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 mith » Wed Sep 16, 2020 1:47 pm

Yes, I agree, thus the quotes around hardest.

What I mean though is that among the puzzles in my previous post, it's one of the few with high SER that doesn't appear to have something like an MSLS or an Exocet that would lower the rating if implemented in SE. (I'd be curious where these fall in the BpB classification though!)
mith
 
Posts: 369
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby denis_berthier » Wed Sep 16, 2020 2:40 pm

mith wrote:Yes, I agree, thus the quotes around hardest.
What I mean though is that among the puzzles in my previous post, it's one of the few with high SER that doesn't appear to have something like an MSLS or an Exocet that would lower the rating if implemented in SE.

I see, thanks

mith wrote:I'd be curious where these fall in the BpB classification though!

The presence of an Exocet is irrelevant to the BpB classification.
There's a relatively strong correlation between it and the SER, as shown in table 11.2, p. 314 of PBCS2. As you have the SERs, this may give you some idea of the BpB classifications of your puzzles.

If you need a precise value, computing the BpB classification may take a long time. I've spent much time long ago doing it for a significant part of the hardest collection. Now, I do it only when I have a good reason (SER >= 11.8), e.g. searching for a possible puzzle that wouldn't be in B7B.

What's new since I made CSP-Rules public (https://github.com/denis-berthier/CSP-Rules-V2.1) is, anyone can now do these computations.
In the SudoRules config file, choose the following settings (by deleting the leading semi-colons on each line
(don't forget to add semi-colons to inactivate the patterns activated by default in section "1) Choose ordinary resolution rules" of the file

Code: Select all
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; 2c) for computing the BpB classification
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;;; choose p (here p = 5):
 (bind ?*Whips* TRUE)
 (bind ?*Braids* TRUE)
 (bind ?*whips-max-length* 5)
 (bind ?*braids-max-length* 5)

;;; choose one of the following forms of T&E(1)
 (bind ?*TE1* TRUE) ;;; for T&E at level 1
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

Re: The hardest sudokus (new thread)

Postby mith » Thu Oct 01, 2020 2:22 am

Found another 11.8 today, this one a pearl. I'll post a full update tomorrow.

Code: Select all
+-------+-------+-------+
| . . . | . . 9 | . . . |
| . . . | . 8 . | . 7 6 |
| . . . | 7 . . | . 8 5 |
+-------+-------+-------+
| . . 5 | . . . | . . 7 |
| . 9 . | . . . | . 5 . |
| 1 . . | . . 2 | 3 . . |
+-------+-------+-------+
| . 1 . | . . 4 | . . . |
| 3 4 . | . 9 . | . . . |
| . . 6 | 8 1 . | . . . |
+-------+-------+-------+
.....9.......8..76...7...85..5.....7.9.....5.1....23...1...4...34..9......681....

minlex ........1....23.4.....452....1.3.....3...4...6..7....8..6.....9.5....62.7.9...1..  ED=11.8/11.8/2.6 (DLFC+DFC) q1=95515 q2=98004


(MSLS gets it down to 9.1.)
mith
 
Posts: 369
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby denis_berthier » Thu Oct 01, 2020 5:32 am

mith wrote:
Code: Select all
minlex ........1....23.4.....452....1.3.....3...4...6..7....8..6.....9.5....62.7.9...1..  ED=11.8/11.8/2.6 (DLFC+DFC) q1=95515 q2=98004


SER 11.8 but only in B5B (or less - I didn't try).
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

Re: The hardest sudokus (new thread)

Postby champagne » Wed Oct 07, 2020 12:52 pm

mith wrote:Found another 11.8 today, this one a pearl. I'll post a full update tomorrow......


Hi mith,

I am updating the data base, as you produced many versions of your findings, could you (private message??) send the summary of puzzles to add to the data base, if possible with only

puzzle (81) followed by serate ratings ER,EP,ED
eg
.....1.2.....3...4..56..1....6..7..1.4.....8.1..9..7....1..69..8...2.....3.5.....;11.4;1.2;1.2

I added games till game 389, I'll add recent games after the process of your file
champagne
2017 Supporter
 
Posts: 7178
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby mith » Wed Oct 07, 2020 11:18 pm

Sure, I can do that. I'm running an update for the ones I've run serate on, I'll put it together once that's done.
mith
 
Posts: 369
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby champagne » Wed Oct 14, 2020 5:48 am

Having received a significant contribution from mith, 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_2010.zip

the link to that storage place is here

The file contains now 3 103 972 puzzles ( 3 049 068 in the former release dated October 2019)

On interesting fact is the arrival of new very high SE ratings (11.8 and 11.7 ...) already posted in this thread

The readme file included in the zip file have more details.
champagne
2017 Supporter
 
Posts: 7178
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby mith » Fri Oct 16, 2020 4:23 am

New 11.8 tonight:

Code: Select all
........1.......2...3..45......6...7..5..8....9.3..4...39..58...8.9.....2...1..7.  ED=11.8/1.2/1.2
mith
 
Posts: 369
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby denis_berthier » Fri Oct 16, 2020 5:11 am

If the "database" is a real database, it shouldn't be too difficult to -issue a diff at the same time, so that we don't have to search for what's really new.
I asked the previous time, to no avail.
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

Re: The hardest sudokus (new thread)

Postby champagne » Fri Oct 16, 2020 6:09 am

denis_berthier wrote:If the "database" is a real database, it shouldn't be too difficult to -issue a diff at the same time, so that we don't have to search for what's really new.
I asked the previous time, to no avail.

Anybody working in this field can do it easily with its own storage, but if this can help the new entries are at the bottom of the file o2_index.
Here, all new puzzles of very high rating have been posted in this thread (or in the pattern game)
champagne
2017 Supporter
 
Posts: 7178
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby denis_berthier » Fri Oct 16, 2020 3:41 pm

champagne wrote:
denis_berthier wrote:If the "database" is a real database, it shouldn't be too difficult to -issue a diff at the same time, so that we don't have to search for what's really new.
I asked the previous time, to no avail.

Anybody working in this field can do it easily with its own storage, but if this can help the new entries are at the bottom of the file o2_index.
Here, all new puzzles of very high rating have been posted in this thread (or in the pattern game)


Not everybody wants to read the 49 pages of this thread plus the 3054 of the patterns game. And not everybody wants to use 19th century methods for finding the proper information by having to combine two files.
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

Re: The hardest sudokus (new thread)

Postby denis_berthier » Fri Oct 16, 2020 3:41 pm

mith wrote:New 11.8 tonight:

Code: Select all
........1.......2...3..45......6...7..5..8....9.3..4...39..58...8.9.....2...1..7.  ED=11.8/1.2/1.2


I tested it: it's in B5B (or lower).
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

Re: The hardest sudokus (new thread)

Postby champagne » Fri Oct 16, 2020 7:37 pm

denis_berthier wrote:
champagne wrote:
denis_berthier wrote:If the "database" is a real database, it shouldn't be too difficult to -issue a diff at the same time, so that we don't have to search for what's really new.
I asked the previous time, to no avail.

Anybody working in this field can do it easily with its own storage, but if this can help the new entries are at the bottom of the file o2_index.
Here, all new puzzles of very high rating have been posted in this thread (or in the pattern game)


Not everybody wants to read the 49 pages of this thread plus the 3054 of the patterns game. And not everybody wants to use 19th century methods for finding the proper information by having to combine two files.

and each user has it's own spec of the "diff" needed. 58 thousand new puzzles does not help that much)
I did my best to keep this data base.
I am now an old man (78)
it's time for anybody willing to replace me to take over the job
champagne
2017 Supporter
 
Posts: 7178
Joined: 02 August 2007
Location: France Brittany

Re: The hardest sudokus (new thread)

Postby mith » Fri Oct 16, 2020 7:49 pm

Since I'm the one adding most of the puzzles at the moment, I don't mind putting up my own version of the database at some point.
mith
 
Posts: 369
Joined: 14 July 2020

Re: The hardest sudokus (new thread)

Postby denis_berthier » Sat Oct 17, 2020 12:15 am

champagne wrote:
denis_berthier wrote:Not everybody wants to read the 49 pages of this thread plus the 3054 of the patterns game. And not everybody wants to use 19th century methods for finding the proper information by having to combine two files.

and each user has it's own spec of the "diff" needed. 58 thousand new puzzles does not help that much)
I did my best to keep this data base.
I am now an old man (78)
it's time for anybody willing to replace me to take over the job


Your work of collecting all the findings has been useful. I just meant the presentation could be improved. And I don't think the diff would be that large: only the upper part of the collection is really useful.
denis_berthier
2010 Supporter
 
Posts: 1997
Joined: 19 June 2007
Location: Paris

PreviousNext

Return to General