The New Sudoku Players' Forum

[Mauriès Robert]

Hi Champagne,
You write:

Here no problem (although, again, I prefer the hidden pair 15r13c5 in the logic). I thought that this was an elimination of the anti track.

There are several ways to make these eliminations at once, I gave you two, here is another one based on the hidden set E=15r13c5.
The anti-track P'(E) is invalid because the elimination of E would lead to a contradiction in B1. From then on we can say that the tracks P(5r1c5) and P(15r3c5) are conjugated, which makes it possible to eliminate, by interaction of the tracks, all those candidates which do not belong to either of these two tracks.
Having said that, I don't know if these logics are TLGO!
Robert

[yzfwsf]

Hi champagne,
Do you have any new progress on this issue? I've been following this topic, thank you.

[champagne]

Hi champagne,
Do you have any new progress on this issue? I've been following this topic, thank you.

Hi yzfwsf,

I have to answer, “yes and no”, announcing a long development.

You gave several examples of “small TLG0 logics, and I wanted to explore this field.
I have leading indicators for the multi floors patterns, but nothing for MSLS

IMO, it is better to start with easy cases and then to increase the difficulty, so I decided to explore the “low size” side.

I implemented a truth links solver “as of Allan Barker” to try to work out leading indicators,
I implemented solving rules “in Allan Barker mood” for basic moves (naked and hidden …; XY wing, XYZWing; XWing…) with as leading indicator the solving potential of a unit, of a band ..

This is tested and works well, although I did not make benchmarks against the previous solving approach.

I then tried to implement the Fish Guide. This is what I am doing slowly for the time being. My surprise, reading the fish guide has been to see how it was close to what I wrote, not only on the TLG0 approach, but also on the TLG-1 plus uncovered cases.

I have seen also (if I got it) the same conjecture as mine: a valid sudoku cannot produce a rank -1 logic.

I thought a while of a “generic implementation” of the fish guide, but the number of possibilities to explore is already very big (n truths out of 9 rows + 9 column + 9 box), so I started an implementation “on identified patterns”.

I also as you know open the door to a “tlg rank 0 logic with a triple point”.

Where I am,

I think better to first implement the rest of the fish guide and to test it
Then to continue with small pattern in the families shown by you
And at the end, to try to apply this to wider scopes.

As I am now an old man working slowly with others constraints; I can not say how long it will take before I am in a position to improve the search for big TLG’s

[yzfwsf]

New ones

msls-db3.png (36.04 KiB) Viewed 1143 times

msls-db32.png (37.42 KiB) Viewed 1143 times

[Hajime]

What is the best way to choose the home set and away set in the MSLS method

[yzfwsf]

I found this irregular shape of MSLS using Death Blosoom's code. For the regular shape of MSLS, I detected whether the rank of the structure formed by the intersection of m rows*n col is zero, and additional cells may be added during the detection. Therefore, I did not use the concept of home set and away set.

For Death Blosoom Extension Type:
http://forum.enjoysudoku.com/death-blossom-complex-type-t38421.html

[StrmCkr]

3trues 15b2,6n5 \ 3links 15c5,3n6

rank0.png

is also a M-Ring or Almost locked Candidate

the new death-blossoms converted to msls i am definitely interested in well done on the your new code and upgrades to it.

Code: Select all

+----------------------+------------------------+----------------------+
| 12789  3      14897  | 5         (278)  (78)  | 6       147    1248  |
| 1278   128    6      | (2378)    4      9     | 1237    5      1238  |
| 5      28     478    | (23678)   1      36-78 | 2347    347    9     |
+----------------------+------------------------+----------------------+
| 169    7      159    | 29-36     2356   356   | 123459  8      12345 |
| 1689   1589   2      | 4         35678  35678 | 13579   1379   135   |
| 3      4      589    | 2789      2578   1     | 2579    79     6     |
+----------------------+------------------------+----------------------+
| 4      (159)  (1579) | (367)     356-7  2     | 8       136-9  135   |
| 278-9  28-59  3      | 1         5678   45678 | 459     469    45    |
| (18)   6      (158)  | (38)      9      345-8 | 1345    2      7     |
+----------------------+------------------------+----------------------+

{cells are in 0..80 format left to right top down}
Aals - 2rc
set a) 36789 @ 57,75
set b) 456789 @ 67,68,69,70,71
set C ) 23678 @ 4,5,12,21
RC 3,6,7,8 ->

30 <> 3,6
61 <> 9
63 <>9
64 <> 5,9
77 <> 8

[yzfwsf]

to " yzfwsf":
is it a standard tool within your solver of a separate approach?
and this pushes me to think of a similar search.

Now I am implementing another zero rank search engine Blossom Loop in my solver.

6 Truths/13 Eliminations