A Pure Jellyfish Collection

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

Postby tarek » Sun Feb 03, 2008 1:38 am

The -J shuffling worked........

the problem is with bn, hn, in

Several attempts were not successful.

I tried this
Code: Select all
-q"{FNBTHW2W3}{W4}-G" -f"%%(h2)x,%%(b2)x,%%(i2)x,%%(I2)x,%%(W4)x"
on the following puzzle
Code: Select all
..3..4..9....7..5.2..6..8....9.....4.4.....3.8.....7....4..3..2.2..6....7..8..6..
the output was
Code: Select all
0,0,0,3
:(

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby gsf » Sun Feb 03, 2008 7:36 am

tarek wrote:The -J shuffling worked........

the problem is with bn, hn, in

Several attempts were not successful.

I tried this
Code: Select all
-q"{FNBTHW2W3}{W4}-G" -f"%%(h2)x,%%(b2)x,%%(i2)x,%%(I2)x,%%(W4)x"
on the following puzzle
Code: Select all
..3..4..9....7..5.2..6..8....9.....4.4.....3.8.....7....4..3..2.2..6....7..8..6..
the output was
Code: Select all
0,0,0,3
:(

there are some subtleties when BTHW are specified without sizes
they expand group-wise and cause extra grouping assigmnents
this is not well-documented

but you can use -T0x10 (really a debugging option) to see the groupings
so that you can use the correct hN, bN and iN indices

here is the first line of -T0x10 output for -q{FNBTHW2W3}{W4}-G
Code: Select all
{FNB2B3T2H2}{T3H3}{T4H4W2W3}W4

you can see that B expanded to B2B3 and TH expanded to T2H2T3H3T4H4 and also added some {...} grouping
making the W4 group index 4

I just reposted the 2008-02-02 solver, it has the %#sQ format for the expanded -q solver constraints
so either -T0x10 or -f%#sQ will show the expanded constraints
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

re: gsf's software -- extra groupings

Postby Pat » Sun Feb 03, 2008 10:25 am

gsf wrote:there are some subtleties when BTHW are specified without sizes
they expand group-wise and cause extra grouping assigmnents
this is not well-documented

but you can use -T0x10 (really a debugging option) to see the groupings
so that you can use the correct hN, bN and iN indices

here is the first line of -T0x10 output for -q{FNBTHW2W3}{W4}-G
Code: Select all
{FNB2B3T2H2}{T3H3}{T4H4W2W3}W4


you can see that B expanded to B2B3 and TH expanded to T2H2T3H3T4H4 and also added some {...} grouping
making the W4 group index 4

I just reposted the 2008-02-02 solver, it has the %#sQ format for the expanded -q solver constraints
so either -T0x10 or -f%#sQ will show the expanded constraints



thanks -- now that we know it, we have the simple method of specifying the exact groupings we wish
    e.g. avoid -q{NFBHTW2W3}{W4}-G
    in favour of -q{NFB2B3H2T2H3T3H4T4W2W3}{W4}-G
User avatar
Pat
 
Posts: 3425
Joined: 18 July 2005

Postby tarek » Sun Feb 03, 2008 12:15 pm

I've just made the adjustments to your multiple -q operations in accordance with what Pat mentined.

It works like a charm, it captures the current rules totally (the point 8 difference is minor)....

with -JN option also working like a charm.... Points 4.1 & 4.2 are probably redundant & can be replaced by max(W4)>=max(W4) (H1T1W4 & the other for H1T1B2B3H2T2T3H3T4H4W2W3W4), this needs 2 runs (which is possible with your new version) but can't handle more than 1 input, therefore for non batch mode it has to be done in 3 lines
1 line for points 1-3
1 line for point 4
1 line for point 5-8

I'll experiment & c what happens.

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby tarek » Wed Feb 06, 2008 4:12 pm

I tried using gsf's solver for the puzzles in this thread (this point has been reached after several trial & error attempts)

the current rules can be emulated (save for a minor difference in point 8) using the following (windows cmd):
Code: Select all
-BG
-q"FN;v_1=V"
-q"{FNB2B3H2T2T3H3H4W2W3}{W4}-G;v_2=V&&h2;h_2=h2;b_2=b2;i_2=i2"
-q"{FN}{W4}-G;v_3=V&&h2;h_3=h2;b_3=b2;i_3=i2"
-q"{FNB2B3H2T2T3H3H4W2W3X}{W4}-G;v_4=V&&h2;h_4=h2"
-f"%%v,%%(v_1?(v_2?(v_3?((b_2>b_3||b_2==b_3&&i_2>=i_3)?(S?(C1?((v_4&&h_4)?80:70):60):50):((b_2==b_3)?42:41)):30):20):10)x"
This has caused puzzle#daj95376-003 to score 4.2
& Puzzle#Mauricio-004 to score 80

the use of the -J otipn has identified these problems. however, incorporating them into the process so that a list of puzzles can be processed with a double-click to give you a list of puzzle with each individual score proved to be difficult.

I have therefore used the followingg set of options
Code: Select all
-BG
-q"FN;v_1=V"
-q"{FNB2B3H2T2T3H3H4W2W3}{W4}-G;v_2=V&&h2;h_2=h2;b_2=b2;i_2=i2"
-q"{FN}{W4}-G;v_3=V&&h2;h_3=h2;b_3=b2;i_3=i2"
-q"{FNB2B3H2T2T3H3H4W2W3X}{W4}-G;v_4=V&&h2;h_4=h2"
-f"%%v,%%(Poster)f,%%(v_1?(v_2?(v_3?(S?(C1?((v_4&&h_4)?80:70):60):50):30):20):10)x"

which skips point 4 all together & run manually each puzzle that scores >=50 with EACH of the following
Code: Select all
-q"{[B]NF}{[S]W4}-G" -J -e"min(W4)+max(W4)+1" -f- -Ff"%%(min())x,%%(max())x"
-q"{[B]NFBHTW2W3}{[S]W4}-G" -J -e"min(W4)+max(W4)+1" -f- -Ff"%%(min())x,%%(max())x"
The passes should have a similar minimum count.

If the FNW4 pass has a greater minimum then that puzzle would be demoted to score 40. For the 19 posted puzzles, this happened only to Puzzle#tarek-001 & Puzzle#Mauricio-004.

I can confirm that for the puzzles scoring >= 50, Only 1 Jellyfish is needed to solve except for Puzzle#tarek-006 which needs 2.

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby gsf » Wed Feb 06, 2008 6:09 pm

tarek wrote:I tried using gsf's solver for the puzzles in this thread (this point has been reached after several trial & error attempts)

teaching to fish is much better than giving fish
a few more rounds like this and I won't be able to grok input to my own solver!
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby tarek » Wed Feb 06, 2008 9:14 pm

here are the 19 posted puzzles, with scores according to a combination of gsf's solver & manual checking using the -J option
Code: Select all
#Puzzle#Chronology#ID#Score#JF
7...2......1..9.6..8.4..2....6....5.9.......1.3....7....2..6.9..4.7..8......3...6#001#tarek-001#4#1
7...6......3..1.8..2.7..6....8....4.1.......3.9....2....6..8.1..5.2..7......7...8#002#tarek-002#7#1
6...7......1..5.3..2.4..6....3....1.9.......2.4....7....9..3.5..7.6..9......1...3#003#tarek-003#8#1
...197....4.....7...9...3..8...3...27..9.2..43...6...1..2...4...1.....9....524...#004#tarek-004#7#1
.5.....7.2.......6...456.....7.3.4....65.12....9.7.1.....812...9.......5.1.....2.#005#tarek-005#7#1
.4..28.7187.6...2............48..26...1.4.9...86..24............6...9.4313.27..9.#006#daj95376-001#7#1
2....358.....2...1...4.56.....9.4.68..5.1.9..61.8.2.....15.8...3...4.....263....4#007#daj95376-002#7#1
.......63..763.48...3..8..1.3...41.5.1..8..3.5.63...7.2..7..3...74.562..65.......#008#daj95376-003#6#1
4..2.7.8...3..85.6.786...4.2.....61.....2.....17.....4.5...346.6.15..2...3.1.2..8#009#daj95376-004#6#1
1.........2.345.....5...4...6.7...3..8..1..5..1...9.4...7...5.....283.6.........8#010#TTHsieh-001#3#0
..3..4..9....7..5.2..6..8....9.....4.4.....3.8.....7....4..3..2.2..6....7..8..6..#011#tarek-006#8#2
.......1....8.9257.2....9...829.1.45.47.8.19.19.4.372...1....8.9381.5....5.......#012#daj95376-005#6#1
.1.....2.34.....56...2.6.....76285.....5.7.....69342.....4.5...56.....84.7.....1.#013#JPF-001#2#0
.1.....2.23.....45...5.2.....63748.....9.6.....42586.....4.7...74.....38.5.....1.#014#JPF-002#2#0
.1.....2.34.....56...6.4.....57862.....9.5.....74235.....3.7...96.....32.5.....1.#015#JPF-003#2#0
...........1...2..34..5..67.........15.2.3.7883.....41.........42.7.8.5671.....83#016#Mauricio-001#8#1
.............1.....123.456...........23...78..47.6.12...........318.764..58...23.#017#Mauricio-002#8#1
.............1.....123.456...........27...38..36.5.42...........682.715..74...83.#018#Mauricio-003#8#1
.............1.....123.456...........35...78..81.2.35...........57...63..638.721.#019#Mauricio-004#4#1

The score now has 4 (not 4.1 & 4.2) to show that the number of Jellyfish needed following singles is more than the number needed using intermediate techniques, the rest of the scores correspond to the scores outlined in the head post. a score of 8 doesn't correspond to the head post but the NFBHTW2W3X techniques in gsf's solver which means singles+fish up to swordfish +subsets+x-cycles
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby tarek » Tue Sep 30, 2008 3:05 pm

I generated this during the course of Patterns Game 0052, It is a Diamond Benchmark Pure Jellyfish ... So something that you can fish out just by Eyeballing the puzzle:D
Code: Select all
 2 . . | . . . | . . 3 
 . 8 . | . 3 . | . 5 . 
 . . 3 | 4 . 2 | 1 . . 
-------+-------+------
 . . 1 | 2 . 5 | 4 . . 
 . . . | . 9 . | . . . 
 . . 9 | 3 . 8 | 6 . . 
-------+-------+------
 . . 2 | 5 . 6 | 9 . . 
 . 9 . | . 2 . | . 7 . 
 4 . . | . . . | . . 1
I will be posting here some more from that thread and also on the Pure Swordfish thread.
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby tarek » Thu Oct 02, 2008 7:03 am

These have been collated from Patterns Gmaes 14-51 ....
As SE technique Heirarchy is different than most players heirarchy, I include here puzzles that can be solved by subsets & basic unfinned fish only ...
Code: Select all
000000000008605900053407180039000870000030000041000320014309250007501400000000000 Patterns_Game_0021_026-m_b_metcalf 5.2/1.2/1.2
120000034003000200500060007200504001004000500800601002300010005002000600650000028 Patterns_Game_0023_070-JPF         5.2/1.2/1.2
100203004000000000002010500400000003006070800500000001008050600000000000300409005 Patterns_Game_0024_008-Mauricio    5.2/5.2/2.6
000000000004703200012509370097000140000000000065000930039107420001208500000000000 Patterns_Game_0030_002-m_b_metcalf 5.2/5.2/5.2
000005900010040080300600000200030600030406070001080009000003006060070010004500000 Patterns_Game_0034_076-gsf         5.2/1.2/1.2
010000002304000000050060400000500001006070080000003700007002600000080040100300005 Patterns_Game_0037_103-JPF         5.2/1.2/1.2
010200030400050006000001000700080100020509060005060008000900000500070004060005020 Patterns_Game_0039_117-JPF         5.2/1.2/1.2
010200030400010005000006000200040500040102070003050008000300000900020004070008010 Patterns_Game_0039_131-Mauricio    5.2/2.0/2.0
102030405000000000400805003003000200200070004004000600800106007000000000706040508 Patterns_Game_0042_089-gsf         5.2/3.4/2.6
001200000030004500600000070700600020000000000040005008050000003007100060000003400 Patterns_Game_0045_049-JPF         5.2/1.2/1.2
001200000030004500600000070700600020000020000040000008080000003009700060000005804 Patterns_Game_0046_082-TTHsieh     5.2/1.2/1.2
000001200030040050206500700500000600040000020001000008003006102060030070009700000 Patterns_Game_0048_090-JPF         5.2/2.0/2.0
001020300040506010700010006080000070107000805020000060200050004050102080009060500 Patterns_Game_0050_070-TTHsieh     5.2/5.2/3.4

My favourite out of the bunch is Patterns_Game_0045_049-JPF

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby Pat » Thu Oct 02, 2008 7:25 am

tarek wrote:These have been collated from Patterns Games 14-51 ....

As SE technique Heirarchy is different than most players heirarchy, I include here puzzles that can be solved by subsets & basic unfinned fish only ...
    21 ...........86.59...534.718..39...87.....3.....41...32..143.925...75.14........... 29 Patterns_Game_0021_026-m_b_metcalf
    22 12.....34..3...2..5...6...72..5.4..1..4...5..8..6.1..23...1...5..2...6..65.....28 28 Patterns_Game_0023_070-JPF
    23 1..2.3..4...........2.1.5..4.......3..6.7.8..5.......1..8.5.6...........3..4.9..5 21 Patterns_Game_0024_008-Mauricio
    24 ...........47.32...125.937..97...14...........65...93..391.742...12.85........... 28 Patterns_Game_0030_002-m_b_metcalf
    25 .....59...1..4..8.3..6.....2...3.6...3.4.6.7...1.8...9.....3..6.6..7..1...45..... 24 Patterns_Game_0034_076-gsf
    26 .1......23.4.......5..6.4.....5....1..6.7..8......37....7..26......8..4.1..3....5 22 Patterns_Game_0037_103-JPF
    27 .1.2...3.4...5...6.....1...7...8.1...2.5.9.6...5.6...8...9.....5...7...4.6...5.2. 24 Patterns_Game_0039_117-JPF
    28 .1.2...3.4...1...5.....6...2...4.5...4.1.2.7...3.5...8...3.....9...2...4.7...8.1. 24 Patterns_Game_0039_131-Mauricio
    29 1.2.3.4.5.........4..8.5..3..3...2..2...7...4..4...6..8..1.6..7.........7.6.4.5.8 25 Patterns_Game_0042_089-gsf
    30 ..12......3...45..6......7.7..6...2...........4...5..8.5......3..71...6......34.. 20 Patterns_Game_0045_049-JPF
    31 ..12......3...45..6......7.7..6...2.....2.....4......8.8......3..97...6......58.4 21 Patterns_Game_0046_082-TTHsieh
    32 .....12...3..4..5.2.65..7..5.....6...4.....2...1.....8..3..61.2.6..3..7...97..... 24 Patterns_Game_0048_090-JPF
    33 ..1.2.3...4.5.6.1.7...1...6.8.....7.1.7...8.5.2.....6.2...5...4.5.1.2.8...9.6.5.. 28 Patterns_Game_0050_070-TTHsieh


but for "Pure" Jellyfish,
only Jellyfish (+ "singles") allowed
i.e. only 21 24 30
User avatar
Pat
 
Posts: 3425
Joined: 18 July 2005

Postby tarek » Thu Oct 02, 2008 7:33 am

Pat wrote:but for "Pure" Jellyfish,
only Jellyfish (+ "singles") allowed
i.e. only 21 24 30
True, Those are the only ones that will make the list ... I included the rest due to their rarity .... The SE ranks Jellyfish Higher than Many other techniques .... Making an ER of 5.2 a very strong Jellyfish that is difficult to bypass.

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby tarek » Fri Oct 03, 2008 1:27 am

Here is the latest list with Puzzles that scored >=4 (Pat's Extra Lenient Criteria) in chronological order:

Puzzle#Chronology#ID#Score#Jellyfish_needed
7...2......1..9.6..8.4..2....6....5.9.......1.3....7....2..6.9..4.7..8......3...6#001#tarek-001#4#1
7...6......3..1.8..2.7..6....8....4.1.......3.9....2....6..8.1..5.2..7......7...8#002#tarek-002#7#1
6...7......1..5.3..2.4..6....3....1.9.......2.4....7....9..3.5..7.6..9......1...3#003#tarek-003#8#1
...197....4.....7...9...3..8...3...27..9.2..43...6...1..2...4...1.....9....524...#004#tarek-004#7#1
.5.....7.2.......6...456.....7.3.4....65.12....9.7.1.....812...9.......5.1.....2.#005#tarek-005#7#1
.4..28.7187.6...2............48..26...1.4.9...86..24............6...9.4313.27..9.#006#daj95376-001#7#1
2....358.....2...1...4.56.....9.4.68..5.1.9..61.8.2.....15.8...3...4.....263....4#007#daj95376-002#7#1
.......63..763.48...3..8..1.3...41.5.1..8..3.5.63...7.2..7..3...74.562..65.......#008#daj95376-003#6#1
4..2.7.8...3..85.6.786...4.2.....61.....2.....17.....4.5...346.6.15..2...3.1.2..8#009#daj95376-004#6#1
..3..4..9....7..5.2..6..8....9.....4.4.....3.8.....7....4..3..2.2..6....7..8..6..#011#tarek-006#8#2
.......1....8.9257.2....9...829.1.45.47.8.19.19.4.372...1....8.9381.5....5.......#012#daj95376-005#6#1
...........1...2..34..5..67.........15.2.3.7883.....41.........42.7.8.5671.....83#016#Mauricio-001#8#1
.............1.....123.456...........23...78..47.6.12...........318.764..58...23.#017#Mauricio-002#8#1
.............1.....123.456...........27...38..36.5.42...........682.715..74...83.#018#Mauricio-003#8#1
.............1.....123.456...........35...78..81.2.35...........57...63..638.721.#019#Mauricio-004#4#1
2.......3.8..3..5...34.21....12.54......9......93.86....25.69...9..2..7.4.......1#020#tarek-007#8#1
...........86.59...534.718..39...87.....3.....41...32..143.925...75.14...........#021#m_b_metcalf-001#4#1
...........47.32...125.937..97...14...........65...93..391.742...12.85...........#024#m_b_metcalf-002#8#1
..12......3...45..6......7.7..6...2...........4...5..8.5......3..71...6......34..#030#JPF-007#4#1

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby Pat » Fri Oct 03, 2008 2:02 am

tarek wrote:The SE ranks Jellyfish Higher than Many other techniques .... Making an ER of 5.2 a very strong Jellyfish that is difficult to bypass.



what we call a Jellyfish puzzle [ not necessarily "pure" ]
    -q{NFB2B3H2T2H3T3H4T4W2W3}{W4}-G

    I2>0
could conceivably rate anywhere in the range 4.2 - 5.4
in the peculiar hierarchy used by SuDoku Explainer
User avatar
Pat
 
Posts: 3425
Joined: 18 July 2005

Postby tarek » Fri Oct 03, 2008 2:11 am

SE's heirarchy is very strange ....

One good thing about it though (as Mauricio did) is to use on your Pure Jellyfish to see if it fits the Rigorous criteria ....

A pure jellyfish that has an ER=5.2 should be a strong candidate for the "8 score"

tarek
User avatar
tarek
 
Posts: 2612
Joined: 05 January 2006

Postby Mauricio » Fri Oct 03, 2008 7:44 am

A present for Tarek:
Code: Select all
+-------+-------+-------+
| . . 1 | . . . | 2 . . |
| . 2 . | 1 . 3 | . 4 . |
| 5 . . | . . . | . . 6 |
+-------+-------+-------+
| . 1 . | 3 . 6 | . 7 . |
| . . . | . . . | . . . |
| . 3 . | 4 . 7 | . 2 . |
+-------+-------+-------+
| 1 . . | . . . | . . 7 |
| . 6 . | 7 . 4 | . 3 . |
| . . 8 | . . . | 9 . . |
+-------+-------+-------+

Can you confirm this puzzle requires 3 Jellyfishes and that it fits in the most rigorous criteria?
Mauricio
 
Posts: 1174
Joined: 22 March 2006

PreviousNext

Return to General

cron