| View previous topic :: View next topic |
| Author |
Message |
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
 Posted: Wed Feb 25, 2009 1:02 am Post subject: m.b.metcalf's patterns game puzzle |
 |
|
this puzzle by metcalf is rated 7.0 by SE
| Code: |
+-------+-------+-------+
| . . 1 | . 2 . | . 3 . |
| 4 5 . | . . 3 | . 6 . |
| . . 3 | . . . | 5 . 7 |
+-------+-------+-------+
| . 1 . | . 8 . | . . . |
| 5 . . | 2 . 6 | . . 8 |
| . . . | . 9 . | . 5 . |
+-------+-------+-------+
| 6 . 4 | . . . | 2 . . |
| . 9 . | 7 . . | . 8 6 |
| . 8 . | . 6 . | 3 . . |
+-------+-------+-------+
|
Play this puzzle online at the Daily Sudoku site
this next one is by tarek... SE...6.8
| Code: | . . .|1 . 7|. . .
. . .|9 . 5|. . .
. . 3|. 8 .|5 . .
-----+-----+-----
. 9 .|. . .|. 7 .
. . 5|2 . 6|4 . .
1 . .|. . .|. . 8
-----+-----+-----
7 . .|. . .|. . 9
. . 6|. 4 .|3 . .
. 2 .|. . .|. 1 . |
enjoy! |
|
| Back to top |
|
 |
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Wed Feb 25, 2009 1:57 pm Post subject: Re: m.b.metcalf's patterns game puzzle |
|
|
| storm_norm wrote: | this puzzle by metcalf is rated 7.0 by SE
| Code: |
+-------+-------+-------+
| . . 1 | . 2 . | . 3 . |
| 4 5 . | . . 3 | . 6 . |
| . . 3 | . . . | 5 . 7 |
+-------+-------+-------+
| . 1 . | . 8 . | . . . |
| 5 . . | 2 . 6 | . . 8 |
| . . . | . 9 . | . 5 . |
+-------+-------+-------+
| 6 . 4 | . . . | 2 . . |
| . 9 . | 7 . . | . 8 6 |
| . 8 . | . 6 . | 3 . . |
+-------+-------+-------+
|
Play this puzzle online at the Daily Sudoku site
enjoy! |
I guess the word is ENJOY. I have never had such a long puzzle. I had 10 single digit steps and 2 bivalue steps.
Ted  |
|
| Back to top |
|
|
wapati
Joined: 10 Jun 2008
Posts: 472
Location: Brampton, Ontario, Canada.
|
| Posted: Wed Feb 25, 2009 2:45 pm Post subject: Re: m.b.metcalf's patterns game puzzle |
|
|
| tlanglet wrote: | | storm_norm wrote: | | this puzzle by metcalf is rated 7.0 by SE |
I guess the word is ENJOY. I have never had such a long puzzle. I had 10 single digit steps and 2 bivalue steps.
Ted |
I started with 7 steps, one of which was a sashimi sword (yay!).
I pruned it down to a skyscraper, W-wing, x-wing and a finish of skyscraper.
Lots of fun in this one! |
|
| Back to top |
|
|
wapati
Joined: 10 Jun 2008
Posts: 472
Location: Brampton, Ontario, Canada.
|
| Posted: Wed Feb 25, 2009 3:01 pm Post subject: Re: m.b.metcalf's patterns game puzzle |
|
|
| storm_norm wrote: |
this next one is by tarek... SE...6.8
| Code: | . . .|1 . 7|. . .
. . .|9 . 5|. . .
. . 3|. 8 .|5 . .
-----+-----+-----
. 9 .|. . .|. 7 .
. . 5|2 . 6|4 . .
1 . .|. . .|. . 8
-----+-----+-----
7 . .|. . .|. . 9
. . 6|. 4 .|3 . .
. 2 .|. . .|. 1 . |
enjoy! |
2 of the 3 sashimi swordfish that are visible from the start are very useful, and finish off with a W-wing.  |
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Thu Feb 26, 2009 2:03 am Post subject: |
|
|
I worked long and hard to get started on this puzzle. Finally, I found a chain that deleted <7> from r3c2. My attempt at Euraka notation may be faulty, especially the last node, but here it is:
(7=1)r3c9 - (1=3)r5c9 - (3=8)r5c1 - (7)r5c2; r3c2<>7.
After a lot of clean-up, an ER<3>, finned X-wing<4>, x-wing <4> and a finned x-wing<6> completed the job.
Wapati, I did not look for sashimi swordfish but will do so immediately!
Ted |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Thu Feb 26, 2009 3:05 am Post subject: |
|
|
deleted because better info was provided
Last edited by storm_norm on Thu Feb 26, 2009 5:05 pm; edited 1 time in total |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Thu Feb 26, 2009 9:46 am Post subject: |
|
|
| tlanglet wrote: | I worked long and hard to get started on this puzzle. Finally, I found a chain that deleted <7> from r3c2. My attempt at Euraka notation may be faulty, especially the last node, but here it is:
(7=1)r3c9 - (1=3)r5c9 - (3=8)r5c1 - (7)r5c2; r3c2<>7.
|
Nice try. From what I can tell, you actually performed a SIN. I don't know how to do them in Eureka notation, but here's my stab at completing your elimination.
| Code: | ..... - (3 )r5c2 ....
/ \
(7=1)r3c9 - (1=3)r5c9 - (3=8)r5c1 - (8=7)r5c2; => r3c2<>7
|
I would have used the old Implication notation used by Jeff for forcing networks, of which SINs are a subset.
| Code: | [r3c9]<>7 [r3c9]=1 [r5c9]=3 ([r5c2]<>3) [r5c1]=8 [r5c2]=7; => [r3c2]<>7
|
|
|
| Back to top |
|
|
Asellus
Joined: 05 Jun 2007
Posts: 865
Location: Sonoma County, CA, USA
|
| Posted: Thu Feb 26, 2009 10:41 am Post subject: |
|
|
Ted,
There is no need for a branched AIC or SIN. This is a simple AIC: an ALS chain of two bivalue cells (one cell ALS) and a 2-cell {378} ALS in r5c12:
(7)r3c2 - (7=1)r3c9 - (1=3)r5c9 - ALS[(3)r5c12=(7)r5c2] - (7)r3c2; r3c2<>7
All instances of any two digits within any ALS have a strong inference because they cannot both be false (since that "starves" the ALS). In this case, the two <3>s in r5c12 are grouped and these grouped <3>s have a strong inference with the single <7> in r5c2. Since all of the <3>s in the group can "see" the <3> at r5c9, they are weakly linked with it, as required.
You can think of the strong inference within an ALS as a "pseudo-bivalue" if you like, in which case your AIC has the structure of an XY-Wing.
Note: We also have the strong inferences ALS[(8)r5c12=(7)r5c2] and ALS[(8=3)r5c12], in which both the <8>s and <3>s are grouped. However, neither of these is useful because the <8>s in the r5c12 group are conjugate and are thus unable to link to any other <8>s. |
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Thu Feb 26, 2009 2:37 pm Post subject: |
|
|
Thanks for all the help folks! First attempts are always fun in any case, but it seems that this situation had secondary issues as well.
Danny, your two part notation seems to state the intended logic. I guess this is also similar to the approach of a "branched AIC" as noted by Asellus (which I will have to research).
Asellus, I followed your detailed explanations, but I am still not comfortable with ALSs so I never considered that possibility when trying to write the Euraka notation. What would the Euraka notation for a branched AIC look like?
Ted |
|
| Back to top |
|
|
wapati
Joined: 10 Jun 2008
Posts: 472
Location: Brampton, Ontario, Canada.
|
| Posted: Thu Feb 26, 2009 2:45 pm Post subject: |
|
|
| tlanglet wrote: |
Wapati, I did not look for sashimi swordfish but will do so immediately!
Ted |
Here is the one that isn't needed.
The others are of a VERY similar form and orientation.
| Code: | .------------------------.------------------------.------------------------.
| 245689 4568 2489 | 1 236 7 | 2689 234689 2346 |
| 2468 14678 12478 | 9 236 5 | 12678 23468 123467 |
| 269 #17 3 | 46 8 24 | 5 269 #17 |
:------------------------+------------------------+------------------------:
| 2346 9 24 | 3458 15 348-1 | 126 7 12356 |
| 38 378 5 | 2 #179 6 | 4 39 #13 |
| 1 3467 247 | 3457 579 349 | 269 23569 8 |
:------------------------+------------------------+------------------------:
| 7 13458 148 | 3568 256-1 1238 | 268 24568 9 |
| 589 #158 6 | 578 4 #1289 | 3 258 257 |
| 34589 2 489 | 35678 5679 389 | 678 1 4567 |
'------------------------'------------------------'------------------------' |
|
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Thu Feb 26, 2009 5:19 pm Post subject: |
|
|
| tlanglet wrote: | Danny, your two part notation seems to state the intended logic. I guess this is also similar to the approach of a "branched AIC" as noted by Asellus (which I will have to research).
|
Yes, my intent was to lend assistance to your solution and to express an opinion that it wasn't a chain so much as it was a forcing network -- which is OK by me.
[Withdrawn: letting it slide]
I've stayed away from ALS for a couple of reasons. I always had the feeling that it was a forcing network in disguise. However, I could easily be wrong.
Last edited by daj95376 on Fri Feb 27, 2009 3:02 pm; edited 1 time in total |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Thu Feb 26, 2009 7:34 pm Post subject: |
|
|
| wapati wrote: |
Here is the one that isn't needed.
The others are of a VERY similar form and orientation.
|
Thanks Wapati, I found the other two sashime swordfish plus a finned swordfish on <1>.
I am feeling more comfortable with sashimi after this exercise. Hopefully I will be able to spot them readily in the future.
Ted |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Thu Feb 26, 2009 9:24 pm Post subject: |
|
|
[Withdrawn: letting it slide]
Last edited by daj95376 on Fri Feb 27, 2009 3:03 pm; edited 1 time in total |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Fri Feb 27, 2009 6:32 am Post subject: |
|
|
| Quote: | | Bottom Line: Since Sudopedia doesn't discriminate about the starting and ending inference, I should just let the topic slide |
it isn't up to you if the topic slides or not !!! 
that is up to Keith, Marty, and Sam
and, you can call AIC's whatever you want. call them networks, DIC's, x-wings, cheesecakes, ping pong balls, whatever
they are still AIC's if the inferences alternate between weak and strong, regardless if any eliminations can be made. |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Fri Feb 27, 2009 7:34 am Post subject: |
|
|
[Withdrawn: letting it slide]
Last edited by daj95376 on Fri Feb 27, 2009 3:03 pm; edited 1 time in total |
|
| Back to top |
|
|
storm_norm
Joined: 18 Oct 2007
Posts: 1741
|
| Posted: Fri Feb 27, 2009 8:04 am Post subject: |
|
|
Danny,
I get the sense that AIC's have thrown a "wrench in the gears" in your quest to solve sudoku puzzles.
your affinity for forcing chains, I assume, have something to do with this, yes?
is it because forcing chains are easier to write code for? easier to formulate an algorithm? |
|
| Back to top |
|
|
Asellus
Joined: 05 Jun 2007
Posts: 865
Location: Sonoma County, CA, USA
|
| Posted: Fri Feb 27, 2009 11:32 am Post subject: |
|
|
| Ted wrote: | | Asellus, I followed your detailed explanations, but I am still not comfortable with ALSs so I never considered that possibility when trying to write the Euraka notation. What would the Euraka notation for a branched AIC look like? |
I would encourage you to try to get comfortable with ALSs when constructing AICs since they are very simple and powerful.
For instance, using the strong (3)=7 inference in the ALS above is much easier to understand than is the branched AIC (since you asked what it would look like):
| Code: | (3=8)r5c1
/ \
(7=1)r3c9 - (1=3)r5c9 - [(38)=7]r5c2 |
What makes this branched AIC tricky to understand is the grouped <3> and <8> in r5c2, written "(38)". We often see groups formed with single digits (as in the common grouped links, or as in the grouped <3>s in the 378 ALS node of the AIC above). But, multiple digits can also be grouped. A Group is false if all of its members are false and is true if any or all of its members is true.
Reading from left to right is straightforward: The <3> in r5c9 propagates directly to r5c2 as well as to r5c1 where it results in <8> true. Together the <3> in r5c9 and <8> in r5c1 propagate to r5c2 and render the (38) group false and thus <7> true.
Reading from right to left is a bit trickier: <7> false in r5c2 means the (38) group is true. So, one or both of the <3> and <8> must be true. (Since it is a single cell, they cannot both be true. But, the AIC logic works in any case.) If <8> is true, the inference propagates via r5c1 (<3> true) to r5c9. If <3> is true, the inference propagates directly to r5c9. (In a case where it was possible for both digits to be true, the AIC would propagate along both branches.)
Even though this is a fairly simple branch (compared with what it is possible to come up with), seeing the structure in this way, with the inference involving grouped multiple digits, is not so intuitive. It is much easier, in my opinion, to recognize the 378 ALS and exploit the (3)=7 inference directly, without any branching or consideration of Group logic.
The goal with AICs is to use the inferences evident in readily recognizable patterns to construct the AIC, not to do a series of "if ... then"s or "forcings". Learning to see the strong inferences contained in ALSs adds to ones repertoire of readily recognizable patterns.
[Edit to correct spelling.]
Last edited by Asellus on Fri Feb 27, 2009 10:48 pm; edited 1 time in total |
|
| Back to top |
|
|
tlanglet
Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills
|
| Posted: Fri Feb 27, 2009 1:59 pm Post subject: |
|
|
Asellus, once again I greatly appreciate your willingness to take the time and effort to provide detailed and understandable posts. In my case, I was looking at the code more in the "if ... then"s or "forcings" mode rather than looking for the inferences of patterns.
And now, it is time to hit the books. The subject for today is ALSs.
Ted |
|
| Back to top |
|
|
daj95376
Joined: 23 Aug 2008
Posts: 3854
|
| Posted: Fri Feb 27, 2009 3:20 pm Post subject: |
|
|
| storm_norm wrote: | Danny,
I get the sense that AIC's have thrown a "wrench in the gears" in your quest to solve sudoku puzzles.
your affinity for forcing chains, I assume, have something to do with this, yes?
is it because forcing chains are easier to write code for? easier to formulate an algorithm? |
No, I don't have a problem understanding/using AICs. When I learn a new technique, I want to understand how it works. What I found is that I can relate many techniques to how I would have solved it using a forcing chain.
Since my introduction to AICs was through Myth Jellies' definition, I rely on the starting and ending strong inference to match the understanding I have of that technique. When an AIC starts and ends with a weak inference, like Asellus and Sudopedia support, then I have to stop and re-evaluate what I'm reading. I've been able to return to Myth Jellies' definition by treating the starting and ending weak inference as extraneous. I find this annoying and have this bad habit of opening my yap.
I will strive to keep my yap closed in the future!!! |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
