The New Sudoku Players' Forum

by **stumble** » Sat Dec 22, 2007 7:42 pm

Original puzzle BB12-19-07VeryHard

Code: Select all: .------------------------.------------------------.------------------------. | 134 14589 1345 | 2 3579 139 | 6 45789 589 | | 3 5689 7 | 4 3569 369 | 1 589 2 | | 124 124569 1245 | 8 5679 169 | 457 3 59 | :------------------------+------------------------+------------------------: | 5 1247 9 | 37 23478 234 | 38 128 6 | | 8 24 234 | 3569 1 23469 | 35 25 7 | | 6 127 123 | 357 23578 23 | 9 1258 4 | :------------------------+------------------------+------------------------: | 1247 3 1245 | 169 2469 8 | 457 145679 159 | | 9 147 8 | 136 346 5 | 2 1467 13 | | 124 1245 6 | 139 2349 7 | 3458 14589 13589 | '------------------------'------------------------'------------------------'

I used my ‘easy’ weapons down to a point where I start looking for AICs,Xychains. If they’re there I couldn’t find them. I started looking for Swordfish and practicing coloring in SudoCue. I THINK I made two Swordfish work on ‘2’ and ‘4’ candidates. Could you guys check them out, see if it’s correct or if I got right answer by accident?

Swordfish on ‘2’
C1 C3 C5

Code: Select all: .------------------.------------------.------------------. | 1 89 5 | 2 7 3 | 6 4 89 | | 3 689 7 | 4 569 69 | 1 589 2 | | 24* 69 24* | 8 569 1 | 7 3 59 | :------------------+------------------+------------------: | 5 12 9 | 7 38 4 | 38 12 6 | | 8 24 2-34 | 69 1 69 | 35 25 7 | | 6 7 13 | 5 38 2 | 9 18 4 | :------------------+------------------+------------------: | 7 3 124* | 19 249* 8 | 45 6 159 | | 9 14 8 | 136 46 5 | 2 7 13 | | 24* 5 6 | 139 249* 7 | 48 89 1389 | '------------------'------------------'------------------'

r5c3<>2

Swordfish on ‘4’
C1 C3 C7

Code: Select all: .------------------.------------------.------------------. | 1 89 5 | 2 7 3 | 6 4 89 | | 3 689 7 | 4 569 69 | 1 589 2 | | 24* 69 24* | 8 569 1 | 7 3 59 | :------------------+------------------+------------------: | 5 12 9 | 7 38 4 | 38 12 6 | | 8 24 3-4 | 69 1 69 | 35 25 7 | | 6 7 13 | 5 38 2 | 9 18 4 | :------------------+------------------+------------------: | 7 3 124* | 19 29 8 | 45* 6 159 | | 9 14 8 | 136 46 5 | 2 7 13 | | 24* 5 6 | 139 29 7 | 48* 89 1389 | '------------------'------------------'------------------'

r5c3<>4

by **daj95376** » Sat Dec 22, 2007 10:03 pm

Both of your Swordfish work, but at the expense of applying lesser techniques first.

Code: Select all: Locked Candidate (2) in [c2b4] => [r5c3]<>2 [r8c5]=h4 => [r8c2]=1 => [r4c2]=2 => [r5c2]=4 => [r5c3]<>4

by **stumble** » Sun Dec 23, 2007 12:07 am

daj95376 wrote:Both of your Swordfish work, but at the expense of applying lesser techniques first.

Code: Select all
Locked Candidate (2) in [c2b4] => [r5c3]<>2 [r8c5]=h4 => [r8c2]=1 => [r4c2]=2 => [r5c2]=4 => [r5c3]<>4

Thanks for the reassurance on my Swordfish. I'm afraid it took me nearly 3 minutes to figure out how you canceled the '2', I'm going to have to sift for locked candidates more carefully.

Your second method, some kind of chain, I don't recognize. Could you put a name to it so I can look it up and read about it? I can't see how it works.

by **daj95376** » Sun Dec 23, 2007 2:27 am

stumble wrote:Your second method, some kind of chain, I don't recognize. Could you put a name to it so I can look it up and read about it? I can't see how it works.

It's a basic Implication Chain that isn't used much anymore. However, it's perfectly suited for a sequence of singles. In your case, there's a hidden single in the second PM that leads to a cascade of naked singles that perform the same elimination as your Swordfish.

by **Pat** » Sun Dec 23, 2007 10:53 am

stumble wrote:Swordfish on 4
C1 C3 C7

Code: Select all
.------------------.------------------.------------------. | 1 89 5 | 2 7 3 | 6 4 89 | | 3 689 7 | 4 569 69 | 1 589 2 | | 24* 69 24* | 8 569 1 | 7 3 59 | :------------------+------------------+------------------: | 5 12 9 | 7 38 4 | 38 12 6 | | 8 24 3-4 | 69 1 69 | 35 25 7 | | 6 7 13 | 5 38 2 | 9 18 4 | :------------------+------------------+------------------: | 7 3 124* | 19 29 8 | 45* 6 159 | | 9 14 8 | 136 46 5 | 2 7 13 | | 24* 5 6 | 139 29 7 | 48* 89 1389 | '------------------'------------------'------------------'

r5c3<>4

well, yes, if the 4 for r379 are only in c137
then you have a valid Swordfish to exclude the digit at r5c3

however, you haven't shown how you managed to reach this position --
your earlier position still had more 4s in those rows !!
( look at box 8 )

have

4

by **stumble** » Sun Dec 23, 2007 8:29 pm

Pat wrote:well, yes, if the 4 for r379 are only in c137
then you have a valid Swordfish to exclude the digit at r5c3

however, you haven't shown how you managed to reach this position --
your earlier position still had more 4s in those rows !!
( look at box 8 )
if you have somehow excluded 4 in r79c5
then the remaining 4 in box 8 is a "hidden single"
and the puzzle is solved

If by 'earlier' you mean the 'original', yeah I skipped some of my steps. And those could well have been incorrect for all I know. Sorry if I was unscientific, I was mainly interested in whether I did the Swordfish correctly or not on that particular phase of the puzzle. I've taken to keeping a running 'history' of some of my puzzles (the inkjet cartridges are eating me up!).
I managed to find a few notes on this one:
XWing on '3' r9c5<>3
Swordfish on '2' r5c3<>2
Swordfish on '4' r7c5<>4, r9c5<>4
No idea if I got these earlier Swordfish right.
I'm pretty aware from my posts so far that I am always trying the 'hard' solutions - I just don't spot the easier ones.

by **stumble** » Sun Dec 23, 2007 8:42 pm

daj95376 wrote:
stumble wrote:Your second method, some kind of chain, I don't recognize. Could you put a name to it so I can look it up and read about it? I can't see how it works.

It's a basic Implication Chain that isn't used much anymore. However, it's perfectly suited for a sequence of singles. In your case, there's a hidden single in the second PM that leads to a cascade of naked singles that perform the same elimination as your Swordfish.

I'm using a SudoCue program that shows me naked singles. After the first Swordfish on '2', sure enough, the naked 4 shows up in r8c5. Well... look at all I learned about Swordfish anyway.

[r8c5]=h4 => [r8c2]=1 => [r4c2]=2 => [r5c2]=4 => [r5c3]<>4
how can you assume that r8c5 was 4 rather than 6? The 4s of both ends of the chain are true (unlike an XY chain). Target r5c3 can't see both ends of the chain. I found mention of a Double Implication in Sudopedia but it still didn't enlighten me. Maybe I should have started doing these things when I was 6.

by **Sudtyro** » Mon Dec 24, 2007 1:31 pm

stumble wrote:[r8c5]=h4 => [r8c2]=1 => [r4c2]=2 => [r5c2]=4 => [r5c3]<>4
how can you assume that r8c5 was 4 rather than 6? The 4s of both ends of the chain are true (unlike an XY chain).

The confusion may be that daj95376 is referring to your third grid position, which has the hidden 4 in b8. His Implication Chain (not an AIC) just shows the cascading singles that result from placing the 4 in r8c5.

by **stumble** » Mon Dec 24, 2007 4:16 pm

Sudtyro wrote:
stumble wrote:[r8c5]=h4 => [r8c2]=1 => [r4c2]=2 => [r5c2]=4 => [r5c3]<>4
how can you assume that r8c5 was 4 rather than 6? The 4s of both ends of the chain are true (unlike an XY chain).

The confusion may be that daj95376 is referring to your third grid position, which has the hidden 4 in b8. His Implication Chain (not an AIC) just shows the cascading singles that result from placing the 4 in r8c5.

Thanks, I finally get it. My initial posting was confused enough - the process by which I bumbled my way to my answer was messy enough so that when I posted, those 4's vanished into thin air between my grids.

The New Sudoku Players' Forum

how's my Swordfish?

how's my Swordfish?