The New Sudoku Players' Forum

[g.r.emlin]

::: adjustment 2m57s 2013-10-02+16:34:00-0000 C -50 :::

[m_b_metcalf]

Code: Select all

 1 . . . 2 . . . 3
 . . 4 . . . . 5 .
 . 6 . . . . 7 . .
 . . . . 1 6 8 . .
 3 . . 8 9 . . . .
 . . . 7 . . . . 2
 . . 5 9 . . 4 . .
 . 7 . . . . . 6 .
 8 . . . . 1 . . 9   ED=9.4/9.4/9.4

Welcome blue (you seem to have rescued g.r.emlin from his madness).

[blue]

Code: Select all

1 . . . 2 . . . 6
. . 3 . . . . 4 .
. 5 . . . . 7 . .
. . . . 6 3 9 . .
2 . . 5 9 . . . .
. . . 4 . . . . 8
. . 7 9 . . 6 . .
. 4 . . . . . 3 .
8 . . . . 1 . . 9    ED=10.0/10.0/10.0


Hi Mike.

I'm only good for a few plays here.
Pat's post and your reply, got me curious ...

[gsf]

g.r.emlin's digs at at&t research have been powered down
this curiously coincides with my termination from the same organization effective oct 10
I depart landing on my feet after an almost 30 year ride along with a group of 20 research colleagues
most with ~30 years experience, many also being at&t fellows
I have a suspicion that g.r.emlin will take residence on one of my home machines
there could be a few referee glitches until we find out

when things settle down I'll retool and rejoin the game

[m_b_metcalf]

Code: Select all

 1 . . . 2 . . . 3
 . . 4 . . . . 5 .
 . 6 . . . . 7 . .
 . . . . 3 7 8 . .
 3 . . 9 6 . . . .
 . . . 1 . . . . 2
 . . 5 8 . . 4 . .
 . 7 . . . . . 9 .
 9 . . . . 1 . . 8   ED=9.3/9.3/9.3

blue, I'll be travelling from Saturday anyway ...

[m_b_metcalf]

::: comment :::

g.r.emlin's digs at at&t research have been powered down
this curiously coincides with my termination from the same organization effective oct 10
I depart landing on my feet after an almost 30 year ride along with a group of 20 research colleagues
most with ~30 years experience, many also being at&t fellows

Here's wishing you a long and fruitful retirement!

[blue]

Code: Select all

6 . . . 2 . . . 9
. . 7 . . . . 4 .
. 5 . . . . 7 . .
. . . . 4 8 6 . .
9 . . 5 6 . . . .
. . . 3 . . . . 8
. . 4 6 . . 5 . .
. 7 . . . . . 3 .
8 . . . . 1 . . 2   ED=8.5/8.5/8.5


Mike, shall I keep them coming, or stand down ?

Best wishes for you, Glenn -- sounds like "me time" now

Best Regards,
Blue.

[m_b_metcalf]

Code: Select all

 1 . . . 2 . . . 3
 . . 4 . . . . 5 .
 . 6 . . . . 7 . .
 . . . . 1 6 8 . .
 3 . . 8 4 . . . .
 . . . 9 . . . . 2
 . . 7 4 . . 9 . .
 . 5 . . . . . 6 .
 8 . . . . 1 . . 7   ED=10.5/10.5/9.9

You can keep 'em coming, but we're running unsupervised now, while g.r.emlin moves house.

[blue]

Code: Select all

9 . . . 6 . . . 8
. . 3 . . . . 4 .
. 5 . . . . 7 . .
. . . . 7 1 9 . .
1 . . 6 9 . . . .
. . . 4 . . . . 2
. . 7 8 . . 3 . .
. 4 . . . . . 5 .
2 . . . . 6 . . 1   ED=10.4/10.4/10.4


Best I've got for diamonds.
Unsupervised, yes.

[Pat]

the "diamonds" have the tough move at start
which makes it relatively-easy to spot this tough move

puzzles which start with one-or-more-"singles"
are more interesting to solve,
i must first do the "singles" and then re-digest the puzzle
before i can find the tough move

anyway i never did figure out how to find "diamonds" for the Patterns Game. generating thousands of puzzles, some "diamonds" may just show up---

[m_b_metcalf]

Code: Select all

 1 . . . 2 . . . 3
 . . 4 . . . . 5 .
 . 6 . . . . 7 . .
 . . . . 3 8 9 . .
 3 . . 9 6 . . . .
 . . . 5 . . . . 2
 . . 5 8 . . 4 . .
 . 7 . . . . . 6 .
 9 . . . . 1 . . 8   ED=9.2/9.2/9.2

[m_b_metcalf]

anyway i never did figure out how to find "diamonds" for the Patterns Game. generating thousands of puzzles, some "diamonds" may just show up---

... most of the thousands are instantly rejected because thay start with singles, etc.

[champagne]

the "diamonds" have the tough move at start
which makes it relatively-easy to spot this tough move

puzzles which start with one-or-more-"singles"
are more interesting to solve,
i must first do the "singles" and then re-digest the puzzle
before i can find the tough move

anyway i never did figure out how to find "diamonds" for the Patterns Game. generating thousands of puzzles, some "diamonds" may just show up---

My experience is not base on thousands, but on millions of puzzles, but it's not the key point.

Usually, the hardest are in the vicinity of other hard, so I would say

start a vicinity generation from any seed

select highest ratings as new seeds and loop

diamonds will come on top of the highest ratings

(but be prepared to filter millions of puzzles, so use a highly performing program)

Usually the highest rating come in the first day, sometimes in the first hour.
Some diamonds are tougher to find, always the same.

[blue]

Code: Select all

1 . . . 2 . . . 9
. . 3 . . . . 4 .
. 5 . . . . 3 . .
. . . . 4 7 5 . .
2 . . 5 6 . . . .
. . . 9 . . . . 1
. . 4 6 . . 7 . .
. 7 . . . . . 3 .
6 . . . . 8 . . 4   ED=8.9/8.9/8.9


anyway i never did figure out how to find "diamonds" for the Patterns Game. generating thousands of puzzles, some "diamonds" may just show up---

As near as I can tell, that's about it.
The pros might have better ideas.

[ Posted before seeing champagne's comments ]

[m_b_metcalf]

Code: Select all

 1 . . . 2 . . . 3
 . . 4 . . . . 5 .
 . 6 . . . . 7 . .
 . . . . 4 3 8 . .
 3 . . 5 8 . . . .
 . . . 7 . . . . 2
 . . 5 9 . . 4 . .
 . 7 . . . . . 6 .
 8 . . . . 1 . . 9   ED=9.1/9.1/9.1