The New Sudoku Players' Forum

[champagne]

Here is a diagram showing ...


The missing "bit of detail", in his description, are how the rest of the diagram -- the D,E,F,U,V,W truths, and thier links into each other and the candidates in the A,B,R,S,T truths -- conspire to make it impossible to have any of these situations occuring in a solution:

• TGT1 and one of {Rx,Ry} and one of {Sx,Sy} true
• TGT1 and one of {Rx,Ry} and one of {Tx,Ty} true
• TGT1 and one of {Sx,Sy} and one of {Tx,Ty} true

He leaves TGT1 out, above. I'm not sure whether he's assuming that its presence is implied, or saying that it doesn't need to be there.

Hi blue,

That diagram is very close to the one produced by Allan Barker himself, and your conclusion reflects the difficulty at that time to explain why it worked.

In Allan general construction, he had in theory to find the rank 0 area of the diagram but the diagram has 6 triple points (link type). The best clue he gave was that these triple point where linked (2 groups of three).

It is just to have a simpler explanation that the exocet logic came.

[denis_berthier]

1) you got confused by your "line" vocabulary and your elimination structure doesn't correctly represent the basic Exocet data: you don't need R, S and T "lines" (corresponding to the three columns in David's initial post). But, for each column ci = c1, c2, c3 and each base digit nj = n1, n2, n3, you need a cinj CSP-variable (2D-cell), as explained in my post here: http://forum.enjoysudoku.com/pattern-based-classification-of-hard-puzzles-t30493-69.html

Here is a diagram showing how logel's "rule fragment" fits into a general 3-digit JExocet pattern.
[...]
The top 3 rows, are column truths for one of the "target" S columns.
The bottom 3 rows, are column truths for the other "target" S columns.
The center row holds the 3 [added by DB] column truths for the S column that intersects the base cell box.
The rows on either side of that, are the base cell truths.

At this point, it may be useful for clarity to notice that we are saying the same thing in different terms.

[blue]

1) you got confused by your "line" vocabulary and your elimination structure doesn't correctly represent the basic Exocet data: you don't need R, S and T "lines" (corresponding to the three columns in David's initial post). But, for each column ci = c1, c2, c3 and each base digit nj = n1, n2, n3, you need a cinj CSP-variable (2D-cell), as explained in my post here: http://forum.enjoysudoku.com/pattern-based-classification-of-hard-puzzles-t30493-69.html

(...)
At this point, it may be useful for clarity to notice that we are saying the same thing in different terms.

Almost the same, but quite: logel's R, S and T aren't "one for each column", but "one for each digit" (in one target cell column).

Regards,
Blue.

[denis_berthier]

Almost the same, but quite: logel's R, S and T aren't "one for each column", but "one for each digit" (in one target cell column).

Actually, in his graph, R, S, T can be any "line" (i.e. 2D-cell).
The point is, we don't need only 3 but 9.

[blue]

Actually, in his graph, R, S, T can be any "line" (i.e. 2D-cell).
The point is, we don't need only 3 but 9.

Well, if they can be "any line", then it isn't necessarily an exocet, and we need some undetermined number in addition to the 3(+2).
When it's an exocet, then R,S,T are what they are (and not what you thought).
That's why I posted the full pattern. It's silly to argue this.

Best Regards,
Blue.

[champagne]

I started the generation of puzzles with no filter using the data base of potential hardest. It could be interesting to give the criteria for our "grey zone";

I don't have anything to suggest, unfortunately -- other than the obvious, that they shouldn't be solvable with singles, line-box, naked/hidden subsets, and basic fish.

Well, no ...

Including single digit template eliminations in the list, would cover "grouped candidate" and finned fish eliminations too.
I have template code, but nothing guaranteed to find all (complex, single digit) fish eliminations.
That may be going too far, though.

To avoid puzzles that are too complicated, it would be nice if they were solvable with SE techniques that exclude the "Dynamic (+)" class. Can you arrange for that with 'skfr' ?

Best Regards,
Blue.

I have now some ideas about volumes, so I can comment on that

As a rough estimate, the generation should bring more than 300 millions puzzles passing the serate "aligned triplet" step
I think that in such a study (looking for appearance of the exocet when the SER rating is going done) we have to keep all puzzles that are not in the "potential hardest" data base.

So I split the file in three blocks

puzzles below SER 6.2 (kept for pat use)
puzzles SER 6.2 to 7.5 what we could call a green area knowing that with the seeds we use, it will be strongly biased

puzzles SER 7.5 to "potential hardest" border.

I have no problem to look for exocets in a file of 300 millions puzzles (likely some minutes), sharing it is another point.
I can easily take a sample (one every ..) to reach the point where I can load it in the skfr project.

[denis_berthier]

A,B,R,S,T can be lines of any type (cell, row, col, box).

As there are 81 cells and 9 rows, 9 columns and 9 blocks, this should be enough to suggest that the 4 notions should never appear in the same sentence with similar roles.
So, the only way to interpret "line" consistently with logel's vocabulary in his definition of an "elimination pattern" is as "rc-, rn-, cn- or bn- cell" (which Allan renamed Truth).

Exocet seem to consider A,B of cell type and R,S,T of row/col type only[...]
The pattern kernel consists of 5 base lines.

Here, the confusion between "lines" in logel's sense and "lines" in David's sense fully appears.

When it's an exocet, then R,S,T are what they are (and not what you thought).

I don't know what you mean by "what they are".
The fact is, you had to replace 3 columns by 3x3 column-number pairs, i.e. by 9 cn-cells, in exactly the same way I did it.

[blue]

Hi Denis,

A,B,R,S,T can be lines of any type (cell, row, col, box).

As there are 81 cells and 9 rows, 9 columns and 9 blocks, this should be enough to suggest that the 4 notions should never appear in the same sentence with similar roles.
So, the only way to interpret "line" consistently with logel's vocabulary in his definition of an "elimination pattern" is as "rc-, rn-, cn- or bn- cell" (which Allan renamed Truth).

I agree with the last sentence. I think you're blowing smoke with the first.
The "lines of type row", are associated with 'rn' cells, and Alan's "row truths", and there are 81 of them.
I agree that the terminology it can be confusing, but for now, I'm willing to read between the lines as necessary.

When it's an exocet, then R,S,T are what they are (and not what you thought).

I don't know what you mean by "what they are".
The fact is, you had to replace 3 columns par 3x3 column-number pairs, i.e. by 9 cn-cells, in exactly the same way I did it.

Actually, I didn't replace anything, I kept logel's labels, (A,B,R,S,T), and added in the missing "truths & links" for a JExocet.
About R,S and T being "what they are" in a JExocet -- I should have said JExocet, not exocet: the TGT candidate links to one candidate each, in the R,S and T lines. That means R, S and T could only be lines/truths for the three 'cn' variables associated with a (single) target column, with the links being between candidates in a target cell.
After that, assuming we don't have any problem understanding that the A and B "lines" hold the candidates in each base cell (once logel has said that they are of type "cell"), everything else falls into place when we look at the other links: the R2,S2,T2 candidates are base digit candidates in the row containing the base cells, and RX,SX,TX can only be the "1 or 2" candidates in the 'S' cells (... the S cells in the column containing the target cell).

Regards,
Blue.

[blue]

Hi champagne,

I'm sorry, but I'm rapidly losing interest in the topic of exocets.
I'm going to have to back out of our proposal.

Again, sorry,
Blue.

[denis_berthier]

Hi Blue,

A,B,R,S,T can be lines of any type (cell, row, col, box).

As there are 81 cells and 9 rows, 9 columns and 9 blocks, this should be enough to suggest that the 4 notions should never appear in the same sentence with similar roles.
So, the only way to interpret "line" consistently with logel's vocabulary in his definition of an "elimination pattern" is as "rc-, rn-, cn- or bn- cell" (which Allan renamed Truth).

I agree with the last sentence. I think you're blowing smoke with the first.
The "lines of type row", are associated with 'rn' cells, and Alan's "row truths", and there are 81 of them.
I agree that the terminology it can be confusing, but for now, I'm willing to read between the lines as necessary.

This discussion started when I tried to estimate the complexity of logel "pattern" for J3-Exocet.
My main point was that the "5 lines" was not correct (11 is better but still not enough); that's a large gap and IMO it requires a little more than reading between the lines. But we agree on the essential.
My secondary point was to draw logel's attention to his vocabulary. I've always been very attentive to using the proper words; I don't think this is blowing smoke.


Everyone, I've been using J3-Exocet, J4-Exocet several times to specify the number of base digits when needed. Any reaction ?

[champagne]

Hi champagne,

I'm sorry, but I'm rapidly losing interest in the topic of exocets.
I'm going to have to back out of our proposal.

Again, sorry,
Blue.

that's fine for me.

I am interested in seeing what happens so I'll continue on my side (I have covered close to 20% of the creation of puzzles). It will be easier without trying to share the file

Best regards

champagne

[champagne]

A first indication on what happens on Jexocets for lower ratings

I have analysed a lot of 31 millions puzzle in the area SER 7.6 -> not "potential hardest" (A)
and a lot of 12 millions puzzles in the area SER 6.2 to 7.5 (B)

I got the following results

Jexocets/total number of puzzles
(0) 76% (last status of the "potential hardest" file)
(A) 13%
(B) 3.6%

Jexocets 3 digits / Jexocets 4 digits

(0) 0.35%
(1) 60%
(2)105%

this corresponds more or less to puzzles generated by the first 10% of the file of "potential hardest"

[eleven]

Oh, much more than i had expected. Did you look for them after applying basics ?

[champagne]

Oh, much more than i had expected. Did you look for them after applying basics ?

Here, the search starts after basics limited to triplet and swordfish

I am not really surprised by the results (may be I could expect less in the "green area")

I should be in a position to deliver the stats for the full lot by the end of the next week. I am very busy till to morrow late.

[eleven]

Don't hurry
I am also very interested, how much they could progress the puzzle, i.e. the differences in ER before and after applying them.