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 Paquita » Tue Nov 20, 2018 9:32 am

After a rocky start I now am busy trying to find hard sudokus.

So far, I found this (SE rates)

98.7.....7...6.9....5..9.7.5....6.8...4.3.6.....2....1.5...7.4....6....3....1.2.. 11.7;11.7;6.6
Thanks to champagne for learning me some things I needed, like the canonical form.
I also found some with probable lower rates, that I yet have to rate with SE, those may follow in a later post
Paquita
 
Posts: 3
Joined: 11 November 2018

Re: The hardest sudokus (new thread)

Postby champagne » Thu Dec 06, 2018 11:08 am

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_807.zip

the link to that storage place is here


The file contains now 2 048 264 puzzles (1 959 402 in the former release)

The main facts in between are:
. the cleaning of wrong entries out of files sent by David Filmer in a format open to manual errors in the updating process (I fell in the trap)
. entry of a new supplier paquita. Again, some difficulties at the start to be in the right format, but the 133 added puzzles have been carefully checked.
paquita additions rate 11.4 to 11.8, (1x11.8 2x11.7).
. several additions out of private cache in the recent pattern games (Dobrichev, Robbie, Papy999 and me)
. update of the transfer of games puzzles
. Addition in the 26 clues area made by me





the count per clue in the data base is now the following
Code: Select all
clues   CompteDeclues
20   79
21   10960  +18
22   88276  +237
23   393651 +597
24   704724 +1893
25   445719
26   404613


The 24 clues area has been reasonably searched,
a lot of work has still to be done with 25 26 clues to reach the same level.



The count per er is the following
Code: Select all
er   CompteDeer
11.9   9      -1
11.8   54     0
11.7   116    +2
11.6   252    +62
11.5   614    +32
11.4   1505   +79   
11.3   8907   +307


The top entries have been cleaned from wrong additions and "paquita" added new puzzles

and the combined table is the following

Code: Select all
clues   20    21     22    23      24      25      26
11.9    1      5      3         
11.8    1     15     29     9         
11.7   31     51      9    16       8       1
11.6    1     38     52    51      87      17       6
11.5          46    101   159     177      90      41
11.4    1     30    305   396     361     173     140
11.3    1    629   1123  2195    2416    1109    1434
11.2    6    991   2721  5507    6707    3004    3459
11.1   12   1267   5106 16430   23194   11861   10532
       21   3148   9493 24759   32958   16262   15613




Note: The file is not yet seen in the google drive, I try to understand why
EDIT solved, once more a silent lock of a process by a pending windows 10 update
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby champagne » Mon Dec 10, 2018 1:48 am

Warning for the use of the last PH files:

Testing the brute force on the files recently uploaded I have seen a problem in the canonical values.
The main file "01 file" is correct,
The canonical form (02index) is wrong(1% of the content), likely in lots extracted on my side from ACCESS (I know how this can appear)
I'll clean the data base and produce a next update beginning of 2019.
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby champagne » Mon Jan 07, 2019 7:11 am

A announced earlier 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_1901.zip

the link to that storage place is here


The file contains now 2 105 149 puzzles (2 048 264 in the former release done one month ago)

The main reasons for this new release after one month are the following:

a) By chance, doing a test on the brute force, I discovered that an old lot was processed with a text delimiter causing a wrong canonical form.
I discarded the corresponding records and added the proper records. Meantime, some morphs of the old puzzles had been added, so the redundant puzzles disappeared.

b) paquita, the new supplier came with puzzles rating on the high side, it was interesting to publish them before my 3 months interruption of this task.

And I added some puzzles with 26 clues, the area where I searched recently.

Due to the redundant discarded puzzles, the tables below can show surprising values.



the count per clue in the data base is now the following

Code: Select all
clues Count
20      79
21   10961  +1
22   87995  -281
23  390377  -3274
24  699589  -5135
25  435017  -10702
26  480919  +76306


here it is clear that the redundant puzzles were mainly in the 24/25 clues area

The count per er is the following
Code: Select all
er   count
11.9     9
11.8    54
11.7   119  +3
11.6   286  +34
11.5   703  +89
11.4  1630 +125
11.3  9353 +446



The changes in the highest entries are mainly dues to paquita additions and partly offset by the redundancy cleaning


and the combined table is the following

Code: Select all
clues   20      21       22       23      24      25      26
11.9             1        5        3         
11.8     1      15       29        9         
11.7            31       51       11      17       8       1
11.6     1      38       53       55      95      36       8
11.5            46      104      171     209     127      46
11.4           130      308      404     406     234     148
11.3     1     630     1128     2222    2512    1245    1615
11.2     6     991     2706     5497    6682    2908    3987
11.1    12    1267     5100    16413   23095   11606   12123
                     
        21    3149     9484    24785   33016   16164   17928
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby ssxsssxs » Mon Mar 04, 2019 11:14 pm

two of my 11.8
Code: Select all
001000002030004050600020700000005000008070100090300000007060008040000090800000200    ED=11.8/11.8/10.7
010200000300004050000060700100500040008000009040003000009000600020400010000080007    ED=11.8/1.2/1.2
User avatar
ssxsssxs
 
Posts: 159
Joined: 22 January 2014

Re: The hardest sudokus (new thread)

Postby champagne » Tue Mar 05, 2019 1:41 pm

the 11.8/11.8/11.7 seems new.
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby coloin » Thu Mar 07, 2019 1:37 am

ssxsssxs wrote:...
Code: Select all
001000002030004050600020700000005000008070100090300000007060008040000090800000200    ED=11.8/11.8/10.7

ahhh shame its not symetrical - otherwise you could enter the pattern in the patterns game .... :roll:

well done ....
coloin
 
Posts: 1804
Joined: 05 May 2005

Re: The hardest sudokus (new thread)

Postby tarek » Thu Mar 07, 2019 8:31 pm

coloin wrote:
ssxsssxs wrote:...
Code: Select all
001000002030004050600020700000005000008070100090300000007060008040000090800000200    ED=11.8/11.8/10.7

ahhh shame its not symetrical - otherwise you could enter the pattern in the patterns game .... :roll:

well done ....

Are you being cryptic?
Code: Select all
..5.......7...1..83......9......28......9..4..6.8....7...2....1..4.5..3..2...76..
User avatar
tarek
 
Posts: 3026
Joined: 05 January 2006

Re: The hardest sudokus (new thread)

Postby ssxsssxs » Sat Mar 09, 2019 2:58 am

coloin wrote:
ssxsssxs wrote:...
Code: Select all
001000002030004050600020700000005000008070100090300000007060008040000090800000200    ED=11.8/11.8/10.7

ahhh shame its not symetrical - otherwise you could enter the pattern in the patterns game .... :roll:

well done ....

i think it is symetrical.
User avatar
ssxsssxs
 
Posts: 159
Joined: 22 January 2014

Re: The hardest sudokus (new thread)

Postby ssxsssxs » Fri Mar 29, 2019 1:11 am

patterns game 0344 is the same with the second one of my post in
the-hardest-sudokus-new-thread-t6539-600.html#p275842
User avatar
ssxsssxs
 
Posts: 159
Joined: 22 January 2014

Re: The hardest sudokus (new thread)

Postby mysdk » Mon Apr 15, 2019 10:27 am

ssxsssxs wrote:two of my 11.8
Code: Select all
001000002030004050600020700000005000008070100090300000007060008040000090800000200    ED=11.8/11.8/10.7
010200000300004050000060700100500040008000009040003000009000600020400010000080007    ED=11.8/1.2/1.2



The second puzzle is more difficult than the previous.
mysdk
 
Posts: 2
Joined: 15 April 2019

Re: The hardest sudokus (new thread)

Postby champagne » Mon May 06, 2019 10:33 am

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_1905.zip

the link to that storage place is here


The file contains now 2 135 371 puzzles (2 105 149 in the former release done one month ago)
Not a big increase in volume, but interesting contributions in the top part of the data base.
one contributor ssxsssxs "SSX" with a 11.8 11.8 10.7
but mostly during last year additions from paquita "PAQ" a relatively silent member of the forum with now 151 puzzles rating 11.6 and more with a 11.8 11.8 10.6 and the first 11.8 with 24 clues.


With a contribution of 30983 entries, paquita is now the third contributor to the data base.


the count per clue in the data base is now the following


Code: Select all
clues Count
20      79
21   11005 +44
22   88531 +536
23  393612 +3235
24  709032 +9443
25  449086 +14069
26  483459 +2540



The count per er is the following

Code: Select all
er   count
11.9     9
11.8    58 +4
11.7   121 +2
11.6   329 +43
11.5   901 +198
11.4  2077 +447
11.3 12100 +2747


The changes in the highest entries are mainly dues to paquita's additions


and the combined table is the following

Code: Select all
        20      21     22     23     24     25     26
11.9             1      5      3         
11.8     1      16     30     10      1      
11.7            31     52     12     17      8      1
11.6     1      39     60     60    102     58      9
11.5            48    110    192    292    211     48
11.4           132    329    485    570    407    154
11.3     1     634   1213   2500   3311   2564   1853
11.2     6     996   2748   5707   7404   3913   4119
11.1    12    1292   5470  19048  30750  23049  14278                     
                     
        21    3189  10017  28017  42447  30210  20462


with the new 11.8/24 clues from paquita
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby Obi-Wahn » Sun Aug 04, 2019 11:05 am

Hi, champagne,

I was checking your latest database against mine to see if there are any new puzzles scoring at least jellies in my rating system. In the process I noticed that a bunch of puzzles labeled as GP 13_10 are missing in your database. I use my own normalizing function to compare puzzles and find isomorphs and I'm pretty sure they aren't included anymore.
So I did a little research and found that after the january cleanup the numbers in your index file skip from 1150324 to 1208173 and that's the area where the missing puzzles were located.

Is there a chance that you accidentally removed puzzles that weren't duplicates?

btw, I found 78 new entries for my database which contains now 2085 puzzles, but sadly no new stars appeared. They still linger at 20.
User avatar
Obi-Wahn
 
Posts: 63
Joined: 05 January 2007
Location: Darmstadt, Germany

Re: The hardest sudokus (new thread)

Postby champagne » Sun Aug 04, 2019 11:14 am

Obi-Wahn wrote:Hi, champagne,

I was checking your latest database against mine to see if there are any new puzzles scoring at least jellies in my rating system. In the process I noticed that a bunch of puzzles labeled as GP 13_10 are missing in your database. I use my own normalizing function to compare puzzles and find isomorphs and I'm pretty sure they aren't included anymore.
So I did a little research and found that after the january cleanup the numbers in your index file skip from 1150324 to 1208173 and that's the area where the missing puzzles were located.

Is there a chance that you accidentally removed puzzles that weren't duplicates?

btw, I found 78 new entries for my database which contains now 2085 puzzles, but sadly no new stars appeared. They still linger at 20.

Hi Obi-whan,
I think that this was the second cleaning done on old lots coming out of my generation.
To make it simple in the transfer from my Access working data base to the Potential hardest data base, there is a window for a handling error.The result is a wrong canonical form, and a risk of redundant puzzles.
I have been caught at least twice. The redundant puzzles have been cleaned and the still valid removed/replaced to have the correct canonical form.
champagne
2017 Supporter
 
Posts: 6959
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby Obi-Wahn » Sun Aug 04, 2019 11:57 am

Yes, I did understand that. But you said that the error only affected 02_index.txt. But now there are some puzzles missing from 01_file1.txt that would still belong there, because they aren't redundant.

For example:
Code: Select all
98.7..6..75.....9...6.....74..8...5.....6...3.....21...7.9...4...9.3...2.....1...;11.40;11.40;10.40;GP;13_10;1150446;23;
9..8..7...8..6..5...4..3..88..9....2.7....1....3..4....2...8.6....5....7....1.4..;11.30;11.30;10.70;GP;13_10;1150328;23;
98.7..6..75..9......4......5...7.8.....3...4......2..14.......3.9..6.5....51...2.;11.30;11.30;10.50;GP;13_10;1150449;23;
98.7.....6...9.8....5..4...7......8..9..3.7....2.....1.6..7.3.....5....4.....1.2.;11.30;1.20;1.20;GP;13_10;1150356;22;
98.7..6..75..4......3..9...8...5.7.....2...6......3..16.......4.9..7.8.....1...2.;11.30;1.20;1.20;GP;13_10;1150447;23;
98.7.....65....4....3.6..9.7..6..9....5.3..2......7..1.4.8..5......12..3.........;11.20;11.20;9.80;GP;13_10;1150360;23;
98.76....5....48....3....7.6...7.5...5.........72....1.4..8.9.....1....3.......2.;11.20;1.20;1.20;GP;13_10;1150452;22;
98.7.....6...9.5....4..3...7..5..6......2..3......6..1.9.6..8....1.....2.....2.4.;11.10;11.10;7.90;GP;13_10;1150352;22;
98.7..6..75.....9...6.9....5..8...4...8.3...2.....1....4.9...5.....6...3.....21..;10.70;10.70;3.40;GP;13_10;1158371;23;*


I copied these from the ph_1812 01_file1.txt.
User avatar
Obi-Wahn
 
Posts: 63
Joined: 05 January 2007
Location: Darmstadt, Germany

PreviousNext

Return to General