The hardest sudokus (new thread)

Everything about Sudoku that doesn't fit in one of the other sections

Re: The hardest sudokus (new thread)

Postby tarek » Mon Jul 20, 2020 3:57 pm

mith wrote:And one hard one by q2
I'm glad somebody still has interest in these quicker rating tools.

It was my responsibility to rate puzzles according to q1, q2 and suxratt but now we have a huge list of SE rated puzzles that haven't been screened to update the list in the head post. I'm hoping to update that soon

Tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Re: The hardest sudokus (new thread)

Postby mith » Mon Jul 20, 2020 4:18 pm

I've been using q2 as a filter before running SE, so I have those ratings for my puzzles at least (as well as for some of the database puzzles in minlex form); edited in for the other two. I'm hoping to incorporate suexratt eventually as well.

I've got about 50 new ones requiring at least D*FC+MFC; I'll post that group at some point, once the current run isn't generating them so regularly.
mith
 
Posts: 996
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby champagne » Mon Jul 20, 2020 5:41 pm

mith wrote:A couple 11.7s I found today:

Code: Select all
.....1..2..3....4..2..5.6......6...5.7..2.1..8..9.......48......5...67..9......3. q2=98904 FNBP C21.m/M2.3.56862  ED=11.7/11.7/2.6
.....1..2....3..4...56..7.......2..1.5.7..8..7...4.....69.....7.8.....3.5..9..6.. q2=98842 FNBP C22.m/M2.3.34992  ED=11.7/1.2/1.2


Should be new in the database, unless I have massively screwed something up in my checking (which is entirely possible).

And one hard one by q2:

Code: Select all
.....1..2....3..4...56..7.......2.3..5....8..9...4...1.6..1.....87..6...5..7..6.. q2=99385 FNBP C22.m/M3.608.874  ED=11.4/11.4/4.2

sorry, these are morphs of old puzzles from eleven
champagne
2017 Supporter
 
Posts: 7453
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby mith » Mon Jul 20, 2020 5:44 pm

Hm. Would you mind sending me which ones so I can figure out why I didn't catch them? (I'm using ph_1910; I actually don't see a single puzzle with ED of 4.2, so I'm especially confused about that one.)
mith
 
Posts: 996
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby champagne » Tue Jul 21, 2020 4:26 am

mith wrote:Hm. Would you mind sending me which ones so I can figure out why I didn't catch them? (I'm using ph_1910; I actually don't see a single puzzle with ED of 4.2, so I'm especially confused about that one.)

....56.8...71.....6.....4.......85...3......29...4..6..1.7.......2.....38...9..5. 11.7 11.7 2.6 21 elev 30
..34...8..5...9...6...3...12...6...3.....59.....8...4...2.7....7.....3...16.....7 11.5 1.2 1.2 22 elev 80
.2.4.....4...89........7..4..1..8.6....7....8.3..6.5...6.....1...5...3..9..8....7 11.4 11.4 6.6 22 elev H12

these are SE ratings. With SE ratings

The 0.2 deviation on ER for the second is not common , but a 0.1 has been often seen in the games
I can not explain the deviation on the last puzzle. skfr ED is also 6.6 for you morph
champagne
2017 Supporter
 
Posts: 7453
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby mith » Tue Jul 21, 2020 1:20 pm

Aha, the minlex program I was using apparently had a bug. Thanks! (The rating difference is just from using SukakuE without 1.2.1 mode, apparently. I'll run the puzzles I post in 1.2.1 mode as well for consistency.)
mith
 
Posts: 996
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby tarek » Tue Jul 21, 2020 1:57 pm

mith wrote:Aha, the minlex program I was using apparently had a bug. Thanks! (The rating difference is just from using SukakuE without 1.2.1 mode, apparently. I'll run the puzzles I post in 1.2.1 mode as well for consistency.)

skfr is another screening tool mith as it is the closest thing to a non-java Sudoku Explainer. Champagne is keeping the list on the ratings of v1.2.1 similar to what is done in the Patterns game. There will be a degree of disparity between the ratings of the current Suakau Explainer v1.17.7 compared to v.1.2.1 mode due to the addition of resolutions rules (Even more when choosing the new rating system)

tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Re: The hardest sudokus (new thread)

Postby tdillon » Wed Jul 22, 2020 4:36 am

Hi champagne,

Here's a little island of puzzles to add to the database. These came up during searches for minisat-hard puzzles, which I'll post about separately.

Ratings using: serate 1.4.1 (Sudoku Explainer) 2019-09-22

Code: Select all
5..2.......1.8..6..9...3...3..5......2...9.....6.4...7....1..74......5.6..7....8.        11.8/11.8/10.8
5..2.......1.8..6..9...3...3..5......2...9.....6.4...7....1..74......9.6..7....8.        11.8/11.8/10.8
5..2.......1.8..6..9...3...3..5......2...9.....6.4...7....1..74......5....7...18.        11.7/11.7/11.2
5..2.......1.8..6..9...3...3..5......2...9.....6.4...7....1..74......9....7...18.        11.7/11.7/11.2
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......5.6..7....8.        11.7/11.7/10.8
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......9.6..7....8.        11.7/11.7/10.8
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......5....7...18.        11.6/11.6/11.1
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......5....7...81.        11.6/11.6/11.1
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......9....7...18.        11.6/11.6/11.1
5..2.......4.8..6..9...3...3....5....2.9.......8.6...7....1..74......9....7...81.        11.6/11.6/11.1
5....3.....4.8..6..9.2.....3....5....2.9.......6.4...7....1..74......5.6..7....8.        11.6/11.6/3.4
tdillon
 
Posts: 66
Joined: 14 June 2019

Re: The hardest sudokus (new thread)

Postby champagne » Wed Jul 22, 2020 5:04 am

tdillon wrote:Hi champagne,

Here's a little island of puzzles to add to the database. These came up during searches for minisat-hard puzzles, which I'll post about separately.

Ratings using: serate 1.4.1 (Sudoku Explainer) 2019-09-22


a) Sorry again they all are already in the data base
b) All puzzles of the data base with ER>=11 have been rated using serate (Sudoku Explainer used in the games)to have an homogeneous set of rating. This is a rule that I think good to keep. Only the lowest part of the data base has been loaded with some SKFR ratings.
I don't have free cycles to redo the rating, so I wold like to receive files with the appropriate rating. ,

Here below, the very old puzzles of the data base
Code: Select all
1.......2.3.4...5...6...7...5.8.3.......74......9...8.7.....6...9...8.3...2.....1   11.6   11.6   11.1   21   tax   jpf-04/14/02
1.......2.3.4...5...6...7...5.8.3.......7.......95..8.7.....6...9...8.3...2.....1   11.7   11.7   10.8   21   tax   jpf-04/08
1.......2.9.4...5...6...7...5.3.4......96.........8.4...2...6...3...9.8.7.......1   11.7   11.7   11.2   21   tax   coloin-04-10
1.......2.9.4...5...6...7...5.3.4.......6........58.4...2...6...3...9.8.7.......1   11.8   11.8   10.8   21   tax   coloin-04-10
.......8...6...12....2..6.5..15..9..8....3....4..7....3....8.....21....6.7..4....   11.6   11.6   3.4   21   col   H5
3.......8.7.5...1...6...4...9.2.1.......4.......97..2.4.....3...5...2.7...8.....6   11.6   11.6   11.1   21   tax   tarek-ultra-0313


EDIT only six puzzles, your file contains puzzles with the same canonical form
Last edited by champagne on Wed Jul 22, 2020 5:55 am, edited 1 time in total.
champagne
2017 Supporter
 
Posts: 7453
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby tdillon » Wed Jul 22, 2020 5:44 am

Sorry for wasting time ... I guess I've misunderstood the requirements for canonicalization.

I wasn't sure what form is used for the puzzles in ph_1910.zip, but I was under the incorrect impression that fsss2 --solrowminlex produces a canonical form. So I solrowminlex'd the puzzles in 01_file1.txt, and the "new" puzzles, and checked for any overlap. (actually I didn't recheck this after adding the last puzzle, and I see there *is* overlap for this one).

What tool produces the canonical form you use?
tdillon
 
Posts: 66
Joined: 14 June 2019

Re: The hardest sudokus (new thread)

Postby champagne » Wed Jul 22, 2020 6:18 am

tdillon wrote:Sorry for wasting time ... I guess I've misunderstood the requirements for canonicalization.

I wasn't sure what form is used for the puzzles in ph_1910.zip, but I was under the incorrect impression that fsss2 --solrowminlex produces a canonical form. So I solrowminlex'd the puzzles in 01_file1.txt, and the "new" puzzles, and checked for any overlap. (actually I didn't recheck this after adding the last puzzle, and I see there *is* overlap for this one).

What tool produces the canonical form you use?


For specific reasons, I use a "max lexical form", but the code is derived from the old gsf's min lexical code. as far as I remember, this code in used in gridchecker in the original form.
Using any "valid' canonical form works. If you don't have it, I can load somewhere the original code from gsf (C++ license as below)
Code: Select all
      *               This software is part of the ast package               *
      *          Copyright (c) 2005-2009 AT&T Intellectual Property          *
      *                      and is licensed under the                       *
      *                  Common Public License, Version 1.0                  *
      *                    by AT&T Intellectual Property                     *


What I do to keep the original code and avoid redundancy is the following :

I add the canonical form to the puzzle
I load both in a data base, checking for redundancy on the canonical form
champagne
2017 Supporter
 
Posts: 7453
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby 1to9only » Wed Jul 22, 2020 8:34 am

champagne wrote:the original code from gsf

rjamil already posted here: http://forum.enjoysudoku.com/gsf-s-software-t6284.html#p249786 in multi parts archive files.
User avatar
1to9only
 
Posts: 4177
Joined: 04 April 2018

Re: The hardest sudokus (new thread)

Postby mith » Sat Aug 01, 2020 6:23 pm

Ok, now that I've resolved a couple bugs in my scripts, these should actually be new (compared to ph_1910):

Code: Select all
........1.....2.34..3.5.6.....6..7...7..8....9....1.4..8.5.....3.1....9.4....3..2  ED=11.7/11.7/8.0 (DLFC+DFC) q2=97177
........1.....1..2...34..5...36......574.....8....2....46....7..7.5...6.9...6.1..  ED=11.7/11.7/2.6 (DCFC+DFC) q2=95198
........1.....2.3.....4.5.6..6.5...4.51...7..8..2......6.8...9..7..1.6..9....3...  ED=11.6/11.6/8.0 (DCFC+DFC) q2=98089
mith
 
Posts: 996
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby coloin » Sat Aug 01, 2020 11:11 pm

mith wrote:
Code: Select all
........1.....2.34..3.5.6.....6..7...7..8....9....1.4..8.5.....3.1....9.4....3..2  ED=11.7/11.7/8.0 (DLFC+DFC) q2=97177
........1.....1..2...34..5...36......574.....8....2....46....7..7.5...6.9...6.1..  ED=11.7/11.7/2.6 (DCFC+DFC) q2=95198
........1.....2.3.....4.5.6..6.5...4.51...7..8..2......6.8...9..7..1.6..9....3...  ED=11.6/11.6/8.0 (DCFC+DFC) q2=98089

yes these are new and they are distinct from
these ones posted by ssxsssxs which also are not [yet] in database
Code: Select all
12.3.....6...7..1...4..5...3.....4...6..2..3...8.....9...9....8.7..3..6......25..  ED=11.7/1.2/1.2
12.3.....34..1......5..6...2.....5...1.....2...7..8..9...5..6.7....3..4......79..  ED=11.7/11.7/2.6
coloin
 
Posts: 2487
Joined: 05 May 2005
Location: Devon

Re: The hardest sudokus (new thread)

Postby mith » Mon Aug 03, 2020 4:17 pm

Here's a 21 clue:

Code: Select all
........1.....2.3...4.5.6.7...2.8....7..6...59.........16...4...5..1.7..8..3.....  ED=11.7/11.7/2.6 (DCFC+DFC) q2=97905


And my highest by q2:

Code: Select all
........1..2..3.4..4..5.6.......4.7...41....38...9......17....2.5..6....9.....8..  ED=11.5/11.5/10.0 (DCFC+MFC) q2=99400
mith
 
Posts: 996
Joined: 14 July 2020

PreviousNext

Return to General