The New Sudoku Players' Forum

[JPF]

pearls :
We could have 2 types of pearl :

• weak : the first move can't be a single
• strong : SS doesn't solve one cell

assuming each player makes all of the moves possible within the rules
the last player will have 0 or more puzzles in a cache that can't be played
rather than the game morph into solitaire spread out over #puzzles*8 hours
some form of bonus for the last player with extra puzzles might be more pleasant for the other players
maybe a .... SCORE_MORE*#extra.puzzles

That's fine for me.

JPF

[gsf]

Code: Select all

. . 3 | 4 . . | . 8 9
4 . 7 | 1 . . | . . .
. 9 . | . . . | . . .
------+-------+------
. 3 . | 9 . 4 | . . 5
. . 4 | . . . | . . .
6 1 . | . . . | . . 3
------+-------+------
. . . | . 2 . | . 9 7
. . . | 3 . 8 | 5 . .
9 . . | . . 5 | . 2 .  ER=000064900000100000108009700309000070700080005050000208001900803000008000004210000


[JPF]

ER=000064900000100000108009700309000070700080005050000208001900803000008000004210000

New world record ?

JPF

[gsf]

Code: Select all

. . 3 | 4 . . | . 8 9
4 . 7 | 1 . . | . . .
. 9 . | . . . | . . .
------+-------+------
. 3 . | 9 . 4 | . . 5
. . 4 | . . . | . . .
6 1 . | . . . | . . 3
------+-------+------
. . . | . 2 . | . 9 7
. . . | 3 . 8 | 5 . .
9 . . | . . 5 | . 2 .  ER=000064900000100000108009700309000070700080005050000208001900803000008000004210000


the one thing the referee doesn;t do is validate se ratings (for fear the referee may get bogged
down doing that and not paying attention to the game)
edit this entry is ignored by the referee because the ER is not a floating point number

[gsf]

pearls :
We could have 2 types of pearl :

• weak : the first move can't be a single
• strong : SS doesn't solve one cell

that's doable
just for accounting, or should there be bonuses: SCORE_PEARL_WEAK, SCORE_PEARL_STRONG?
if bonus then any suggestion for bonus values? 20/50, 25/100?
adding a pearl bonus could add a bit of strategy

maybe a .... SCORE_MORE*#extra.puzzles

That's fine for me.

any suggestion for SCORE_MORE? 10? 20?

[m_b_metcalf]

Off Topic.

For collectors, two Pearls:

Code: Select all

000012300000400000406008200508000060900070005020000109003500608000006000009340000 # SE 9.0
000012300000800000805004100308000060600050002040000901004600503000001000006940000 # SE 8.2

thanks for the pearl ref
for the first move using SS constraint order, I get locked-candidates for the 9.0 and naked quad for the 8.2
and from the ref it looks like that makes them not pearl

I don't use SS, nor do I intend to. According to the definition:

PEARLS, RUBIES, GARNETS, EMERALDS, ETC.

For quick reference, definitions of above terms :

PEARL :

A Pearl is a minimal sudoku where the first solvable cell is difficult to solve. The more difficult, the greater the pearl and its value.

both qualify according to SE. If I step through SE with them, the SE 9.0 puzzle requires a number of steps, starting at 2.6 and rising to 9.0, before solving the first cell. The 8.2 one is similar, going from 3.4 via 7.3, before the first cell is solved.

If it's hard to get agreement between programs, maybe we should just drop the idea.

Regards,

Mike Metcalf

[gsf]

both qualify according to SE. If I step through SE with them, the SE 9.0 puzzle requires a number of steps, starting at 2.6 and rising to 9.0, before solving the first cell. The 8.2 one is similar, going from 3.4 via 7.3, before the first cell is solved.

If it's hard to get agreement between programs, maybe we should just drop the idea.

aha, I read too fast and mistook "first solvable cell" for "first candidate elimination"
your're right that we will never get exact agreement between solvers
but we can get close enough
although the referee will not use SE to determine pearls -- there's better ways to burn cycles
thanks

[TTHsieh]

...adding a pearl bonus could add a bit of strategy...

Just like Sudoku Explainer evaluating the SE ratings.
If there is a simple and unique way to evaluate the pearl, then we can add the pearl's rule.

gsf, I think you can do it.

TTHsieh

[gsf]

...adding a pearl bonus could add a bit of strategy...

Just like Sudoku Explainer evaluating the SE ratings.
If there is a simple and unique way to evaluate the pearl, then we can add the pearl's rule.

gsf, I think you can do it.

with a bit of tweaking my solver can efficiently filter out simple sudoku constraint set pearls
there are no pearl entries in this thread yet, but these were found in the original thread

Code: Select all

000100000020003010004050006700000800008070005040001070000200000030009040005060009 # G  99 0004 JPF
020000080400030006007000400000801000010000050000502000006000900300090007080000020 # G  99 0016 TTHsieh
020000080400030005006000100000703000090000060000504000008000200500070004060000090 # G  99 0016 TTHsieh
020000080400070003006000400000502000050000090000908000004000100600030007080000020 # G  99 0016 TTHsieh
000000000001203400056070120060000080009000300030000040082060950003409200000000000 # Y  11 0017 JPF
000000000001203400056010720060000010008000300010000070074090850005408200000000000 # G  99 0017 JPF
001000200003000400540030061000605000007040800000801000130020079006000300002000100 # W2  8 0028 TTHsieh
001000200003000400540030061000701000007040500000809000380020054005000300002000900 # T4  6 0028 TTHsieh
001000200003000400420050013000607000007040800000102000950030028008000500002000300 # W2  8 0028 JPF
001000200003000400520060017000604000007020800000309000710090082009000700002000500 # W2  8 0028 JPF
001000200003000400240050031000304000006020100000507000160080027005000900002000600 # T4  6 0028 JPF
001000200003000400450030061000708000008020900000109000310060097004000600007000100 # Y  11 0028 JPF
000010000002000300034506210006000700700000008009000400027109630003000800000040000 # B3  4 0030 JPF
200000004010030090000605000005000200040300010003090600000807000090010020600000005 # X  10 0031 m_b_metcalf
100000002030040050000603000007000800040300060009060700000809000020030040500000001 # G  99 0031 Mauricio
100000002030010040000506000007000800050400030006090700000308000040020090200000008 # G  99 0031 JPF
010000020008000600000504000005020400060090030700000008020000090004000700000816000 # G  99 0032 m_b_metcalf
100000004020080050003000600000102000050040070000906000006000300070020080400000001 # G  99 0033 Eioru
200000009010070060008000300000204000030050010000906000005000200040060070900000005 # W3  9 0033 m_b_metcalf
100000004020080050003000600000102000050090070000807000006000300070020080400000001 # G  99 0033 Eioru
200000009010070080008000300000904000030050010000207000005000200040080070900000005 # W3  9 0033 m_b_metcalf
100000004020080050003000600000103000050040070000902000006000300070020080400000001 # Y  11 0033 Eioru
100000004020080050003000600000103000050040070000502000006000300070020080400000001 # G  99 0033 Eioru
100000004020080050003000600000102000050040070000506000006000300070020080400000001 # G  99 0033 Eioru
600000002050070090008000300000607000040020050000809000003000800070090040200000001 # G  99 0033 m_b_metcalf
200000008070040050001000300000102000090030070000405000006000100050060040800000002 # Y  11 0033 coloin
200000001090070030008000400000102000060030020000405000001000800070090060300000007 # G  99 0033 coloin
600000008010090040002000500000205000030040010000109000006000200040080030500000009 # W3  9 0033 m_b_metcalf
500000002070040090003000600000906000080020070000308000001000300050090080600000005 # G  99 0033 m_b_metcalf
400000002080030060005000100000702000090010080000608000002000500030060090700000004 # G  99 0033 m_b_metcalf
100000004020080050003000600000402000050090070000807000006000300070020080400000001 # G  99 0033 m_b_metcalf
100000002030040050006000700000108000090060030000902000003000100080050040200000007 # Y  11 0033 JPF
100000002090080010005000300000502000060010070000907000002000500040090060300000001 # G  99 0033 m_b_metcalf
100000000070386000005000200060900070020040030080005010003000400000712080000000006 # G  99 0034 m_b_metcalf
700000000010234000008000500020300080030090050060005010006000900000813040000000002 # B2  4 0034 wintder
100000000050638000002000400040700030080090050060001020003000900000452080000000006 # G  99 0034 m_b_metcalf
100900003024000050000000060000301008000050000900802000040000000060000420300007001 # G  99 0037 gsf
100200003045000060000000070000302008000060000900804000050000000070000540800001002 # G  99 0037 JPF
100200003045000060000000070000302001000060000800104000050000000070000450200009008 # G  99 0037 JPF
100000002030400050006000700080010000000639000000050030007000100050003080200000006 # G  99 0038 m_b_metcalf


the first comment column is my solver shorthand for the hardest constraint up to and including the first placement
col 2 is the constraint ordinal in the simple sudoku constraint list, 99 for G (guessing)
col 3 is the game number in the original thread
col 4 is the submitter

the first post will reflect the pearl addition

[JPF]

Code: Select all

 . . . | . 1 2 | 3 . .
 . . . | 4 . . | . . .
 4 . 1 | . . 5 | 6 . .
-------+-------+-------
 7 . 5 | . . . | . 1 .
 6 . . | . 2 . | . . 7
 . 3 . | . . . | 8 . 6
-------+-------+-------
 . . 3 | 6 . . | 1 . 9
 . . . | . . 1 | . . .
 . . 6 | 5 9 . | . . .     ER=9.5



JPF

[gsf]

Code: Select all

ER=9.5


I'm guessing you took my

Code: Select all

ER=000064900000100000108009700309000070700080005050000208001900803000008000004210000

as an entry
but that was malformed and the referee ignored it
because of that it has pegged you for consecutive entries

I modified the referee to treat any post with { grid [ER|SE]=* } as an attempted submission
attempted submissions from another player will act as separators for consecutive submissions

btw, that invalid ER=000* post did not match the current pattern

[JPF]

:::status:::

[g.r.emlin]

::: game 0003 status :::

[m_b_metcalf]

col 3 is the game number in the original thread

gsf, there was a massive SE 9.8 Pearl in the original thread in Game 2 (on page 4) that I don't see listed. Maybe the early games are not recognised as such? (Just to help with gremlin eradication.)

[Edit: sorry, it's here.]

Regards,

Mike Metcalf

[JPF]

I'm guessing you took my

Code: Select all

ER=000064900000100000108009700309000070700080005050000208001900803000008000004210000

as an entry

No ; actually, I forgot the new rule after a busy working day

maybe a .... SCORE_MORE*#extra.puzzles
...
any suggestion for SCORE_MORE? 10? 20?

Btw, SCORE_MORE=10 is too low as the last entries are by definition the most difficult to find.
The last points are around 50.
What about SCORE_MORE=50 ?

PS : this game (3) is still ongoing !

JPF