| View previous topic :: View next topic |
| Author |
Message |
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
 Posted: Sat Nov 21, 2009 8:14 am Post subject: Free Press Nov 20, 2009 |
 |
|
The reason I've stopped doing these, is the difficulty is so variable. | Code: | Puzzle: FP112009
+-------+-------+-------+
| . . 3 | 4 . . | . 8 . |
| . 9 . | . 5 7 | 4 . . |
| . . 8 | . 3 . | 2 . . |
+-------+-------+-------+
| . . . | 2 . . | 9 . . |
| 4 8 . | . . . | . 5 3 |
| . . 9 | . . 3 | . . . |
+-------+-------+-------+
| . . 5 | . 9 . | 3 . . |
| . . 4 | 3 7 . | . 9 . |
| . 3 . | . . 6 | 1 . . |
+-------+-------+-------+ |
Keith |
|
| Back to top |
|
 |
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sat Nov 21, 2009 8:16 am Post subject: |
|
|
After basics: | Code: | +----------------+----------------+----------------+
| 1567 1567 3 | 4 16 2 | 567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 1567 4 8 | 69 3 19 | 2 167 1567 |
+----------------+----------------+----------------+
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 1567 1567 9 | 67 4 3 | 8 1267 1267 |
+----------------+----------------+----------------+
| 8 26 5 | 1 9 4 | 3 267 267 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+----------------+----------------+----------------+ |
At this point, Sudoku Susser says "guess" (use Nishio). I am able to make some headway, but not much.
Keith |
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Sat Nov 21, 2009 3:59 pm Post subject: |
|
|
I give up...............
Ted |
|
| Back to top |
|
|
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sat Nov 21, 2009 5:57 pm Post subject: |
|
|
There is a skyscraper that takes out <7> in R1C2. Leading to: | Code: | +----------------+----------------+----------------+
| 1567 15-6 3 | 4 16# 2 | 567# 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 1567 4 8 | 69 3 19 | 2 167 1567 |
+----------------+----------------+----------------+
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16# 19 | 67# 5 3 |
| 156 1567 9 | 67 4 3 | 8 1267 1267 |
+----------------+----------------+----------------+
| 8 26b 5 | 1 9 4 | 3 267 267 |
| 16 126 4 | 3 7 8 | 56a 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+----------------+----------------+----------------+ |
I have no idea what this is called, but: (Is true means is 6.)
Either the X-wing # is true, or a is true. If a is true, b is true, eliminating 6 in R1C2.
Now, I am stuck.
Keith |
|
| Back to top |
|
|
Mogulmeister
Joined: 03 May 2007
Posts: 1151
|
| Posted: Sat Nov 21, 2009 6:18 pm Post subject: Extreme |
|
|
Bent fin ? Extended fin ? Nice move anyway.
This looks for all the world like an Extreme. I found a couple of AICs but it is very slow progress and have not got a solution.
Last edited by Mogulmeister on Sat Nov 21, 2009 6:26 pm; edited 1 time in total |
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Sat Nov 21, 2009 6:25 pm Post subject: |
|
|
Keith, I call that pattern a Kraken x-wing.
I found it plus multi-coloring on 7 that deletes the 7 from the same cell, r1c2. (Note that my code after basics still had a 7 in r1c2 & r6c1.)
At that point, I hit the wall.
Ted |
|
| Back to top |
|
|
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sat Nov 21, 2009 6:36 pm Post subject: |
|
|
Ted, the skyscraper results in removing those two 7s.
I think daj suggested this:
| Code: | +----------------+----------------+----------------+
| 1567 15 3 | 4 16c 2 | 567 8 9 |
| 2 9 16d | 8 5 7 | 4 3 16b |
|15-67 4 8 | 69e 3 19 | 2 167 1567 |
+----------------+----------------+----------------+
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16a 19 | 67 5 3 |
| 156 1567 9 | 67 4 3 | 8 1267 1267 |
+----------------+----------------+----------------+
| 8 26 5 | 1 9 4 | 3 267 267 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+----------------+----------------+----------------+ |
If a is 1, it forces 1 in b. If a is 6, c is 1. So, bc are pincers in 1.
So then, one or both of bc are not 6, and we can build the skyscraper bd and ce. de are pincers in 6, making the elimination shown.
Not much help, though.
Keith |
|
| Back to top |
|
|
Mogulmeister
Joined: 03 May 2007
Posts: 1151
|
| Posted: Sat Nov 21, 2009 6:51 pm Post subject: |
|
|
If you put a 6 in r1c1 you force a contradiction as it will also force a 6 in r1c5
Follow the sequence of events going from putting a 6 in r1c1 to r1c7 (which is forced to 7) to r5c7 which is forced to being a 6 which in turn forces r5c5 to being a 1 which forces r1c5 to be a 6. So r1c1 can not be 6! |
|
| Back to top |
|
|
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sat Nov 21, 2009 7:22 pm Post subject: |
|
|
| Mogulmeister wrote: | If you put a 6 in r1c1 you force a contradiction as it will also force a 6 in r1c5
Follow the sequence of events going from putting a 6 in r1c1 to r1c7 (which is forced to 7) to r5c7 which is forced to being a 6 which in turn forces r5c5 to being a 1 which forces r1c5 to be a 6. So r1c1 can not be 6! |
I just came to the same conclusion as follows: | Code: | +----------------+----------------+----------------+
|1567a 15 3 | 4 16c 2 | 567b 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 157 4 8 | 69 3 19 | 2 167 1567 |
+----------------+----------------+----------------+
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16d 19 | 67e 5 3 |
| 156 1567 9 | 67 4 3 | 8 1267 1267 |
+----------------+----------------+----------------+
| 8 26 5 | 1 9 4 | 3 267 267 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+----------------+----------------+----------------+ |
If a is 6, b is 7.
If a is 6, c is 1, d is 6, e is 7.
Contradiction (two 7s in C7), a cannot be 6.
Not pretty, but it works to solve the puzzle.
Keith
edit: Another thing to look at. Any value in R1C5 forces 1 in R2C9. |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Sat Nov 21, 2009 7:46 pm Post subject: |
|
|
I have an interesting (to me) almost Remote Pair. The rest is just playing with my solver.
After basics:
| Code: | r4c8<>7, <16> Remote Pair, r3c8=7 => r67c8<>7
+--------------------------------------------------------------+
| 1567 1567 3 | 4 16 2 | 567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 1567 4 8 | 69 3 19 | 2 167 1567 |
|--------------------+--------------------+--------------------|
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 1567 1567 9 | 67 4 3 | 8 126-7 1267 |
|--------------------+--------------------+--------------------|
| 8 26 5 | 1 9 4 | 3 26-7 267 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 55 eliminations remain
|
| Code: | r34c8 Skyscraper <> 7 r1c2,r6c1
+--------------------------------------------------------------+
| 1567 156-7 3 | 4 16 2 | 567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| *1567 4 8 | 69 3 19 | 2 *167 156 |
|--------------------+--------------------+--------------------|
| 3 *167 16 | 2 8 5 | 9 *167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 156-7 1567 9 | 67 4 3 | 8 126 126 |
|--------------------+--------------------+--------------------|
| 8 26 5 | 1 9 4 | 3 26 7 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 49 eliminations remain
|
| Code: | Kraken X-Wing c57 \r15 <> 6 r1c2 w/remote cell r8c7 -or-
finned Franken Swordfish c57b7\r158 <> 6 r1c2 w/fin cell r7c2
+--------------------------------------------------------------+
| 1567 15-6 3 | 4 *16 2 | *567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 1567 4 8 | 69 3 19 | 2 167 156 |
|--------------------+--------------------+--------------------|
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 *16 19 | *67 5 3 |
| 156 1567 9 | 67 4 3 | 8 126 126 |
|--------------------+--------------------+--------------------|
| 8 #26 5 | 1 9 4 | 3 26 7 |
| *16 *126 4 | 3 7 8 | *56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 47 eliminations remain
|
| Code: | gM-Wing: (6=1)r2c3 - r1c12 = (1-6)r1c5 = (6)r4c4 => r3c1<>6
+--------------------------------------------------------------+
| b1567 b15 3 | 4 c16 2 | 567 8 9 |
| 2 9 a16 | 8 5 7 | 4 3 16 |
| 157-6 4 8 | d69 3 19 | 2 167 156 |
|--------------------+--------------------+--------------------|
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 156 1567 9 | 67 4 3 | 8 126 126 |
|--------------------+--------------------+--------------------|
| 8 26 5 | 1 9 4 | 3 26 7 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 46 eliminations remain
|
| Code: | (6)r1c5 = r5c5 - (6=7)r5c7 - r1c7 = (7)r1c1 => r1c1<>6
+--------------------------------------------------------------+
| 157-6 15 3 | 4 16 2 | 567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 157 4 8 | 69 3 19 | 2 167 156 |
|--------------------+--------------------+--------------------|
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 156 1567 9 | 67 4 3 | 8 126 126 |
|--------------------+--------------------+--------------------|
| 8 26 5 | 1 9 4 | 3 26 7 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 45 eliminations remain
|
|
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Sun Nov 22, 2009 2:38 am Post subject: |
|
|
| keith wrote: | There is a skyscraper that takes out <7> in R1C2. Leading to: | Code: | +----------------+----------------+----------------+
| 1567 15-6 3 | 4 16# 2 | 567# 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 1567 4 8 | 69 3 19 | 2 167 1567 |
+----------------+----------------+----------------+
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16# 19 | 67# 5 3 |
| 156 1567 9 | 67 4 3 | 8 1267 1267 |
+----------------+----------------+----------------+
| 8 26b 5 | 1 9 4 | 3 267 267 |
| 16 126 4 | 3 7 8 | 56a 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+----------------+----------------+----------------+ |
I have no idea what this is called, but: (Is true means is 6.)
Either the X-wing # is true, or a is true. If a is true, b is true, eliminating 6 in R1C2.
Now, I am stuck.
Keith |
I'd call it a "almost x-wing"
[x-wing(6)r15c57] = (6)r8c7 - (6)r8c12 = (6)r7c2 |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Sun Nov 22, 2009 2:43 am Post subject: |
|
|
| Quote: | | Code: | (6)r1c5 = r5c5 - (6=7)r5c7 - r1c7 = (7)r1c1 => r1c1<>6
+--------------------------------------------------------------+
| 157-6 15 3 | 4 16 2 | 567 8 9 |
| 2 9 16 | 8 5 7 | 4 3 16 |
| 157 4 8 | 69 3 19 | 2 167 156 |
|--------------------+--------------------+--------------------|
| 3 167 16 | 2 8 5 | 9 167 4 |
| 4 8 2 | 679 16 19 | 67 5 3 |
| 156 1567 9 | 67 4 3 | 8 126 126 |
|--------------------+--------------------+--------------------|
| 8 26 5 | 1 9 4 | 3 26 7 |
| 16 126 4 | 3 7 8 | 56 9 256 |
| 9 3 7 | 5 2 6 | 1 4 8 |
+--------------------------------------------------------------+
# 45 eliminations remain
|
|
nice.
this is one of those 5 inference chains that isn't named yet.
in this form
A=A-(A=B)-B=B |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Sun Nov 22, 2009 5:49 am Post subject: |
|
|
| storm_norm wrote: | | Quote: | | Code: | (6)r1c5 = r5c5 - (6=7)r5c7 - r1c7 = (7)r1c1 => r1c1<>6
|
|
nice.
this is one of those 5 inference chains that isn't named yet.
in this form
A=A-(A=B)-B=B
|
I run across them periodically but, since they aren't a recognized pattern, I normally ignore them. I believe this is the first that I've posted. Glad you like it.
Regards, Danny |
|
| Back to top |
|
|
Mogulmeister
Joined: 03 May 2007
Posts: 1151
|
| Posted: Sun Nov 22, 2009 11:28 am Post subject: |
|
|
| I believe in some quarters it is referred to as an AIC Type 2. |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Sun Nov 22, 2009 5:30 pm Post subject: |
|
|
| Mogulmeister wrote: | | I believe in some quarters it is referred to as an AIC Type 2. |
hi Mogulmeister,
what I was talking about was the actual technique name.
such as M-wing, w-wing, xy-wing
the number of inferences in these named techniques are identical, they each contain 5, 3 strong inferences and 2 weak inferences.
the reason that some have a specific name is because they are readily found and have been documented as such in this forum and other forums.
anyways,
the example above also has 5 inferences except the individual candidates are arranged differently and therefore constitute another one of these short chains that hasn't been documented as being readily found.
although, you can see that in the pattern
A=A-(A=B)-B=B
there is one bi-value cell available in the middle of the chain to be spotted by the solver. the question then becomes, how much of an effort the solver wishes to take in order to find the rest of the chain.
W-wings, M-wings, xy-wings have been shown to provide easier reasoning.
the example above might not prove to be as easy and therefore hasn't been named. |
|
| Back to top |
|
|
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sun Nov 22, 2009 7:56 pm Post subject: |
|
|
| Quote: | | the example above also has 5 inferences |
I think I disagree, if you mean it has the same number of inferences as a W-wing.
In the following, a, b, ... are any candidates, X, Y are particular candidates. - is a weak link, = is a strong link.
The W-wing is
XY - AY = bY - XY
and the end cells are pincers on X. Looks like three inferences to me.
The current example is 16 = 16 - 67 - 567 = 567
Since you are also using the strong inference 67 in the center cell, it looks like 5 links to me. In the same notation as above
aX = bX - (X=Y) - cY = dY
The "pincers" are also not so obvious: aX and dY need to "see" each other, and then the eliminations are:
a cannot contain Y, and d cannot contain X.
(To understand this statement, work the chain from left to right, starting with aX is not X. Then work it from right to left, with dY is not Y.)
If I am missing something, please correct me, with prejudice if need be!
Keith  |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Sun Nov 22, 2009 8:05 pm Post subject: |
|
|
this is actually 5 inferences. remember that in the bi-value XY cells on the ends, the X and Y have a Strong inference between them.
| Code: | (X=Y) - Y = Y - (Y=X)
| | | | |
1 2 3 4 5 |
|
|
| Back to top |
|
|
keith
Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA
|
| Posted: Sun Nov 22, 2009 8:30 pm Post subject: |
|
|
| storm_norm wrote: |
this is actually 5 inferences. remember that in the bi-value XY cells on the ends, the X and Y have a Strong inference between them.
| Code: | (X=Y) - Y = Y - (Y=X)
| | | | |
1 2 3 4 5 |
|
Norm,
Thank you. My mistake.
Keith |
|
| Back to top |
|
|
Mogulmeister
Joined: 03 May 2007
Posts: 1151
|
| Posted: Mon Nov 23, 2009 5:29 am Post subject: |
|
|
I understand what you mean Norm: a generic, pattern based description of a sudoku elimination strategy.
To misquote the owner of your eponymous handle:
"Guess we'll have to bomb that bridge when we get to it."  |
|
| Back to top |
|
|
Mogulmeister
Joined: 03 May 2007
Posts: 1151
|
| Posted: Mon Nov 23, 2009 7:16 am Post subject: |
|
|
Be gentle with me - I haven't been on this board for 2 years but....
We got to here
Now consider the <16> pair at r24c3. There is an absurdity created if you let r4c3 become 6. Why ?
r2c3 is forced to be 1 which leaves a locked pair <57> at r13c1 thus removing both 1 and 5 from r1c2 - an impossibility.
(Edit - and as Danny pointed out a 6 in r4c3 removes all 6's from block 1!)
So r4c3 is not 6.
After that it is just singles and the puzzle is over.
Last edited by Mogulmeister on Mon Nov 23, 2009 11:10 am; edited 5 times in total |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
