You are viewing motris

 
 
01 April 2011 @ 12:14 am
Friday Puzzle #95 - The Voice of Reason  
While I usually post original puzzles here every Friday, this week I have an announcement that merits a change in my normal formula. In an effort to reach more mobile users, who may not have the time to solve their puzzles on paper, Nextoy has recently announced the "Will Shortz Presents KENKEN on the Road" Audiobook series. This really is a bold new direction in the history of this already classic puzzle and is sure to greatly expand their audience with the poetic beauty of their puzzles.

I'm sure my regular readers will be asking themselves why I am mentioning this here. Don't I have a beef with KENKEN that led to my own book of Calcu-doku called TomTom Puzzles? Well, Nextoy and I have recently buried the hatchet. It seems my outspoken nature introduced them not just to my strong opinions about puzzle design but also to - and I'll quote from their correspondence - "the dulcet tones of a logic master", someone with a "smooth baritone voice with a rich timbre that just says 'solve!'". Once they agreed that on all matters KENKEN I would henceforth be called "The Voice of Reason", our deal was set in stone. I'm going to be the voice behind the KENKEN audiobooks!!!

So I don't have any new puzzles for you this week. I've simply been spending too much time in the recording studio putting the finishing touches on some of the Kenerator's latest devious creations. But I snuck out some tracks that seemed particularly well-themed and interesting from our recent collaboration. Solving an audiobook KENKEN is a huge change from the format you are most used to, so prepare to start your learning curve all over again.


Puzzle 23, 4x4 easier* (MP3 file) - "Bueller ... Bueller ..." - getting the right intonation of the first clue, a clear homage to John Hughes' work, was the most important part of the delivery here.


Puzzle 55, 5x5 mediumer* (MP3 file) - "Squeeze Play" - this was a fun puzzle to share on tape with the baseball season starting. Really brilliant theme.


Puzzle 102, 5x5 harder* (MP3 file) - "Countdown" - this one took several takes, as it was easy to make a mistake while reading off so many common clues.


(Note - for those who are not familiar with KENKEN, it's very important to know the standard format of their grids to solve these puzzles. In a KENKEN puzzle, you enter the numbers 1 to N into each cell of an N x N grid so that no number repeats in any row or column. Also, the grid is divided into cages, each with a numeric value and operation. These clues are always printed in the first cell belonging to that cage when proceeding left to right from top to bottom, which is the same order in which they are being read. Cages with subtraction and division are limited to two cells. Singleton values (1-cell cages) are presented as single numbers without any operations. A grid with solution is shown here.)

*The difficulty scale of audiobook puzzles may be different than the rating if you could actually see the grid.
 
 
( 23 comments — Leave a comment )
nameelectricshadow4 on April 1st, 2011 06:37 am (UTC)
I think I got the first one?

3214
1342
2431
4123
motrismotris on April 1st, 2011 06:38 am (UTC)
Good job!
nameelectricshadow4 on April 1st, 2011 06:40 am (UTC)
If I realized you were serious about the later ones having a theme, I would have waited to comment.
Hooligan: beginner's luckjedusor on April 1st, 2011 08:51 am (UTC)
Hah! Cute.
standupphilosopherstandupcanada on April 1st, 2011 12:18 pm (UTC)
I'll try to post the grids on my site later today for the hard of hearing, but I finally got an offer on this bridge Ii've been trying to unload to take care of first. Thanks Thomas - enjoy the day
Caziquecazique on April 1st, 2011 02:55 pm (UTC)
you are the awesomest person in the world today
(Anonymous) on April 3rd, 2011 11:24 am (UTC)
Brilliant, thanks!

Cheers
Rob
(Anonymous) on April 3rd, 2011 08:06 pm (UTC)
Brilliant puzzles!
For the record, I found the third one easier than the other two - those 1-cell cages helped a lot.
RV
motrismotris on April 3rd, 2011 08:09 pm (UTC)
Yes, the third one is probably the hardest for me to rate. I had a very hard time proving it had just one answer, but not a hard time finding that one answer if you spend the clues in a greedy way. So I guess depending on the way you go about solving it the experience can differ.
lunchboylunchboy on April 5th, 2011 07:45 pm (UTC)
Nice! If only all online April Fool's gags had an actual payoff. "Squeeze Play" was the easiest for me, though the last one fell reasonably quickly once I stopped trying to be super-rigorous and just made educated guesses along the lines of "Well, these clues seem like they ought to go here and here."
motrismotris on April 5th, 2011 08:38 pm (UTC)
Yeah, as I already responded to RV, the "being super-rigourous" trap is why I rated the last one as the hardest. Each requires a different kind of insight, and I think mostly spends the kinds of insights that will exist with diagramless KenKen, so I'm pretty happy with the set even though the ratings may be way off.
mathgrant.blogspot.com on April 6th, 2011 05:16 pm (UTC)
Words cannot describe how beautiful your voice is. I could fall asleep and have the happiest dreams ever while listening to your voice. I'm not gifted enough to solve these puzzles without paper, though, which feels like an insult to your abilities.

Puzzle 102 was tricky!
(Anonymous) on April 8th, 2011 11:01 am (UTC)
Uniqueness
Unless I missed something, you didn't use the trick that a potential board is illegal if it has multiple solutions.

All the wrong clue configurations had no solutions at all.
motrismotris on April 8th, 2011 03:09 pm (UTC)
Re: Uniqueness
You're not missing anything. As a constructor I avoided having the solver decide this is "invalid if it has multiple solutions" by never letting that option arise.

I personally think valid diagramless puzzles in this style should never use such "logic" to rule out choices. As I see it, if a grid has multiple solutions, then the puzzle has multiple solutions, and there is no ruling out those grids for a solver, although some subset of solvers may think along your lines and ignore such invalid configurations.
(Anonymous) on April 9th, 2011 08:25 pm (UTC)
Re: Uniqueness
I don't agree. I agree that using the solver of normal puzzles shouldn't use information of this type. (At least in pure solving. Speed solving is a different story.) However, I don't believe the space of "valid" puzzles includes grids with multiple solutions. Solving a grid includes proving the solution is fair and unique. Finding a valid grid includes proving all others grids are not valid for any reason.

I suppose the difference between our views is that you viewed the audio puzzles as a "regular" KENKEN puzzle with fuzzy clues, in which case any clue placement with a solution is valid. On the other hand, I understood the puzzles as an "audiobook form" of KENKEN, where someone reads the puzzle, "accidentally" losing information in the process. (See something similar in http://rec-puzzles.org/index.php/Children.) The puzzle is then both to recover the correct grid, which you know can't have multiple solutions, and solve the grid, without using that same knowledge.
motrismotris on April 10th, 2011 06:17 am (UTC)
Re: Uniqueness
I guess I have two points. First, I generally agree on the possibility of what you describe, but it was not my goal here and would not be my normal starting point. If I did allow this choice, then I would certainly edit the rules to specifically state what is a sometimes unwritten rule that "the correct solution will be the single arrangement of givens that itself has just one valid solution." If this is clear at the outset, then I'd be more comfortable using it as the step. What I don't want is a solver to approach the puzzle, think they are done, not have my answer, and think it's because the puzzle is broken when I'd be using the broken-ness of their solution in a possibly circular fashion to say they are wrong, because I haven't stated an often unstated goal of puzzle design.

Outside of the theoretical question, I think it is harder to construct "satisfying" puzzles here that have extra non-unique solution states. There are obvious clue constellations that could offer this possibility (two similar 3+ cages in rows/columns that give a 12/21 pair of solutions, for example) but it seems challenging to incorporate them from my three puzzles worth of experience here. I'd imagine a computer search would turn up countless examples, but I don't think their solutions would resolve to the clean steps I eventually embedded here. I can write reasonable solutions for all of these to first resolve the grid and then (with very little difficulty) solve it. I'm not sure if I could do the same with ambiguous grids with a single arrangement with a "valid" single solution.
(Anonymous) on April 10th, 2011 02:04 pm (UTC)
Re: Uniqueness
I am reposting since I think the previous post was filtered because of a link.

I don't agree. I agree that using the solver of normal puzzles shouldn't use information of this type. (At least in pure solving. Speed solving is a different story.) However, I don't believe the space of "valid" puzzles includes grids with multiple solutions. Solving a grid includes proving the solution is fair and unique. Finding a valid grid includes proving all others grids are not valid for any reason.

I suppose the difference between our views is that you viewed the audio puzzles as a "regular" KENKEN puzzle with fuzzy clues, in which case any clue placement with a solution is valid. On the other hand, I understood the puzzles as an "audiobook form" of KENKEN, where someone reads the puzzle, "accidentally" losing information in the process. (See something similar in the rec.puzzles archive on the page index.php/Children.) The puzzle is then both to recover the correct grid, which you know can't have multiple solutions, and solve the grid, without using that same knowledge.
motrismotris on April 10th, 2011 03:17 pm (UTC)
Re: Uniqueness
LJ did initially screen your comment but I thought I'd opened it up before I responded to it above. Tell me if you don't see this earlier comment or my reply (otherwise I'll probably clean the thread of the repost).
(Anonymous) on April 10th, 2011 07:09 pm (UTC)
Re: Uniqueness
I see it. I must have opened the post link before you answered, and posted after. I had some trouble connecting to livejournal today, so it was in that state for a while.
motrismotris on April 10th, 2011 07:56 pm (UTC)
Re: Uniqueness
No problem. LJ has had some connectivity issues recently, and they are also now more aggressive at pre-blocking any message especially from an anonymous poster with a web link, so I'm not surprised this issue came up.
(Anonymous) on April 9th, 2011 08:16 pm (UTC)
Further to previous facebook comment, these are lovely puzzles! Thanks!

I found the second one hardest - just because the initial logic was non-immediately obvious. Was there an American sports clue I missed?

Ronald
motrismotris on April 10th, 2011 06:03 am (UTC)
Not really. While squeeze play is a baseball reference, the use of "squeeze" was being used in a non-baseball way to clue the need to fit all those clues in the grid which forces a 6-cell 9+ clue that is squeezed into multiple rows and columns. Using checkerboard logic with domino tiling then forces the single option where "3" is positioned on a corner and not one in from a corner that would cause a problem.

Admittedly "squeeze play" isn't the best clue if my goal was to ease the necessary discovery, but I wasn't going for tremendously informative titles.
(Anonymous) on April 10th, 2011 09:12 pm (UTC)
Yes, exactly how I did it. Very nice :)
( 23 comments — Leave a comment )