View previous topic :: View next topic |
Author |
Message |
Marty R.
Joined: 12 Feb 2006 Posts: 5770 Location: Rochester, NY, USA
|
Posted: Tue Apr 04, 2006 5:08 pm Post subject: A humbling puzzle |
|
|
Just when I think I'm getting proficient, a puzzle like this comes along to quickly disabuse me of that notion. I've solved exactly five cells, all in the initial cross-hatching. I feel figuratively constipated; I can't eliminate anything.
There was an early jellyfish, but the one candidate I could erase was of zero help.
Code: | ----------------------------------------------------------
|6 1239 239 |2349 8 2349 |145 7 25 |
|5 279 2789 |1 69 24679 |48 3 268 |
|2378 1237 4 |237 5 2367 |18 9 268 |
----------------------------------------------------------
|23789 2379 6 |3589 4 389 |3589 258 1 |
|389 349 5 |6 2 1 |7 48 389 |
|1 2349 2389 |3589 7 389 |6 2458 3589 |
----------------------------------------------------------
|3479 8 1379 |479 169 4679 |2 156 3569 |
|2349 6 1239 |2489 19 5 |389 18 7 |
|279 5 1279 |2789 3 26789 |89 168 4 |
---------------------------------------------------------- |
|
|
Back to top |
|
|
keith
Joined: 19 Sep 2005 Posts: 3355 Location: near Detroit, Michigan, USA
|
Posted: Tue Apr 04, 2006 5:19 pm Post subject: Not much help |
|
|
Look at <6> in B3 and C9. There is an XY-wing in R8C8, but that is not much help.
Keith |
|
Back to top |
|
|
George Woods
Joined: 28 Mar 2006 Posts: 304 Location: Dorset UK
|
Posted: Tue Apr 04, 2006 10:54 pm Post subject: |
|
|
No expert but the 6 in box9 col 9 can be removed |
|
Back to top |
|
|
ravel Guest
|
Posted: Wed Apr 05, 2006 9:12 am Post subject: |
|
|
Pretty hard puzzle. Like Keith said, you can eliminate a 6 and 3 9's.
You can also show, that r1c9=2:
if r1c9=5, triple 148 in r123c7, r8c7=3, r9c7=9
no number for r7c9 left
But then ... ?? |
|
Back to top |
|
|
Marty R.
Joined: 12 Feb 2006 Posts: 5770 Location: Rochester, NY, USA
|
Posted: Wed Apr 05, 2006 4:56 pm Post subject: |
|
|
Thanks guys. I'll see where that takes me.
Quote: | You can also show, that r1c9=2:
if r1c9=5, triple 148 in r123c7, r8c7=3, r9c7=9
no number for r7c9 left |
Ravel, I'm not sure why there's no number for r7c9. The way I follow it is that with the "5" in r1c9 and the "3" and "9" in r8c7 and r9c7, a "6" is forced into r7c9. That then leaves no "6" and a duplicate "2" in box 3. Of course, either way proves that a "5" can't dwell in r1c9.
Once again, thanks to all. |
|
Back to top |
|
|
Guest
|
Posted: Thu Apr 06, 2006 7:45 am Post subject: |
|
|
The 6 in r7c9 can be eliminated with simple box(3)/line(col 9) intersection. |
|
Back to top |
|
|
Marty R.
Joined: 12 Feb 2006 Posts: 5770 Location: Rochester, NY, USA
|
Posted: Thu Apr 06, 2006 5:00 pm Post subject: |
|
|
Anonymous wrote: | The 6 in r7c9 can be eliminated with simple box(3)/line(col 9) intersection. |
Thanks, I was clearly asleep at the xx. |
|
Back to top |
|
|
ravel Guest
|
Posted: Thu Apr 06, 2006 5:24 pm Post subject: |
|
|
I managed to solve the puzzle in the meantime, but i needed 4 complicated and long steps (with more than 10 cells involved) to get 3 more numbers and 2 forcing chains after that.
I wonder, if there is an easier solution. |
|
Back to top |
|
|
Guest
|
Posted: Thu Apr 06, 2006 9:23 pm Post subject: |
|
|
I have written down my solution now:
After xy-wing and r1c9<:
Code: | *--------------------------------------------------------------------*
| 6 1 39 | 349 8 349 | 5 7 2 |
| 5 279 2789 | 1 69* 2679 | 4 3 68* |
| 2378* 237 4 | 237 5 2367 | 1 9 68* |
|----------------------+----------------------+----------------------|
| 23789 2379 6 | 3589 4 389 | 389 258 1 |
| 389* 349 5 | 6 2 1 | 7 48* 39 |
| 1 2349 2389 | 3589 7 389 | 6 2458 359 |
|----------------------+----------------------+----------------------|
| 3479 8 1379 | 479 169 4679 | 2 156 359 |
| 2349 6 1239 | 2489 19* 5 | 38 18* 7 |
| 279 5 1279 | 278 3 2678 | 89 168 4 |
*--------------------------------------------------------------------*
| If you look at the marked cells, you can see, that there are only 2 possibilities in this loop, where r3c1 and r5c1 are either 8 or not 8, the rest has to be one of the 2 candidates. From this follows, that
r469c8<>8 (either r5c8 or r8c8 must be 8)
r4c1<>8 (either r3c1 or r5c1 must be 8)
r2c6<>6 (either r2c5 or r2c9 must be 6)
r7c5<>9 (either r2c5 or r7c5 must be 9)
Next eliminate 1 from r9c8.
If r9c8=1, then r8c8=8 and
1. r8c5=1, r7c5=6, r7c8=5
2. r5c8=4, pair r46c8=25, r7c8<>5
=> r9c8=6
Eliminate 8 from r9c7:
If r9c7=8,
1. pair r9c45=27, r9c1=9
2. r5c8=8 and r7c9=9,r5c9=3 -> r5c1=9
Then we come here:
Code: | *-----------------------------------------------------------*
| 6 1 39 | 349 8 349 | 5 7 2 |
| 5 279 2789 | 1 69 279 | 4 3 68 |
| 2378 237 4 | 237 5 2367 | 1 9 68 |
|-------------------+-------------------+-------------------|
| 2379 2379 6 | 3589 4 389 | 38 25 1 |
| 389 349 5 | 6 2 1 | 7 48 39 |
| 1 2349 2389 | 3589 7 389 | 6 245 359 |
|-------------------+-------------------+-------------------|
| 3479 8 379 | 479 16 4679 | 2 15 35 |
| 2349 6 239 | 249 19 5 | 38 18 7 |
| 27 5 1 | 278 3 278 | 9 6 4 |
*-----------------------------------------------------------*
|
From this point, susser found a better solution than mine was. I give it in my words:
I dont know, how this fish monster is called, but a 9 in r8c3 leaves an x-wing in r17c46, which kills all 9's in box. So r8c3<>9.
if r5c1=9:
r5c8=8
r5c9=9, r4c7=8
=>r5c1<>9
if r6c3=2:
r8c3=3
r5c1=8,r5c8=4,r8c8=1,r8c7=3
=> r6c3<>2
If r2c3=8:
r7c3=7
r8c3=2,r9c1=7
=> r2c3=8, r2c9=8
This finally solves the puzzle. |
|
Back to top |
|
|
ravel Guest
|
Posted: Thu Apr 06, 2006 9:26 pm Post subject: |
|
|
It was me |
|
Back to top |
|
|
David Bryant
Joined: 29 Jul 2005 Posts: 559 Location: Denver, Colorado
|
Posted: Thu Apr 06, 2006 10:07 pm Post subject: One (long) chain will do the trick. |
|
|
Hi, Ravel!
Thanks for the explanation. Here's a slightly simpler way to solve this puzzle.
After incorporating all the excellent suggestions people had made in this forum, and making all the obvious moves that followed I arrived at this position.
Code: | 6 139 39 349 8 349 5 7 2
5 279 2789 1 69 2679 4 3 68
2378 237 4 237 5 2367 1 9 68
23789 2379 6 3589 4 389 389 258 1
389 349 5 6 2 1 7 48 39
1 2349 2389 3589 7 389 6 2458 3589
3479 8 1379 479 169 4679 2 156 359
2349 6 1239 2489 19 5 38 18 7
279 5 1279 278 3 2678 89 168 4 |
We can prove that r2c5 <> 9 as follows.
r2c5 = 9 ==> r8c5 = 1 ==> r7c5 = 6
r8c5 = 1 ==> r8c8 = 8 ==> r8c7 = 3 & r9c7 = 9
r8c8 = 8 ==> r5c8 = 4 ==> r5c1 = 8
With these changes applied the grid looks like this.
Code: | 6 139 39 34 8 34 5 7 2
5 27 278 1 9 267 4 3 68
237 237 4 237 5 2367 1 9 68
2379 2379 6 3589 4 389 8 25 1
8 39 5 6 2 1 7 4 39
1 2349 239 3589 7 389 6 25 39
3479 8 1379 479 6 479 2 15 59
249 6 29 2489 1 5 3 8 7
27 5 127 278 3 278 9 16 4 |
Now we can see the contradiction:
{3, 4} pair in r1c4&6 ==> r1c3 = 9
r9c7 = 9 ==> r7c9 = 5 ==> r7c8 = 1
r1c3 = 9 ==> r8c3 = 2 ==> r6c3 = 3 ==> r7c3 = 7
and there are now no candidates left for r9c1.
With r2c5 = 6 the puzzle is easily completed. dcb |
|
Back to top |
|
|
ravel Guest
|
Posted: Fri Apr 07, 2006 9:31 am Post subject: |
|
|
Hi David,
fine solution.
I had a similar proof to show that r3c1<>9, but after setting the 2 numbers.
So i am approved now, that this is a hard puzzle and that there is more than one way to skin a sudoku |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|