make any mention of how many options there may be for which two of the base digits can be true. The difference between an Exocet and a partial Exocet therefore appears to be an afterthought (and I would need case by case examples to understand precisely what you mean).

I don't catch exactly your point. It's clear that

a 3 digits base has three possible pairs
a 4 digits base has 6 possible pairs

to stay within the size of interest

My solver (2 years old) works with "partial exocets" so it's not at all an afterthought

It is sure that only one pair is valid, the game is to establish which one.

David P Bird wrote:
Now the question in my mind is how does using "partial" help our discussions to any great extent? At the moment I can't see an overwhelming need for it, and it just leads to potential confusion.

So to use your expression how strongly are you glued to wanting to talk about "partial Exocets" ?

Let me just answer in a different way

I agree that the most common exocet, the Jexocet, is worth catching in priority the efforts (what is done), but I don't want to be glued to restrict the search and the logic to that one.

Moreover, as soon as the Jexocet will have been deeply studied, there will not be much to discuss about it. IMO, we are not so far from that point.

Players will just have to apply it.

Look at the SK loop. It's for long a mature concept. Jexocet should be soon in the same status

by **David P Bird** » Wed May 30, 2012 11:03 am

Champagne We not only have terminology problems to overcome but also our language barrier.

Let me put it more plainly: the term partial JExcocet is meaningless. A pattern is either a JExocet or it isn't.

Any pattern that would comply with the pattern conditions if a particular placement or elimination could be made is an Almost JExocet.

Deductions can be made using Almost JExocets showing that if an assumption was true a JExocet inference would produce the same result in a cell as would happen as if it was false (there is still much to explore here).

Does that equate to your "partial Exocet"?

Alternatively we can say a fully populated Swordfish has 9 instances of a candidate in the pattern cells providing 6 possible solutions. A partially populated Swordfish may only have 2 or 4 possible solutions. Is that closer?

If you say that neither of these is equivalent and your partial term only applies to the templating approach you use, then we should split the Exocet and JExocet threads.

I'm afraid putting things another way won't help us resolve the basic questions much, although I appreciate the points you make.

To rephrase my question, how essential is it to you to define the "partial Exocet" condition? Can't you use another form of words when necessary to overcome the problem? After all, I don't think it will happen that often.

When I used the term "afterthought" I was referring to how precise your definition is. If the conditions you stipulate sometimes find Exocet patterns and sometimes don't but find partial Exocets instead, then they aren't precise enough.

Believe me, I'm trying to be constructive here.

David

by **champagne** » Wed May 30, 2012 1:46 pm

David

I see our roads diverging but it's not a problem. I'll continue to help you as much as possible in the Jexocet search;

Some preliminary statements.

I don't remember having used the expression "partial Jexocet".
That thread is titled "exotic patterns a resume" and I opened it to exchange on all tools that can be used in hardest puzzles.
I hope it will stay true and I would not see it limited to the Jexocet.

I suspected a while that I had been caught by a "false friend" . I took my dictionary but it says me that partial is the right translation for what I describe. Reversely, I don't like "almost" as you will see.

I see no similarity between the "partial exocet" and the situation you describe for the swordfish.

I have serious reasons to stay interested by "partial exocet".
In my solver, the highest step of logic is the use of AAHS and AALS in the same way others are using ALS and AHS.
Any elimination of a pair of candidates in such objects creates new weak links. So I like very much any tool leading to such eliminations.

Now back to another rewording of the definition

let's assume we have a base defined as for an exocet (4 digits abcd) and a target t1;t2
if
a true in the base forces t1;t2 contains a
b true in the base forces t1;t2 contains b
then ab true in the base forces ab true in the target

this is the generic rule leading to an exocet if this is true for each digit

We can then apply all the logic specific to one pair in an exocet pattern to see if that pair is valid.

we name it as we want, but I prefer "partial exocet" to "almost exocet" in that case where one pair out of six follow the logic
Any other name could be used

we have usually many more "partial exocets" in a PM than exocets.

let's now assume that we have also

c true in the base forces t1;t2 contains c

then the logic valid for a pair in an exocet can be applied to the pairs ab ; ac ; bc 3 pairs out of six.

You can not be interested in that, which is ok, it does not makes that logic false.

by **ronk** » Wed May 30, 2012 3:29 pm

champagne wrote:Now back to another rewording of the definition

let's assume we have a base defined as for an exocet (4 digits abcd) and a target t1;t2
if
a true in the base forces t1;t2 contains a
b true in the base forces t1;t2 contains b
then ab true in the base forces ab true in the target

this is the generic rule leading to an exocet if this is true for each digit
We can then apply all the logic specific to one pair in an exocet pattern to see if that pair is valid.
...
let's now assume that we have also

c true in the base forces t1;t2 contains c
then the logic valid for a pair in an exocet can be applied to the pairs ab ; ac ; bc 3 pairs out of six.

You now have the beginnings of a real definition, but there are a couple of problems still:

There is nothing to indicate the possibility of the same digit ultimately being in both t1 and t2, as in puzzle #12177
After stating the base has "4 digits abcd", 'd' is never mentioned again

BTW what means ';'? Always logical OR, always logical AND, always a list separator, or sometimes any one of these? To be unambiguous, '|', '&', and ','' respectively, would be better.

by **champagne** » Wed May 30, 2012 3:58 pm

ronk wrote:You now have the beginnings of a real definition, but there are a couple of problems still:
There is nothing to indicate the possibility of the same digit ultimately being in both t1 and t2, as in puzzle #12177
After stating the base has "4 digits abcd", 'd' is never mentioned again

I had not the feeling to add much here to my previous writing , but it appears it had to be written.

Your 2 questions seem to be in fact more or less the same;

We can in that situation tell something for the pairs verifying the rule, but nothing can be said for the rest.

In puzzle 12177 the base has the digits 1236 but the proof in the target fails for digit 2.
For any pair in the base containing the digit 2 we can only say that one of the target will be occupied by the other digit. We have no indication for the other target

In case only 2 digits would verify the exocet elementary rule, (a & b in the wording) we could say nothing for the complementary pair (cd)

'd' is not mentioned again just because if each of 'abcd' would verify the elementary rule, it would be an exocet.
(and because I said nothing for pairs that don't verify the rule)

ronk wrote:BTW what means ';'? Always logical OR, always logical AND, always a list separator, or sometimes any one of these? To be unambiguous, '|', '&', and ','' respectively, would be better.

surely my writing can be improved. Thank you to point where you see ambiguity, i'll try to improve the situation in the future and I'll edit the text if necessary

by **ronk** » Wed May 30, 2012 7:06 pm

champagne wrote:
ronk wrote:You now have the beginnings of a real definition, but there are a couple of problems still:
There is nothing to indicate the possibility of the same digit ultimately being in both t1 and t2, as in puzzle #12177
After stating the base has "4 digits abcd", 'd' is never mentioned again
In puzzle 12177 the base has the digits 1236 but the proof in the target fails for digit 2.
For any pair in the base containing the digit 2 we can only say that one of the target will be occupied by the other digit. We have no indication for the other target

In case only 2 digits would verify the exocet elementary rule, (a & b in the wording) we could say nothing for the complementary pair (cd)

Below is (I think) a "partial exocet" for puzzle #12177, simpler than the one on your web page. Note the targets are r4c2=1236 and r7c1=1236 just as if it were a "full exocet." This can only happen because there is an indication for the other target. Evidently this cause has yet to be formulated in the definition, so I'll try to do so.

___

Puzzle & logic set: Show

The New Sudoku Players' Forum

Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume

Re: Exotic patterns a resume