Wednesday, 18th October 2017
Puzzles Solved Yesterday: 130
Home | Register | Login | Current Puzzle | Archives | Leaderboard | Forum | Tutorial | FAQ
Forum Index
 
An unusual deduction (Friday, 26th July 2013)
Darklady
Kwon-Tom Obsessive
Puzzles: 3489
Best Total: 9m 37s
Posted - 2013.07.28 13:44:35
The lower left corner of the puzzle should look like this after some straightforward deductions:



But, you can actually get from that to this with some uncommon ones:



Do you see how?

I'm sure some of you are already familiar with the relevant patterns here, but I thought this area of the puzzle was unusual enough to be worth noting.
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.28 16:00:20
Quote:
Originally Posted by Darklady

I'm sure some of you are already familiar with the relevant patterns here, but I thought this area of the puzzle was unusual enough to be worth noting.

While I can see at least two different sets of "guesses" that I could try that were simple enough to do in my mind and reject, I don't actually have any of them memorized as a pattern.  I probably lose a few seconds having to mentally try them, I suppose.  I'll have to look at it again and expand my memorized corner patterns.
Darklady
Kwon-Tom Obsessive
Puzzles: 3489
Best Total: 9m 37s
Posted - 2013.07.28 19:01:58
The significant patterns here have nothing to do with corners.
arbor8
Kwon-Tom Obsessive
Puzzles: 3074
Best Total: 14m 2s
Posted - 2013.07.28 19:33:08
There is only 1 line coming to the lower left corner by the lower "2" : either under "1" or over "1". If you try the over ,contradiction follows.
   For me, this is certainly not a pattern. I have to try both the
alternatives.Important is though that the lower "2" allows only 1 line
Darklady
Kwon-Tom Obsessive
Puzzles: 3489
Best Total: 9m 37s
Posted - 2013.07.28 19:48:52
Quote:
Originally Posted by arbor8
There is only 1 line coming to the lower left corner by the lower "2" : either under "1" or over "1"... Important is though that the lower "2" allows only 1 line

That's half of it! There's one more uncommon pattern here.
arbor8
Kwon-Tom Obsessive
Puzzles: 3074
Best Total: 14m 2s
Posted - 2013.07.28 19:59:47


combining with the former deduction.
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 00:15:58
I see a few patterns.  This one I knew, but I have to do another one first:



Same for this one:



This one I didn't, but it's a good one, and can be applied off of the original problem, and may get us partially there.



There is another one I'm still trying to figure out how simple I can make it.  I'll work on it later.
Last edited by LoopGuy - 2013.07.29 00:18:06
Jankonyex
Kwon-Tom Obsessive
Puzzles: 3636
Best Total: 9m 35s
Posted - 2013.07.29 01:12:50
The 2 adjacent to the 3 is an odd 2.
Hence the blank between two 3's is a hidden 2.

Last edited by Jankonyex - 2013.07.29 01:20:57
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 01:36:25
Quote:
Originally Posted by Jankonyex
The 2 adjacent to the 3 is an odd 2.
Hence the blank between two 3's is a hidden 2.

That's interesting.  What is your definition of an "odd 2", and when can I apply it?  Also, I see two different solutions to it:



and:



So there needs to be further restrictions to actually come up with a unique solution.  At most, you get this:



Of course, then you can apply another pattern, so that may be as good as it gets, but I was hoping for something a bit simpler then that.
Last edited by LoopGuy - 2013.07.29 01:45:16
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 02:50:49
Ok, here we go again.  After applying my 1-2 solution above, and removing it from the corner as darklady pointed out, I'm left with this:



And I can't seem to leave anything else out in order to arrive at the solution we started with.  Even putting the 2 between the two threes still leaves me with trial and error.  So I'm stuck.  Is there a simpler pattern then this?
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 02:57:56
And I do realize that having a 2 instead of a 1 is equivalent, as arbor8 says, there has to be a line on top or bottom, but not both:



So maybe I'm better off writing it this way:



and explaining that exactly one of the ? lines need to be filled in.
Last edited by LoopGuy - 2013.07.29 03:01:25
Jankonyex
Kwon-Tom Obsessive
Puzzles: 3636
Best Total: 9m 35s
Posted - 2013.07.29 07:40:36
For years, I keep saying this: "sorry for my poor English". But I would like to use by own terminology. So I just explain it here.

Surrounding a point, there can be either 0 line (no line pass through this point) or 2 lines (line in and line out). Having line to pass through is a "valid point". Let those surrounding four lines be a, b, c and d in clockwise, then corner between a and b are "opposite" to corner between c and d, but "adjacent" to corner between b and c. If a+b = 1, then we call the corner between a and b an "odd corner", else it is an "even corner". If a point have only odd corners, it is an "odd point" or "straight through". Surrounding a block, there are four corners. If it is a "2", it is an "odd 2" if all four corners are odd corners, else it is an "even 2". If a+b >= 1, we call it an "positive corner", if a+b <= 1, it is an "negative corner".

Summarizing observation and definitions expands to many deduction methods for logically solving a puzzle. For example, for any two opposite "2", their o/e corner status can propagate through their common corner. And for example, a positive corner implies a valid point, and a valid point implies that +/- corner are in pair for all opposite corners.(e.g. the 131 pattern can be explained by this rule) There are many deduction methods like counting, coloring, highlander, hidden number, maximum loop extension, closing and separation analysis etc., as you can think of, but just few people discuss it. If you are interested, please find those old thread in this forum. (really old) I just remember the "logical thinking" and "puzzles sharing" thread, but there should be more. I recommend you to finish all my hand made puzzles. They are not difficult and contain some inspiration for solving a puzzle.
arbor8
Kwon-Tom Obsessive
Puzzles: 3074
Best Total: 14m 2s
Posted - 2013.07.29 08:29:44
Your english is fine and easy to understand but your puzzles are very
difficult! I solved #156 in 15 min and needed to use highlanders often.Overall rating is difficult but very solvable
   But #123 offers resistance.Have you any guidelines how to deal with
it?
Darklady
Kwon-Tom Obsessive
Puzzles: 3489
Best Total: 9m 37s
Posted - 2013.07.29 08:48:42
Jankonyex, that's an interesting deduction! My method was a bit different. Here's the other pattern I used:


If there are an odd number of lines passing between the two ? marks, then there must be a line here:


and the 2 will be what Jankonyex calls an "odd 2" - the two line segments around it do not touch each other.

Given the layout in Friday's puzzle, this also means that the segment below the 3 must be a line, else a loop would be formed around this 3 and the 3 in the corner below it. From there, everything else in my example should be trivial.

As a side note, here's a similar pattern - you can make a deduction here if you know that there are an odd number of lines between the ? marks, and you can make a different deduction if you know that there are an even number of lines:

LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 12:26:40
Quote:
Originally Posted by arbor8
Your english is fine and easy to understand but your puzzles are very
difficult! I solved #156 in 15 min and needed to use highlanders often.Overall rating is difficult but very solvable
   But #123 offers resistance.Have you any guidelines how to deal with
it?

I remember #123.  I used nested guesses and multiple browser windows / tabs to solve that one, and it still took me awhile to figure out which guesses forced enough of the puzzle to be solved so that I could determine if the combination of guesses I used was right or could be eliminated.
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.29 12:32:47
Quote:
Originally Posted by Jankonyex
If you are interested, please find those old thread in this forum. (really old) I just remember the "logical thinking" and "puzzles sharing" thread, but there should be more. I recommend you to finish all my hand made puzzles. They are not difficult and contain some inspiration for solving a puzzle.

I remember that thread, and being confused by it, but I think I now get what you are trying to say, so now I'll go read them again.  I've finished all you're user puzzles, and I do think I'm a better solver for it, so thanks for the puzzles and advice.
arbor8
Kwon-Tom Obsessive
Puzzles: 3074
Best Total: 14m 2s
Posted - 2013.07.29 12:33:35
Yes, I was thinking too applying both multiple windows plus paper
to keep my tracking in order. But also, there have to be more
efficient deductions than my recent ones
LoopGuy
Kwon-Tom Obsessive
Puzzles: 761
Best Total: 45m 59s
Posted - 2013.07.30 00:26:17
Quote:
Originally Posted by arbor8
Yes, I was thinking too applying both multiple windows plus paper
to keep my tracking in order. But also, there have to be more
efficient deductions than my recent ones

I used alt-printscreen pasting it into a document, but paper may work as well.  I have found the most important indicator of difficulty is how many other people have solved it, and only 24 people have, which is quite hard.

I remember going through a particular hard puzzle where I posted a partial solution where I had become stuck, and various people gave me a few hints on what I had missed.  You may want to try that in a separate thread.  Just warn people in the title in case they want to work on it themselves before reading the thread.  And it took me 1d 22h 50m 58s, for the record.
arbor8
Kwon-Tom Obsessive
Puzzles: 3074
Best Total: 14m 2s
Posted - 2013.07.30 03:57:15
Got finally solved, in 42min 11. No fine insights, just crunching.
My original pivotpoints were the sole 3 in the NW corner and the NW line of 3's and 1's (with 3 1's and 2 3's , 3rd line from NE corner ) which ends with an odd 3.
Jankonyex
Kwon-Tom Obsessive
Puzzles: 3636
Best Total: 9m 35s
Posted - 2013.08.01 19:33:34
Quote:
Originally Posted by arbor8
Your english is fine and easy to understand but your puzzles are very
difficult! I solved #156 in 15 min and needed to use highlanders often.Overall rating is difficult but very solvable
   But #123 offers resistance.Have you any guidelines how to deal with
it?
Don't use any highlanders and brute-forcing for "Jsl" puzzles series. They are handmade puzzles and can be solved logically without any brute-forcing (and highlanders). On the other hand "Jcsl" series are generated puzzles. I make them public after I've solved one, and provided a loopy.exe playback solution for each of them. They are most difficult puzzles I've found and are usually not very interesting and require some brute-forcing and advanced analysis. For #123 (Jcsl52+), unfortunately I havn't make a solution playback on my computer that time, and I've already forgotten it (already 5 years ago), so I can't give you any guidelines.

So, you may better try to solve all "Jsl" series logically (without fix position, and with more analysis). As I remember I've provided all playback solutions for this series. If you don't understand any step in the solution, you can discuss here with everyone. Have fun with my "Jsl" series
Last edited by Jankonyex - 2013.08.01 19:34:48

Forum Index