Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Post the puzzle or solving technique that's causing you trouble and someone will help

Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Sat Nov 25, 2017 3:33 pm

How would the experts approach this puzzle? I'm just wondering because the couple of software solvers I tried afterwards had a very different idea from my manual line.

400500038503002076001000200010600000000247000000003020008000600750800309130004002

The grid after basic solving:

Code: Select all
+----------------+------------------+-----------------+
| 4   267 279    | 5    1679   169  | 19    3    8    |
| 5   89  3      | 149  189    2    | 149   7    6    |
| 689 67  1      | 3479 36789  689  | 2     459  45   |
+----------------+------------------+-----------------+
| 289 1   24579  | 6    589    589  | 45789 4589 3    |
| 3   89  59     | 2    4      7    | 589   6    1    |
| 689 467 4579   | 19   1589   3    | 45789 2    457  |
+----------------+------------------+-----------------+
| 29  24  8      | 1379 123579 159  | 6     145  457  |
| 7   5   246    | 8    126    16   | 3     14   9    |
| 1   3   69     | 79   5679   4    | 578   58   2    |
+----------------+------------------+-----------------+


My line below:

Hidden Text: Show
First I must say that I kind of screwed up at the beginning because I'd missed marking one of the strong links on the 8s (noticed it soon after). That's why I didn't see the rather obvious X-Cycle and came up with a bit more complicated start step.

1. (2)r4c3-(2=589)r4c156-(589=4)r4c8-(4)r78c8=(4-7)r7c9=(7)r9c7-(7)r4c7=(7-2)r4c3 => r4c3 <> 2
2. (8): r6c5-r6c1=r3c1-r2c2=r2c5-r6c5 => r6c5 <> 8
3. (8-3)r3c5=(3-2)r7c5=(2-4)r7c2=(4-6)r6c2=(6-8)r6c1=(8)r3c1-(8)r3c5 => r3c5 <> 8
4. (9): r3c45-r3c8=r4c8-r4c6=r13c6-r3c45 => r3c45 <> 9
5. (9)r2c7-(9=8)r2c2-(8)r2c5=(8-9)r3c6=(9)r3c8-(9)r2c7 => r2c7 <> 9
6. (9)r5c3-(9=8)r5c2-(8)r2c2=(8)r2c5-(8)r3c6=(8-5)r4c6=(5)r7c6-(5)r9c5=(5)r9c7-(5)r5c7=(5-9)r5c3 => r5c3 <> 9
7. (9): r1c7-r5c7=r5c2-r2c2=r1c3-r1c7 => r1c7 <> 9
stte

(Basic steps in between omitted.)
Last edited by SpAce on Sun Nov 26, 2017 10:38 pm, edited 1 time in total.
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby Leren » Sun Nov 26, 2017 9:13 am

Managed to get the solution down to 6 not too complex moves.

Hidden Text: Show
1. ALS XY Wing: (2=6) r7c2, r8c3 - (6=4) r8c68 - (4=2) r4c1568 => - 2 r7c1

2. Skyscraper: (8) r61c = r3c1 -r3c6 = (8) r4c6 => - 8 r6c5

3. Grouped Skyscraper: (9) r3c8 = r4c8 - r4c6 = (9) r13c6 => - 9 r3c45

4. ALS XY Wing: (1=9) r18c6 - (9=8) r1c23, r3c12 - (8=1) r138c6 => - 1 r7c6

5. W Wing with transport: (1=9) r6c4 - r6c3 = r1c3 - (9=1) r1c7 - r1c6 = (1) r8c6 => - 1 r7c4

6. W Wing with transport : (9=8) r4c6 - r3c6 = r3c1 - (8=9) r2c2 - r5c2 = (9)r5c7 => - 9 r4c8; stte

(Basic steps in between omitted.)

Hodoku took 9 non-basic steps, so 6 steps for me was not too bad. You claim to have taken 8 steps but step 3, a pointing pair, is basic, so you can remove it and claim 7 steps.

The single candidate elimination stte list for this puzzle was : 4 r2c4, 9 r4c8 & 6 r6c1. Proving any of these eliminations might be tricky.

Leren
Last edited by Leren on Mon Nov 27, 2017 8:29 am, edited 2 times in total.
Leren
 
Posts: 2929
Joined: 03 June 2012

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Sun Nov 26, 2017 11:24 pm

Thanks, Leren! I'll try and study your interesting solution. I still suck at grasping complex wing formations and box-based ALSs, so I'm not holding my breath that I will be spotting and using those in the near future. Lined-up ALSs are pretty easy for me to spot, like the one in my first step (I actually found that chain traversing the other direction which used not one but three ALSs), but irregularly shaped box-based ones are not.

I removed the PP move from my list. I certainly counted it as a basic move, but I'd just written it down for myself as a reminder because it made my earlier plan to use an X-Cycle to solve r9c8 unnecessary (how disappointing to spot a pattern that you don't actually need!).

So I had just one step more than you? I guess I should be proud of that! :) I'm actually happy for another reason. This was more or less the first time I worked a little bit strategically instead of just picking any and all odd eliminations I could find. As far as I can tell, all of my steps contributed to the solution.

I assume you used software (Hodoku?) to find the single candidate elimination list? Is that how those impressive single-move solutions are found in general? You let the software find the candidates to attack and then figure out a proof? Or can someone actually guess/know them otherwise?
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Mon Nov 27, 2017 1:39 am

Leren, I looked at your solution in more detail. Very nice. Looks like your first three steps have the same effects as my 1, 2 and 4 (even though the logic in the first one is completely different). I'm quite impressed with your ALS moves (1 and 4). It'll take a while before I can spot things like that. I see that as an eventual possibility, though, because they're just a bit more complex chaining, and I get chains. In general, I just see chains of various complexities instead of named patterns, because I prefer understanding general ideas to memorizing a zillion special cases. If a pattern is not a chain (like non-chain-shaped fishes), I'm unlikely to see it.

A couple of questions/comments:

Is step 3 really a Skyscraper? I saw it as a grouped X-Chain in my solution, and now I can also see it as a Finned X-Wing (Hodoku seems to make same eliminations via a Finned Swordfish as it has more 9s around at that point than either of us). I can't see a Skyscraper, though. (Then again, I don't usually think in terms of Skyscrapers or Kites anyway -- I just see X-Cycles).

What exactly is a W-Wing *with transport*? I think I understand W-Wings although I don't think in those terms (again, I just see them as AICs).

Btw, you seem to have a little notation error in step 4 ("9-1" should be "9=1", I guess?). (I haven't double-checked my own -- probably mistakes there, too.)
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby Leren » Mon Nov 27, 2017 8:57 am

Fixed typo in Step 5 and added Grouped to Step 3. A Skyscraper can always be seen as a Sashimi Finned X Wing or a short (2 Strong links) X Chain - take your pick on what you call it.

Creation of an stte list is a function of my solver (pjb came up with the idea first). It helps with the daily one move wonder puzzles. Also I can limit lots of moves to those that solve a cell. That helps with the dailies, but it's also good for reducing the number of moves used to complete harder puzzles. Transport is just just an added Weak link-Strong link to any move that otherwise would not make an elimination. W Wings are described pretty well on the Hodoku Site here.

Is my solution "better" because it has fewer non-basic moves ? Not everyone would argue that. Another measure of "goodness" in a solution might be the smallest complexity (number of Strong links) of the most complex move.

Leren
Leren
 
Posts: 2929
Joined: 03 June 2012

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Mon Nov 27, 2017 1:50 pm

Leren wrote:added Grouped to Step 3. A Skyscraper can always be seen as a Sashimi Finned X Wing or a short (2 Strong links) X Chain


That part I understand. I also understand in general that things can be viewed from different perspectives and thus have several logical names. But a Grouped Skyscraper? Never heard of it. This is just my personal and possibly uninformed opinion, but I'm not sure if that is a very helpful term to add(?) to the already bloated sudoku nomenclature :)

I'm not saying I'm necessarily right, but having only recently (this spring) started cramming advanced techniques I may have a fresh perspective on what's pedagogically useful and what's not. Most of the patterns and their variants wouldn't be that difficult to learn if they were named and categorized in consistent hierarchies but they're not. There's not much we can do to the historically established names but I'd prefer not to add to the problem. The many and sometimes illogical names for the same concept (or similar names for totally different concepts) make things look much more complicated than they really are, which may discourage potential enthusiasts and complicates communication (for example, some of the exotic patterns or their variations on this forum aren't easily googlable). I understand the benefits from the experts' point of view, as some things can be communicated more concisely with specific names, but I bet it's a pretty small group that understands all the nuances. For others it's hard to follow certain discussions that contain very specialized jargon.

I can only speak for myself, but for me it's much easier to learn a single generic concept - such as X-Chains (which is already a special case of AICs) and its grouped variant - than several instances of it with each having multiple names depending on the perspective. As far as names go, the X-family is one of the worst already, as it has X-Wings that aren't wings and Turbot Fishes that aren't fishes, as well as special cases like Skyscrapers and 2-String Kites (not to even mention Simple Coloring adding to the confusion sometimes). Adding a Grouped Skyscraper to that list, even if it's logical from a certain perspective (and it is), does more damage than good, as far as I'm concerned. Why not call it a Finned X-Wing which is an established concept, if you want to communicate clearly both the size and the shape of the pattern (all of which the generic "Grouped X-Chain" admittedly does not)? A newbie can easily learn about "Finned X-Wing" by googling, but a search on "Grouped Skyscraper" brings nothing by itself and requires translating Skyscraper into an X-Chain which has a known grouped variant. (In fact -- and unfortunately -- googling "Grouped X-Cycle" would be most helpful even though it's not exactly accurate in this case.)

Creation of an stte list is a function of my solver (pjb came up with the idea first). It helps with the daily one move wonder puzzles. Also I can limit lots of moves to those that solve a cell. That helps with the dailies, but it's also good for reducing the number of moves used to complete harder puzzles.


So it's a bit of cheating? :) Just kidding. I understand the motivation and will probably try that approach sometime. Until now I haven't been too concerned about the style points or efficiency of my solve paths. That may change once I learn more. Also, it might actually be a good learning tool to force oneself to attack certain candidates instead of just picking more or less obvious eliminations without any clear goal. Thanks for the clarification!

Transport is just just an added Weak link-Strong link to any move that otherwise would not make an elimination.


Thanks for the clarification on that, too. That's an example of a variant that at least I couldn't find by googling. Searching this forum would have helped more, though, it seems.

W Wings are described pretty well on the Hodoku Site


Yes, I ran into those some time ago. The W-Wing is actually a perfect example of what I said about illogical pattern names above. It's a very confusing pattern name, even if it's somewhat established. I have two problems with it: the W and the Wing :) The first time I saw a reference to it, I thought it was just a short name for WXYZ-Wing just like Y-Wing is for XY-Wing. Only when I finally googled it I learned it was a very different thing. Also, I don't see how it should be called a wing at all and categorized with XY-Wings and XYZ-Wings on the Hodoku site. (Then again, what exactly is the definition of a "wing"? That's somewhat unclear to me as well. I see them as short ALS (or sometimes perhaps AALS?) chains with a pivot cell, but is that accurate and comprehensive?)

Is my solution "better" because it has fewer non-basic moves ? Not everyone would argue that. Another measure of "goodness" in a solution might be the smallest complexity (number of Strong links) of the most complex move.


Yeah, I guess the relative goodness of a solution depends on the perspective like so many things. The elegance of an effective but complex move may be totally lost on newbies while it's greatly appreciated by experts, and vice versa. Could an optimal solution be something like the shortest path your audience can fully understand (or at least learn to understand with acceptable effort)?
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Tue Nov 28, 2017 8:44 am

Leren, I just want to add that my criticism was in no way directed at you or your choice of pattern names. It was a general observation about the unnecessary complexity of the sudoku pattern namespace. It comes from my own recent (and on-going) learning experience, and is a personal opinion as such.
Last edited by SpAce on Thu Nov 30, 2017 9:22 am, edited 1 time in total.
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Thu Nov 30, 2017 9:21 am

One more thing, Leren. Even though I still don't like the name of the W-Wing, I must give more credit to the pattern itself and appreciate that you made me do so. I haven't given it much thought before because it's just a simple AIC and I generally don't have trouble finding them. I've used W-Wings implicitly before without thinking about them as such. However, now that you made me think about it, I started actively looking for them, and it really is one of the easiest patterns to spot. Less time wasted on looking for simple chains is always a good thing. So thanks for that!
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby Leren » Thu Nov 30, 2017 10:31 am

SpAce Wrote : Then again, what exactly is the definition of a "wing"? That's somewhat unclear to me as well

That's a good question, that I've never seen really answered. As far as I know a "Wing" is a move that a solver can "see" via pattern recognition. Of course there is an underlying theorem to a Wing, but you don't have to repeat the proof to make the elimination, just point out the pattern and name it. So, depending on your pattern recognition skills, one solver's Wing is another solver's AIC.

Leren
Leren
 
Posts: 2929
Joined: 03 June 2012

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Thu Nov 30, 2017 12:09 pm

Leren wrote:So, depending on your pattern recognition skills, one solver's Wing is another solver's AIC.


That's quite true. I guess my pattern recognition skills aren't that hot because I mostly see raw chains.
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby JasonLion » Thu Nov 30, 2017 8:07 pm

I've always thought of a Wing as the simplest of a set of complex techniques. Like an X-Wing is the simplest fish, XY-Wings are very short XY-Chains, etc. Just my personal opinion, there doesn't seem to be a clear definition.
User avatar
JasonLion
2017 Supporter
 
Posts: 624
Joined: 25 October 2007
Location: Silver Spring, MD, USA

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby SpAce » Thu Nov 30, 2017 9:35 pm

JasonLion, that seems pretty logical. I like it. Personally I always thought the X-Wing and the Y-Wing were just names chosen by some Star Wars fan because of the rough shape of the patterns :) I've been suspecting that only later those letters and the "Wing" started to have other meanings (X = single digit chains, Y = XY = bivalue chains, but Wing = ?). Does anyone know about the real history? Might be hard because they're probably among the oldest patterns.
SpAce
 
Posts: 68
Joined: 22 May 2017

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby StrmCkr » Tue Dec 12, 2017 7:31 am

usually a "wing" term was implemented to describe the far off endpoints for a "chain"/ visual pattern that allowed the move to do something ie give it lifting power to accomplish the end goal.

in general there is a central hub connecting the "wing" points together... {even an x-wing has 2 off set extension and 2 central hubs}

the end points where generally called pincer's {as they attack a target} or wings across many many different forums {most of them long defunct, with zero backup}.

when, where and why it came to exists in present usage i cannot say for sure, but it did popup into fashion right around the time "xy - wing" was described using the term pincer/wing as i mentioned above.
Some do, some teach, the rest look it up.
User avatar
StrmCkr
 
Posts: 657
Joined: 05 September 2006

Re: Line check for SudoCue Nightmare (Sun Dec 9, 2007)

Postby eleven » Tue Dec 12, 2017 10:46 pm

Don't know, what the inventors meant with their namings.
Since i like birds, i had bird wings in mind, when i connected pivot and pincers of an xy-wing.
And for a w-wing the pivot of the bird wing was doubled.
Those (today's simple) techniques were discovered at the same time by different people with different approach and therefore became different names.
The really interesting part is to discover them by oneself, and then to spot them, not to categorize them for optimized learning.
eleven
 
Posts: 1580
Joined: 10 February 2008


Return to Help with puzzles and solving techniques