View previous topic :: View next topic |
Author |
Message |
| mumdigau
| Joined: 07 Jun 2008 | Posts: 3 | : | | Items |
|
Posted: Sat Jun 07, 2008 9:53 am Post subject: help with Nice Loop |
|
|
Hi,
I just started to work with Nice Loops. Here is an example, I want to ask your help. After some elimination I reached this grid:
Code: |
*--------------------------------------------------------------------*
| 459 1 249 | 8 45 3 | 7 6 2459 |
| 456 8 246 | 9 457 457 | 1 3 245 |
| 7 359 349 | 2 1 6 | 59 459 8 |
|----------------------+----------------------+----------------------|
| 2349 239 5 | 347 8 79 | 239 1 6 |
| 8 239 7 | 6 59 1 | 4 259 359 |
| 1 6 349 | 34 2 459 | 8 579 3579 |
|----------------------+----------------------+----------------------|
| 269 279 8 | 5 3 479 | 269 2479 1 |
| 3569 4 369 | 1 679 2 | 3569 8 3579 |
| 23569 23579 1 | 47 4679 8 | 23569 2479 3479 |
*--------------------------------------------------------------------*
|
I found a discontinuous Nice Loop at R4C1 which can be closed with a weak link either on 3 or 9:
[R4C1]=4=[R4C4]=3=[R6C4]-3-[R3C3]-3/9-[R4C1] ==> R4C1<>3, R4C1<>9
I'd like to ask you if loop and conclusion are correct.
Many thanks for your feedback.
Best regards
mumdigau |
|
Back to top |
|
|
| hobiwan
| Joined: 11 Feb 2008 | Posts: 83 | : | | Items |
|
Posted: Sat Jun 07, 2008 1:24 pm Post subject: |
|
|
Hello mumdigau, welcome to the forums!
As for your question: Two consecutive cells in a Nice Loop must see each other (be in the same row, column or block). Therefore
[R6C4]-3-[R3C3]
cannot be right (r6c4 and r3c3 are in different rows, columns and blocks).
You could go r6c4 -3- r6c3, but from there you don't have a strong link to continue.
A possible short Discontinous nice Loop starts with r2c5 -5- r5c5 ..., another one with r6c8 -5- r6c6 ...
Perhaps you can complete them? |
|
Back to top |
|
|
| mumdigau
| Joined: 07 Jun 2008 | Posts: 3 | : | | Items |
|
Posted: Sat Jun 07, 2008 3:03 pm Post subject: |
|
|
Hi hobiwan,
Of course, you're right. Was a typo, and I solved the grid by means of your proposals. Thanks!
In the mean time I think I've understood the basics of Nice Loops (rather errorprone on manual work, though). But there is one situation left I can't figure out. Have a look on this grid:
Code: |
.------------------.------------------.------------------.
| 67 8 5 | 379 139 4 | 36 2 19 |
| 1 29 4 | 359 2389 6 | 58 358 7 |
| 3 2679 679 | 579 1289 129 | 4 568 19 |
:------------------+------------------+------------------:
| 4 35 2 | 39 6 139 | 159 7 8 |
| 67 3567 367 | 2 139 8 | 159 15 4 |
| 9 1 8 | 4 5 7 | 36 36 2 |
:------------------+------------------+------------------:
| 5 79 1 | 6 2479 29 | 78 48 3 |
| 2 3679 3679 | 8 3479 39 | 17 14 5 |
| 8 4 37 | 1 37 5 | 2 9 6 |
'------------------'------------------'------------------'
|
One continuous Nice Loop I discovered (hopefully correct) is
-[R3C3]=9=[R8C3]-9-[R8C6]=3=[R4C6]=1=[R3C6]-1-[R3C9]-9-[R3C3]=
The conclusions then should be (but aren't):
R4C6=13
R8C2<>9, R8C5<>9
R3C5<>1
R3C2<>9, R3C4<>9, R3C6<>9
Questions:
1. Why is R3C2<>9 wrong (sees both R3C3 and R3C9)?
2. Can R3C6 be included when 9's are discarded even though R3C6 belongs to the loop?
Thanks for your help.
Best regards
mumdigau |
|
Back to top |
|
|
| hobiwan
| Joined: 11 Feb 2008 | Posts: 83 | : | | Items |
|
Posted: Sat Jun 07, 2008 10:54 pm Post subject: |
|
|
mumdigau wrote: | -[R3C3]=9=[R8C3]-9-[R8C6]=3=[R4C6]=1=[R3C6]-1-[R3C9]-9-[R3C3]= |
Unfortunately you have an error here too:
r8c6 =3= r4c6
should be
r8c6 -3- r4c6
(r8c3 -9- r8c6 means, that r8c6<>9, therefore r8c6=3 => r4c6<>3). You are stuck again at this state, because you don't have a strong link to continue.
My solver can find only two continuous nice loops, both not very effective:
Continuous Nice Loop r3c8 =5= r3c4 =7= r1c4 -7- r1c1 -6- r1c7 =6= r3c8 => r3c4<>9
Continuous Nice Loop r3c4 =5= r3c8 =6= r1c7 -6- r1c1 -7- r1c4 =7= r3c4 => r3c8<>8
plus a couple of discontinuous nice loops.
mumdigau wrote: | 2. Can R3C6 be included when 9's are discarded even though R3C6 belongs to the loop? |
I don't know, I am afraid. Somebody else will have to answer this. |
|
Back to top |
|
|
| mumdigau
| Joined: 07 Jun 2008 | Posts: 3 | : | | Items |
|
Posted: Sat Jun 07, 2008 11:49 pm Post subject: |
|
|
hobiwan wrote: | mumdigau wrote: | -[R3C3]=9=[R8C3]-9-[R8C6]=3=[R4C6]=1=[R3C6]-1-[R3C9]-9-[R3C3]= |
Unfortunately you have an error here too:
r8c6 =3= r4c6
should be
r8c6 -3- r4c. |
Yeah, you're right again. Thanks for help. Still have to practice. As I said Nice Loops are very strong, but errorprone (at least for me).
You mentioned your solver. Can it check for correctness of Nice Loops? (Just to verify if one doesn't spoil the grid. I wouldn't use it to search for Loops.) Any link?
Best regards
mumdigau |
|
Back to top |
|
|
| hobiwan
| Joined: 11 Feb 2008 | Posts: 83 | : | | Items |
|
Posted: Sun Jun 08, 2008 7:29 pm Post subject: |
|
|
mumdigau wrote: | You mentioned your solver. Can it check for correctness of Nice Loops? (Just to verify if one doesn't spoil the grid. I wouldn't use it to search for Loops.) Any link?
Best regards
mumdigau |
It can show you when you remove a candidate that belongs to the solution or when you try to set a wrong value in a cell. Unfortunately it's still work in progress so I have not released it yet (and I am not sure if I ever will). Plus the UI is in german. But there should be a lot of free sudoku programs available with the same feature, perhaps somebody else can provide a link. |
|
Back to top |
|
|
| kyriako
| Joined: 12 Jun 2008 | Posts: 16 | : | Location: Germany | Items |
|
Posted: Tue Jul 08, 2008 7:40 pm Post subject: Nice Loops software |
|
|
Hello mumdigau, hello hobiwan,
I have been working with nice loops recently and implemented the method.
Download here:
http://homepages.compuserve.de/meinelg
My question:
how do you produce these nicely formatted sudoku examples? Is it the output of a program or do type them manually?
Kyriakos. |
|
Back to top |
|
|
| hobiwan
| Joined: 11 Feb 2008 | Posts: 83 | : | | Items |
|
Posted: Tue Jul 08, 2008 8:05 pm Post subject: Re: Nice Loops software |
|
|
kyriako wrote: | My question:
how do you produce these nicely formatted sudoku examples? Is it the output of a program or do type them manually? |
Most sudoku programs can copy candidate grids of the current state in text format to the clipboard. As for possible steps, I copy them directly from my own solver. I cannot produce candidate grids with cells marked according to a step, but I suspect that some users here have done this with their solvers too (in my solver its one of those "not yet implemented" menu items ).
Solutions that my solver cant produce I have to type manually. |
|
Back to top |
|
|
|