The New Sudoku Players' Forum

[daj95376]

I could replace the normal step output with the output used in the all steps dialog or a combination of both (which should be easier to parse). Example:
Instead of

Code: Select all

Skyscraper: 9 in r1c2,r9c3 (connected by r19c8) => r3c3,r8c2<>9

it could say:

Code: Select all

9 (2):r3c3,r8c2<>9 (Skyscraper)

(read: eliminated digits + number of eliminations in parenthesis).

For setting a digit it would be:

Code: Select all

Forcing Net Verity: r9c8=4

If you want more of a statistic thats can probably done rather easily too, but I would need a more detailed request.

One minor tweak for all possible steps:

Code: Select all

9 (k):r3c3,r8c2<>9 (Skyscraper r19c8)


Where k = 2 + progress_measure

[StrmCkr]

http://forum.enjoysudoku.com/viewtopic.php?p=60757#p60757
...found it...

[hobiwan]

I decided to make another release befor tackling the rating/heuristics problem. Some changes (for a complete list please see the change log:

Slight changes to the GUI (menus, configuration)
Dual 2-String-Kites and Dual ERs (optional)
Sashimi Fish (optional)
Death Blossom
ALS overlap (optional for adjacent ALS)
Statistics option in batch solve
Drag and Drop for ordering steps
Fish search can be restricted to specific candidates (all steps only)

ALS in Nice Loops/AICs and Forcing Chains/Nets are now optional and disabled by default: They tended to cover all ALS steps and took rather long to compute.

[StrmCkr]

did you add in dual skyscrapers as well?
or is it still left as two seperated views.

[hobiwan]

did you add in dual skyscrapers as well?
or is it still left as two seperated views.

I had troubles with the concept of Dual Skyscrapers so I left them out. You will have to use the two seperate views.

[wintder]

did you add in dual skyscrapers as well?
or is it still left as two seperated views.

I had troubles with the concept of Dual Skyscrapers so I left them out. You will have to use the two seperate views.

Just another penny tossed in.

When you find a sky, because you currently only see half, could you
mirror it and then find the other part?
(I know two mirrors may be involved.)

[StrmCkr]

i found some glitches. in version 1.2

when solving if i add an incorrect clues

the pm board does not delete the clues you entered that is incorrect from line of sight cells.

edit: wintder - the dual view for skys are when two sperate skys use the same set of nodes.

can be in polar oppisit direction.{ie if one is rows the other is columns.}
and they can be directly mirrored.

the complexity is that more then "2" {dual} skys can share the same nodes.

if remeber correctly i had an example with 3+ skys off the same groupd of nodes.

resolution is a nightmare to sort out i would solve each as their own,
only have the dual shown if they share nodes.

[daj95376]

Does anyone have an example/pattern for the discussion on "dual Skyscrapers"? Do either of these qualify?

Code: Select all

 Pattern #1
 +-----------------------------------+
 |  .  .  .  |  .  .  .  |  3  .  3  |
 |  .  .  3  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  3  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  3  .  |
 |  .  3  .  |  3  .  .  |  .  .  .  |
 |  3  .  .  |  .  .  3  |  .  .  .  |
 |-----------+-----------+-----------|
 |  3  .  .  |  3  .  .  |  .  .  3  |
 |  3  .  .  |  .  .  .  |  3  .  3  |
 |  3  3  .  |  .  .  3  |  3  .  .  |
 +-----------------------------------+

 Skyscraper c26r9   <> 3  [r5c4],[r6c1]
 Skyscraper c24r5   <> 3  [r7c1],[r9c6]


Code: Select all

 Pattern #2
 +-----------------------------------+
 |  .  .  1  |  .  .  .  |  .  1  1  |
 |  .  .  .  |  .  1  .  |  .  .  .  |
 |  1  .  .  |  .  .  .  |  .  1  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  1  .  .  |
 |  .  .  .  |  .  .  1  |  .  .  .  |
 |  .  1  .  |  .  .  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  1  .  .  |  .  .  .  |
 |  1  .  1  |  .  .  .  |  .  1  1  |
 |  .  .  1  |  .  .  .  |  .  1  .  |
 +-----------------------------------+

 Skyscraper c19r8   <> 1  [r1c3],[r3c8]
 Skyscraper r39c8   <> 1  [r1c3],[r8c1]


Code: Select all

 Pattern #3
 +-----------------------------------+
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  / r45 /  |  /  /  /  |  / s45 /  | <-- strong link on <4> and <5>
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  /  / t45 |  /  /  /  |  / u45 /  | <-- strong link on <4> and <5>
 |  .  .  .  |  .  .  .  |  .  .  .  |
 +-----------------------------------+

 Skyscraper r28c8   <> 4  [r13c3],[r79c2]
 Skyscraper r28c8   <> 5  [r13c3],[r79c2]


[StrmCkr]

i have examples:

i have to find them i just rebuild my computer again...

damn hacker!

or i'll just use daj's

Pattern #3: two seperated functions diffrent digits.
not a dual. paired yes i usually list it under remote pairs.


this is a dual:(pattern 1) {it shares mutiple same points}
@ are the linked nodes.

normal elimination due to same links.

Code: Select all

|-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  3  .  3  |
 |  .  .  3  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  3  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  3  .  |
 |  . 3@ . |  3@  .  . |  .  .  .  |
 |  3  .  . |  .  .  3#  |  .  .  .  |
 |-----------+-----------+-----------|
 |  3  .  .  | 3#  .  .  |  .  .  3  |
 |  3  .  .  |  .  .  .  |  3  .  3  |
 |  3  3@ .  |  .  . 3#  |  3  .  .  |
 +-----------------------------------+

eliminations are
R7C123
R9C456
&
R6C123
R5C456


this is also 1 with 1 shared cell.

overlapping sections dual coverage in boxes
{neting possible extra eliminations.}

Pattern #2

Code: Select all

 +-----------------------------------+
 |  .  .  1  |  .  .  .  |  .  N  1@ |
 |  .  .  .  |  .  1  .  |  .  .  .  |
 |  1# .  .  |  .  .  .  |  .  1@ .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  1  .  .  |
 |  .  .  .  |  .  .  1  |  .  .  .  |
 |  .  1  .  |  .  .  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  1  .  .  |  .  .  .  |
 |  1@ .  N  |  .  .  .  |  .  N  1@ |
 |  .  .  1@ |  .  .  .  |  .  1@ .  |
 +-----------------------------------+

# is the linked nodes for this example:

elminations are:

R1C123
R3C789

and

R123C3
R789C1

i also belive that the N cells are removed from double coverage since they are linked.

have to check on that one though...

how to define them as duals is an open question...

[hobiwan]

i found some glitches. in version 1.2

when solving if i add an incorrect clues

the pm board does not delete the clues you entered that is incorrect from line of sight cells.

It does remove them, but in all cells where the removed candidate should be placed, it is painted light red (it's a deviation from the correct solution). If you don't like that behaviour, uncheck "Show Deviations" in the Options menu.

im also 90% confident that death bloosom is not activating correctly.

Do you have an example? (Death Blossom is not active by default, you have to enable it manually)

Dual Skyscrapers:
Thanks for the discussion. I am following it closely even if I don't actively participate.

[StrmCkr]

thanks for the tip

i reviewed the deathbloosoms again... figured them out again.
they working correctly now. just a little diffrent the i expected to when i first reviewed them...


ps. i found some intresting things for W-wings as well..

with the aid of ron

he found Ring types and i found grouped...

http://forum.enjoysudoku.com/viewtopic.php?p=63874#p63874

&
http://forum.enjoysudoku.com/viewtopic.php?p=67793#p67793

if your intrested in getting those running as well..

[wintder]

Code: Select all

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

Not a big issue but I notice that with my current settings in the summary window the sword is listed as a franken swordfish while in the solution path window it is more correctly listed as a swordfish.

[hobiwan]

wintder,

could you please send me the contents of your hodoku.hcfg file (in your temp directory) as PM?

Thanks

[StrmCkr]

the error is more specifically

in the search for
moves
configuration.

if you brign the puzzle to this state

Code: Select all

.--------------.---------------.-------------.
| 24   8   36  | 1     5  23   | 7    46  9  |
| 27   9   57  | 4     6  8    | 25   1   3  |
| 245  1   36  | 23    7  9    | 46   8   25 |
:--------------+---------------+-------------:
| 3    57  8   | 57    1  4    | 29   29  6  |
| 1    4   2   | 9     3  6    | 8    5   7  |
| 6    57  9   | 8     2  57   | 1    3   4  |
:--------------+---------------+-------------:
| 8    3   457 | 6     9  27   | 245  24  1  |
| 9    26  45  | 235   8  1    | 36   7   25 |
| 57   26  1   | 2357  4  2357 | 369  69  8  |
'--------------'---------------'-------------'

and do a search for
basic fish or basic/franken

the sword does not show up.

but if you do a search for fish
that is basic +franken/mutant
the sword is found.

i also noticed that no basic swords shows up until i do a search for basic/franken mutant.

there is two on this grid state.

[hobiwan]

wintder and StrmCkr,

thanks for your infos.

wintder, you have Franken Swordfish before Swordfish in your solver configuration. The search for fish is exhaustive, so searching for Franken Fish returns Basic Fish as well. I didn't think of this case, so I took no precautions against it: It is found as a Swordfish, so listet as Swordfish in the solution path, but searched for as Franken Swordfish so listed as Franken Swordfish in the summary. I will correct this in the next version.

StrmCkr, this is obviously an error with fish ordering (resulting from the option "only one fish per elimination"). I have to investigate that a bit more.