Monday, 23rd April 2018
Puzzles Solved Yesterday: 91
Home | Register | Login | Current Puzzle | Archives | Leaderboard | Forum | Tutorial | FAQ
Forum Index
 
Sunday's Kwon-tom Wrong?
itchfizzix
Kwon-Tom Obsessive
Puzzles: 3034
Best Total: 17m 57s
Posted - 2006.07.09 04:32:07
So... I finally finished the Kwon-tom loop for Sunday... but there's no link for the Kwon-tom wrong.  Is it just me?  or others too?
Erick
Kwon-Tom Obsessive
Puzzles: 1292
Best Total: 31m 39s
Posted - 2006.07.09 08:49:49
Yup, it's missing for me too.
foilman
Kwon-Tom Admin
Puzzles: 1720
Best Total: 24m 8s
Posted - 2006.07.09 09:57:43
Very strange... I just checked, and 46 puzzles in the coming weeks are missing wrongs! I'll see if I can add them now...
tilps
Kwon-Tom Obsessive
Puzzles: 4324
Best Total: 20m 22s
Posted - 2006.07.09 11:04:36
Before I go and write a program to find out the answer, does anyone know of a puzzle for which there are no wrongs, no ways in which you can change a single number and still have a single solution.
foilman
Kwon-Tom Admin
Puzzles: 1720
Best Total: 24m 8s
Posted - 2006.07.09 11:10:25
Ok, the wrong is there now...

As to whether there's such a thing as a puzzle that can have no wrong... well I haven't found one yet. But I'm sure such a thing must exist - try a few small puzzles and I'm sure it'll be obvious.
matth314
Kwon-Tom Obsessive
Puzzles: 2516
Best Total: 16m 52s
Posted - 2006.07.09 13:51:02
So, I guess this means that those of us who already completed it have to redo the whole puzzle. It's no big deal, just a bit time-consuming.
Fgnn
Kwon-Tom Obsessive
Puzzles: 717
Best Total: 19m 46s
Posted - 2006.07.09 20:09:23


All Potential Wrongs

Corner:




Edge:




Center:




This may be a good enough example, small, but it can exist.
tilps
Kwon-Tom Obsessive
Puzzles: 4324
Best Total: 20m 22s
Posted - 2006.07.10 00:37:56
Hmm, maybe a stronger condition - a puzzle for which you can't remove any  clues and there exists no wrong. You can take the 0 out of that puzzle and half the one's - and then I think it has wrongs.

Having unneeded clues would make seem to make it harder to make wrongs.
m2e
Kwon-Tom Obsessive
Puzzles: 607
Best Total: 16m 43s
Posted - 2006.07.10 01:32:40
Quote:
Originally Posted by tilps
Hmm, maybe a stronger condition - a puzzle for which you can't remove any  clues and there exists no wrong. You can take the 0 out of that puzzle and half the one's - and then I think it has wrongs.

Having unneeded clues would make seem to make it harder to make wrongs.

Doesn't every puzzle have unneeded clues?
procrastinator
Kwon-Tom Obsessive
Puzzles: 1083
Best Total: 12m 56s
Posted - 2006.07.10 08:55:22
Quote:
Originally Posted by m2e
Quote:
Originally Posted by tilps

Having unneeded clues would make seem to make it harder to make wrongs.

Doesn't every puzzle have unneeded clues?

Every loop has unnecessary clues, but a puzzle doesn't have to include all the numbers, so you can always keep removing them until there are no more unnecessary ones. Foilman's puzzles often contain unnecessary numbers to satisfy his symmetry constraints. On the other hand, I strongly suspect Jankonyex's "extreme" puzzle (which I still haven't quite solved - great puzzle) would be impossible (i.e. non-unique) without any one of the clues. In Fgnn's puzzle, every clue is unnecessary,  a condition that alway leads to wronglessness. I suspect the same is true of the 4x4 practise puzzle in the archives, but haven't checked it thoroughly.


My thoughts on wrongless puzzles: Each possible change needs to create a number of solutions other than one - either zero solutions or multiple solutions. A zero case means the rest of the board didn't depend on that clue, so I postulate that:

1) A lot of zeros means a puzzle is solvable using fairly local reasoning.
2) Multiple solutions means distant interactions.
3) A puzzle is unlikely to have a significant mix of both.
4) Mostly local ~= very easy, mostly distant ~= very hard.
5) Foilman's puzzles are neither very easy nor very hard.
6) There will always be a Kwontom Wrong.
7) There might not always be a Fgnn Wrong or a Jankonyextreme Wrong.
chairman
Kwon-Tom Obsessive
Puzzles: 1395
Best Total: 17m 32s
Posted - 2006.07.10 09:07:37
Let us call a puzzle without unnecessary numbers minimal. All
minimal 3x3 puzzles with the same solution as fgnn's (the border
line) have wrongs. If you take the April first fool puzzle and omit
the zeros, you get a minimal wrongless puzzle, but I guess this
example is not quite satisfactory...
chairman
Kwon-Tom Obsessive
Puzzles: 1395
Best Total: 17m 32s
Posted - 2006.07.10 09:42:33
This seems to be a minimal wrongless puzzle:

m2e
Kwon-Tom Obsessive
Puzzles: 607
Best Total: 16m 43s
Posted - 2006.07.10 10:16:54
Quote:
Originally Posted by chairman
This seems to be a minimal wrongless puzzle:

That not a valid puzzle
Last edited by m2e - 2006.07.10 10:17:53
m2e
Kwon-Tom Obsessive
Puzzles: 607
Best Total: 16m 43s
Posted - 2006.07.10 10:23:58
Quote:
Originally Posted by procrastinator
Quote:
Originally Posted by m2e
Quote:
Originally Posted by tilps

Having unneeded clues would make seem to make it harder to make wrongs.

Doesn't every puzzle have unneeded clues?

Every loop has unnecessary clues, but a puzzle doesn't have to include all the numbers, so you can always keep removing them until there are no more unnecessary ones. Foilman's puzzles often contain unnecessary numbers to satisfy his symmetry constraints. {...}
If Foilman's puzzles are made by this process (which i am pretty sure is the case):
1) draw a loop
2) add a number onto the grid and then see if the puzzle is solveable
3) if not solveable, repeat step 2
4) then grade the puzzle's difficulty
Then isn't it fair to assume that unnecesary numbers will be added? And that puzzles with more unnecesary numbers will be given lower difficulty levels (as a trend, not a rule)?
chairman
Kwon-Tom Obsessive
Puzzles: 1395
Best Total: 17m 32s
Posted - 2006.07.10 10:29:25
Quote:
Originally Posted by m2e
Quote:
Originally Posted by chairman
This seems to be a minimal wrongless puzzle:

That not a valid puzzle

Oops!

I trusted too much on the |3|3| pattern! I should have remembered  this drnull's post.
chairman
Kwon-Tom Obsessive
Puzzles: 1395
Best Total: 17m 32s
Posted - 2006.07.10 10:50:14
Here is another try. This seems to be a minimal wrongless puzzle:

tilps
Kwon-Tom Obsessive
Puzzles: 4324
Best Total: 20m 22s
Posted - 2006.07.10 11:26:13
On puzzle generation.
I actually go the opposite way with my program.  I put All the numbers on, and then start taking them off in random order until all the clues which are left can't be removed and it still be solved with the selected difficulty level.  I'm pretty sure I did that because I thought it would be faster.  Don't remember my logic why though.
foilman
Kwon-Tom Admin
Puzzles: 1720
Best Total: 24m 8s
Posted - 2006.07.10 14:25:49
Quote:
Originally Posted by m2e
If Foilman's puzzles are made by this process (which i am pretty sure is the case):
1) draw a loop
2) add a number onto the grid and then see if the puzzle is solveable
3) if not solveable, repeat step 2
4) then grade the puzzle's difficulty
Good guess! And not far wrong...

This is the incredibly long-winded way my puzzles are designed:-
1) Draw a loop
2) Add a set of numbers (for symmetry) onto the grid and then see how solvable the puzzle is. During this stage I try every possible combination and pick the one that makes it "most solvable".
3) If not completely solvable, repeat step 2.
4) Start taking unnecessary number sets (again, for symmetry) off the board, unless that would make it "too difficult" based on my current difficulty settings.
5) Repeat step 4 until none can be removed.
Last edited by foilman - 2006.07.10 14:28:07

Forum Index