The New Sudoku Players' Forum

[tarek]

There is no standard for the representation of a pencilmark grid in a line format. The 729-character line representation has benefits but having delimiters will certainly shorten the lines in most cases. In the Sukaku explainer project we are working on at the moment we are using the 729-character line and the multi line pencilmark grid. So it appears that the 729-character is becoming more of a standard as more people use it.

If you are using delimiters especially spaces may have issues with sharing and posting them as you can’t control how the message board formats them. The 729-character line has less problems in the way.

We are currently working on a project on github to improve Sudoku explainer and allow rating of Sukaku. The puzzles I posted remain challenging without a rating as we speak because of the time required to process even a single elimination

Tarek

[Pat]

Is there any problem with spaces in a line?

i would use a colon (":")
to separate the cells

( as used in URL for dailysudoku )

[tarek]

Is there any problem with spaces in a line?

i would use a colon (":")
to separate the cells

( as used in URL for dailysudoku )

I’m hoping that the next release of Sukaku explainer will allow Pasting into GUI these multiple formats which in theory should allow pasting any pencilmark grid (sudoku or Sukaku) and continue solving/analysis from that point in the pencilmark grid

[dobrichev]

Another question about these puzzles: has anyone finally found a solution to compute their SER?

No.

SE had one elimination technique at the bottom which failed to rate them. All known vanilla puzzles don't use it.
This enforced us to make changes in the techniques beyond the ones that solve vanilla sudoku.
So far the extension gives some more eliminations but still fails to solve them.

In this context, at the moment we have an extension to SE which gives some number for a puzzle, but this shouldn't be considered as "SE rating" when technique beyond the old one is in use.

The 999_Springs example uses only old techniques and its result can be considered SER.

Technically SE now accepts 729-char pencilmarks as input, can dump the solving steps, and does ~17 hours processing on a 4-core machine to ensure the puzzle is beyond the implemented techniques.

[blue]

Hi Denis,

A more interesting question for me would now be: can one find a Sukaku not in B7B?

Tarek's "singles backdoor size 6" puzzles are all "TE[singles](4)" -- "in BBBB, but not in BBB".
There were two more puzzles of interest, mentioned here:

RuudLeft4: in BBB, but not in BB
tareknasty1: in BBB, but not in BB

The others (with high SE ratings) were "TE[singles](2)", and so in BB.
I can't say whether they're in B7B.

Cheers,
Blue.

[denis_berthier]

Hi Blue,

A more interesting question for me would now be: can one find a Sukaku not in B7B?

Tarek's "singles backdoor size 6" puzzles are all "TE[singles](4)" -- "in BBBB, but not in BBB".
There were two more puzzles of interest, mentioned here:

RuudLeft4: in BBB, but not in BB
tareknasty1: in BBB, but not in BB

Thanks for your answer; it totally solves the question for me. Having Sukakus in T&E(4) or even T&E(3) - i.e. intractable - is enough to show that Sukakus can be terribly harder than Sudokus.

The others (with high SE ratings) were "TE[singles](2)", and so in BB.
I can't say whether they're in B7B.

If there are Sukakus in B3B, I have no doubt that there are also Sukakus in BpB for some p > 7. Unless anyone is interested in knowing about these two particular ones, I see no point in checking.
I fear this also puts an end to my short-lived interest for the hardest Sukakus.

[checkvilelurl]

i am Checkvil. This i my login on this forum. it is good ?

[ssxsssxs]

new 11.7

Code: Select all

120300000600070010004005000300000400060020030008000009000900008070030060000002500  ED=11.7/1.2/1.2

[tarek]

This puzzle solved successfully in the Sukaku explainer and would get the following rating combination:

Code: Select all

11.7/1.2/1.2 DCFC+DFC/HS/HS

with the hardest step at

Code: Select all

+-------------------+-------------------+-------------------+
| 1     2     579   | 3     4689  4689  | 6789  4589  4567  |
| 6     3589  359   | 24    7     489   | 2389  1     2345  |
| 789   389   4     | 1268  1689  5     | 23678 2789  2367  |
+-------------------+-------------------+-------------------+
| 3     159   127   | 15678 15689 16789 | 4     2578  1267  |
| 4579  6     1579  | 4578  2     14789 | 178   3     157   |
| 2457  145   8     | 14567 1456  3     | 1267  257   9     |
+-------------------+-------------------+-------------------+
| 245   1345  136   | 9     1456  167   | 1237  247   8     |
| 24589 7     1259  | 458   3     148   | 129   6     124   |
| 489   13489 136   | 14678 1468  2     | 5     479   137   |
+-------------------+-------------------+-------------------+


You can pick a more difficult one as well

[champagne]

Having received a heavy contribution from paquita, I prepared a new update of the data base

as for all new files , that update is in my "google drive space" with the name ph_1910.zip

the link to that storage place is here

The file contains now 3 049 068 puzzles (2 135 371 in the former release dated May 2019)

On interesting fact is that paquita produced many high ratings in the area >24 clues.

The readme file included in the zip file have more details.

[denis_berthier]

Having received a heavy contribution from paquita, I prepared a new update of the data base
The readme file included in the zip file have more details.

When I'm aware of an update of the hardest, I test the BpB classification of the highest rated puzzles.

The 1910 Readme file says:

Code: Select all

The count per er is the following
er   count
11.9        9
11.8       59
11.7      124

The 1905 Readme file said:

Code: Select all

The count per er is the following
er   count
11.9     9
11.8    54
11.7   119  +3


I conclude that the difference is:

Code: Select all

er     count
11.9    0
11.8    +5
11.7    +5

However, I could find only 1 new 11.8 and 4 new 11.7, namely:

Code: Select all

98.7.....76..8......5......8..9..4....3....2......1..5...46.7......9.6.2.....2.34;11.80;11.80;2.60;PAQ;2019_05_01;2346393;23;
98.7.....6...9.5....4..3...7..9..6......7..42.....1....6.5..8....2.1...3.......1.;11.70;11.70;3.40;PAQ;2019_08_05_a;2487227;22;
..5.69...67.8.....8.....5...4.....3...85..9......2...1..79..6......1..4......3..2;11.70;1.20;1.20;GP;13_11_gexo;1209207;22;*
98.76.5..4.53..6....2..9....5.6..3......4...........1..3..7..58...5...63......7..;11.70;1.20;1.20;PAQ;2019_08_05_a;2489141;24;
98.76....7....59...45.3.6..4....6....6.59.4....93.......4..35.....9...2.....1....;11.70;1.20;1.20;PAQ;2019_08_05_a;2488918;25;


Am I missing anything?

[champagne]

The 1910 Readme file says:

Code: Select all

The count per er is the following
er   count
11.8       59
11.7      124

The 1905 Readme file said:

Code: Select all

The count per er is the following
er   count
11.8    54
11.7   119


I conclude that the difference is:

Code: Select all

er     count
11.8    +5
11.7    +5

However, I could find only 1 new 11.8 and 4 new 11.7,
Am I missing anything?

Digging in the past is time consuming and not always possible. I'll try to do it later.
Just a quick answer for the 11.7 not from Paquita.
This is an adjustment of the data base reintroducing an old puzzle killed by error.

[International_DBA]

What is the copyright, if any, on these puzzles?
Do you mind if I post a few on my blog?
I would say where I got them from with a link back to your group from my blog and also my Facebook group, which has almost 7000 members.

[champagne]

This data base is shared to give samples of "potential hardest puzzles".
I do my best to give the reference to the first known morph of a given puzzle.
Using a puzzle of the collection is totally free. No supplier asked for any form of copyright.
As usual, it is fair to indicate the source of a puzzle with special properties if it is taken out of an external collection.

[ssxsssxs]

Is this 11.7 not a new puzzle? What is the isomorphism puzzle in ph_1910.zip