## Patterns Game Strategies

Interactive on-site game threads go here

### isomorphs

in game 270, Pat wrote:

4..78.9..5..1......8...4.7.6.....5...1..4..3...2.....6.6.4...8......8..9..8.13..2 ED=10.3/10.3/9.8

we can get a slightly-higher rating
by rotating the puzzle

2..31.8..9..8......8...4.6.6.....2...3..4..1...5.....6.7.4...8......1..5..9.87..4 ED=10.3/10.3/9.9
( found by champagne ),
but how would i see that before posting?

rating countless isomorphs is just too expensive

Pat

Posts: 3880
Joined: 18 July 2005

### re: isomorphs

take my posted 10.3
and swap just 2 givens
( the first 7,1 )
4..18.9..5..7......8...4.7.6.....5...1..4..3...2.....6.6.4...8......8..9..8.13..2 ED=10.3/10.3/9.8
this is a Very Similar puzzle
( but not isomorph )
in this case,
rotating the puzzle
leaves us with the same rating
can we find some isomorph of this puzzle
with a higher rating?

Pat

Posts: 3880
Joined: 18 July 2005

### Re: isomorphs

Pat wrote:
but how would i see that before posting?

rating countless isomorphs is just too expensive

Hi pat,

that question is not too hard, although I don't do it and I have been caught as others.

If you suspect a possible morphing trap, just morph both puzzles into a canonical form.

Pat wrote:can we find some isomorph of this puzzle
with a higher rating?

But here, I have no answer. The deviation in serate for morphs of the same puzzle with high ratings in unpredictable. Usually, it is a 0.1 deviation, that's it.
champagne
2017 Supporter

Posts: 7054
Joined: 02 August 2007
Location: France Brittany

### RIP, Patterns Game

Well, it seems that, after nine years, the Patterns Game has finally died. For the record, here is a list of the 30 participants:

Code: Select all
`gsf           coloin        999_Springs   JPF           m_b_metcalf   tarek         Mauricio      Pat           TTHsieh       Eioru         draco         pseudocoup    claudiarabia  mario133      champagne     joel64        ronk          surbier       BryanL        Patrice       daj95376      eleven        papy999       dobrichev     WosRet        blue          GouinJP       ssxsssxs      CSC           billeioru    `

I wanted to send out a PM to all of them to mark the occasion, but the number of recipients in a PM is limited to five, so I gave up. Anyway thanks to all, especially to JPF for the idea and to gsf for looking after g.r.emlin so efficiently.

Attached are summaries of the played submissions and of the winning submissions, for future historians. Submissions up to September last year can be found in a single file kindly hosted by champagne and referenced in the Patterns Game benchmark file thread.

Best wishes,

Mike Metcalf

Edit: fout.txt deleted, see newer version below.
Attachments
fplayed.txt
all submissions
Last edited by m_b_metcalf on Sat Jun 30, 2018 5:53 am, edited 1 time in total.

m_b_metcalf
2017 Supporter

Posts: 10592
Joined: 15 May 2006
Location: Berlin

### 4.9

and after all these years,
game 270 has finally provided the highest-ever 4.9

Pat

Posts: 3880
Joined: 18 July 2005

### Re: 4.9

Pat wrote:and after all these years,
game 270 has finally provided the highest-ever 4.9

Well done. It has a cousin:

Code: Select all
` 1 . . 2 3 . 4 . . 5 . . 6 . . . . . . 7 . . . 8 . 2 . 2 . . . . . 3 . . . 3 . . 5 . . 1 . . . 4 . . . . . 9 . 6 . 8 . . . 4 . . . . . . 4 . . 5 . . 8 . 7 6 . . 1   ED=4.9/4.4/2.6`

m_b_metcalf
2017 Supporter

Posts: 10592
Joined: 15 May 2006
Location: Berlin

### re: game 49

champagne wrote:
---for the game 262

I checked the order of magnitude of the task for a full scan for that pattern.

I filled the box 1, then expanded band 1 stack 1

and solved the redundancy using
gridchecker --pattern --patcanon < in.txt > out.txt
I got 721 ED partial puzzles---

would this be a sensible strategy for game 49 ?

Code: Select all
`. . . . . . O O .. . . O O . . . .. . O . . . O O .. O . . . . . . O. O . . O . . . O. . . . . O . . .O . O . . . O . .O . O . . . . . .. . . O O . . . .`

Pat

Posts: 3880
Joined: 18 July 2005

### Re: re: game 49

Pat wrote:
champagne wrote:
---for the game 262

I checked the order of magnitude of the task for a full scan for that pattern.

I filled the box 1, then expanded band 1 stack 1

and solved the redundancy using
gridchecker --pattern --patcanon < in.txt > out.txt
I got 721 ED partial puzzles---

would this be a sensible strategy for game 49 ?

Code: Select all
`. . . . . . O O .. . . O O . . . .. . O . . . O O .. O . . . . . . O. O . . O . . . O. . . . . O . . .O . O . . . O . .O . O . . . . . .. . . O O . . . .`

Hi pat,
just for memory, the game 262 pattern

Code: Select all
`                    . X X . . . . . X                    X . X . . X . . .                    X X X . . . . . .                    . . . X . . . X .                    . . . . X . X . X                    . X . . . X . . .                    . . . . X . X . .                    . . . X . . . X .                    X . . . X . . . X`

The process can be applied to any pattern. A pattern with 20 clues would generate less puzzles and the scan would be faster, but it remains a slow process.
Normally, the process is better if you take the pair band/stack with the highest number of clues, here band 3 and stack 3.

But you should have good reasons to lock some power to do the job.
champagne
2017 Supporter

Posts: 7054
Joined: 02 August 2007
Location: France Brittany

### neighborhood search w/o minimality constraint

dobrichev (2017.Feb.5) wrote:

Curiously, for this pattern {280}
the neighborhood search w/o minimality constraint
demonstrated fairly good results.

Once I experimented something similar
and concluded that it wouldn't work.
But it works!

i wonder if this would be profitable for the present game (289)

Pat

Posts: 3880
Joined: 18 July 2005

### Highest Non Diamond ratings

Highest Non Diamond ratings

In the last patterns game [289], these ratings were found, and a quick look at some of the other results show that they probably are the max for that rating level
Code: Select all
`4.6/4.6/3.4                                                                                                   4.7/4.7/3.4                                                                                                   5.1/1.2/1.2                                                                                                   5.6/4.7/3.4                                                                                                   5.7/4.7/3.4                                                                                                   5.8/4.6/3.4`

Here are some others from previous games - probably approaching the maximum
Code: Select all
`                                                                                                        183                                                                                                           000000001000002000000130420005001003006007080020690050004000802001049070300700900    4.9/4.7/2.8              000000001000002000000130450005006003001005070020740080004000908003094020900600300    5.4/5.4/4.4              000000001000002000000130450006007002003005070010860090005000709009076010300500800    5.9/4.2/2.6              000000001000002000000130450006003007001004080030790060009000508005047020200800600    6.0/2.6/2.6              000000001000002000000310240005006007006008090010490080004000305002073010600500900    6.1/2.0/2.0              000000001000001000000230450003006004007002080090140020006000708009064030300800900    6.5/6.5/3.0                                                                                                                            184                                                                                                           000000001000023040005406200004070030060200007087001400009005000010300004800090060    5.0/5.0/5.0  [naked quad]                                                                                                              225                                                                                                           000001002003040050050020607000800020074000090800003700007006009090130000605000300    8.0/8.0/7.8                                                                                                                            226                                                                                                           120000000300400500004030060050070040008200900000001002070050010009700300000008007    4.5/4.5/3.8              120000000300200400005060070060020050002600300000008006070050030001300500000004008    5.2/4.6/2.6              120000000300100400004050060040070010002600500000008003070040050009300200000001006    8.1/8.1/3.4                                                                                                                            227                                                                                                           010200000300004005002000400600307004000000010050408007009000700000090006020501090    4.8/4.8/2.6      `
coloin

Posts: 1864
Joined: 05 May 2005

### Re: Highest Non Diamond ratings

coloin wrote:Highest Non Diamond ratings

In the last patterns game [289], these ratings were found, and a quick look at some of the other results show that they probably are the max for that rating level
Code: Select all
`4.6/4.6/3.4                                                                                                   4.7/4.7/3.4                                                                                                   5.1/1.2/1.2                                                                                                   5.6/4.7/3.4                                                                                                   5.7/4.7/3.4                                                                                                   5.8/4.6/3.4`

Code: Select all
`010002003400000020002450000006000071005000600800000200000084900030600017900700050 # 100    4.6 4.6 4.6 0 champagne050007003400000080002410000006000071005000600800000200000084900030200017900700050 # 108    4.7 4.7 4.7 0 Patrice010004003400000050008260000007000085001000600900000200000083500050600094100900070 # 110    5.1 3.4 2.6 0 Patrice000000000012003004045006002000010050000700800027005001000060900000300010098004005 #  44    5.6 5.4 5.4 2 m_b_metcalf000001023000400510000020406060504700008003000900610000049700008580000000306000200 # 119    5.7 5.4 3.0 0 dobrichev`

each occurring just once (5.6 thrice).

Please also see the (now out-of-date) information here (just above).

Regards,

Mike

m_b_metcalf
2017 Supporter

Posts: 10592
Joined: 15 May 2006
Location: Berlin

### Highest Non Diamond ratings

So these are approaching the maxima ... and some indeed are diamonds .... i didnt actually know what a 4.6 puzzle was !!!
Its likely that we can improve on this list

Code: Select all
`120000000300400500004030060050070040008200900000001002070050010009700300000008007 # 226    4.5/4.5/3.8                   010002003400000020002450000006000071005000600800000200000084900030600017900700050 # 100 D  4.6/4.6/4.6   champagne       050007003400000080002410000006000071005000600800000200000084900030200017900700050 # 108 D  4.7/4.7/4.7   Patrice         010200000300004005002000400600307004000000010050408007009000700000090006020501090 # 227    4.8/4.8/2.6                   000000001000002000000130420005001003006007080020690050004000802001049070300700900 # 183    4.9/4.7/2.8                   000000001000023040005406200004070030060200007087001400009005000010300004800090060 # 184 D  5.0/5.0/5.0  [naked quad]     010004003400000050008260000007000085001000600900000200000083500050600094100900070 # 110    5.1/3.4/2.6   Patrice         120000000300200400005060070060020050002600300000008006070050030001300500000004008 # 226    5.2/4.6/2.6                   000000001000002000000130450005006003001005070020740080004000908003094020900600300 # 183    5.4/5.4/4.4                   000000000012003004045006002000010050000700800027005001000060900000300010098004005 #  44    5.6/5.4/5.4   m_b_metcalf     000001023000400510000020406060504700008003000900610000049700008580000000306000200 # 119    5.7/5.4/3.0   dobrichev       000000001000002000000130450006007002003005070010860090005000709009076010300500800 # 183    5.9/4.2/2.6                   000000001000002000000130450006003007001004080030790060009000508005047020200800600 # 183    6.0/2.6/2.6                   000000001000002000000310240005006007006008090010490080004000305002073010600500900 # 183    6.1/2.0/2.0                   000000001000001000000230450003006004007002080090140020006000708009064030300800900 # 183    6.5/6.5/3.0                   000001002003040050050020607000800020074000090800003700007006009090130000605000300 # 225    8.0/8.0/7.8                   120000000300100400004050060040070010002600500000008003070040050009300200000001006 # 226    8.1/8.1/3.4                   `
coloin

Posts: 1864
Joined: 05 May 2005

as i recall,
m_b_metcalf's list is meant for determination of "rarity" bonuses;
thus, it only includes puzzles posted during the game
-- ignoring "more" puzzles (posted after game has closed)

a separate list would include "more"

even then, we'd still be missing puzzles posted Too Late
e.g. 4.9/4.9/2.6 ( above )

Pat

Posts: 3880
Joined: 18 July 2005

### Re: Patterns Game Strategies

Here's the list updated according to my records:
Code: Select all
`120000000300400500004030060050070040008200900000001002070050010009700300000008007 # 226    4.5/4.5/3.8                   010002003400000020002450000006000071005000600800000200000084900030600017900700050 # 100 D  4.6/4.6/4.6   champagne       050007003400000080002410000006000071005000600800000200000084900030200017900700050 # 108 D  4.7/4.7/4.7   Patrice         000100203000405010000023006420000370007000800035000062500670000070509000904008000 # 117    4.8/4.8/3.4 0 m_b_metcalf                 000000001000002000000130420005001003006007080020690050004000802001049070300700900 # 183    4.9/4.7/2.8                   000000001000023040005406200004070030060200007087001400009005000010300004800090060 # 184 D  5.0/5.0/5.0  [naked quad]     010004003400000050008260000007000085001000600900000200000083500050600094100900070 # 110    5.1/3.4/2.6   Patrice         000000000012003004094001007000060020000900700029007001000050800000300040031004002 # 101 D  5.2/5.2/5.2 - gsf                000102030100000400035004600706205008000000000800406207002700590004000003070609000 #  77 D  5.4/5.4/5.4 + Mauricio                000000000012003004045006002000010050000700800027005001000060900000300010098004005 #  44    5.6/5.4/5.4   m_b_metcalf     000001023000400510000020406060504700008003000900610000049700008580000000306000200 # 119    5.7/5.4/3.0   dobrichev     100002300040050020006700004007108000000000000000503900800005200050080070003900006 #  51    5.8/4.6/3.4 0 m_b_metcalf  100002003000040015006000700000700000040008002500029080007000300080005000920010008 #  63    5.9/4.6/3.0 0 champagne                000000012000003004000250360007005000006080900030100046009070000604008000270009000 #  19    6.0/4.6/3.4 0 champagne               000000000001203400024506370013000260000000000087000140052107630006304500000000000 # 133    6.1/5.2/5.2 - Mauricio                  008007009040050060100400700009002004030000090800700600002003005010040030300200100 #  70    6.5/6.5/3.8 0 ronk                  006090500020001040500000702000000020400080300050000000803040007090100080005000600 #  72 D  8.0/8.0/8.0 1 ronk                000000000001203400025607810084000260000000000073000580046802130007504600000000000 # 146    8.1/8.1/7.2 - TTHsieh                  `

I also attach the latest version of my summary of winning posts.

Regards,

Mike
Attachments
fout.txt
Last edited by m_b_metcalf on Tue Aug 22, 2017 8:32 am, edited 1 time in total.

m_b_metcalf
2017 Supporter

Posts: 10592
Joined: 15 May 2006
Location: Berlin

can update 5.8
from game 289

Pat

Posts: 3880
Joined: 18 July 2005

PreviousNext