Difficulty: Harder 9x9 (More like a Thursday in difficulty, despite the 9x9 size)

So, the (first) week of "Thomas Snyder Outdoes the NYT KenKen Puzzle" has come to a conclusion. You can believe this was a very well planned event that became necessary after the NYT announced it was starting a new puzzle, with a release of a set of puzzles starting from the April Fool's-like joke to today's $ puzzle (remembering my specific choice of $ as operator from Thursday). You can also believe this is an idea I came up with, and further developed, on a day-by-day basis by the seat of my pants. What's important to realize is that we've survived. (Exhales). What have we learned:

Me:
1. KenKen (with rule adjustments to appease my inelegance sensors!) can be surprisingly deep while easily constructed. The difficulty of making a specific theme in the numbers like "3" for 3/3/3*3 or a sequence as in Downward Spiral is very low because they are so easily constructed.

2. KenKen always felt "over-constrained" with lots of tiny boxes, in part because you have way too many clues. Without leaving unclued regions, this can be addressed and in an interesting way with larger region shapes which themselves can complement a visual theme at low cost. I did not write a single puzzle this week without a region larger than 10 cells (with 33 cells being my max), yet each puzzle was still sufficiently constrained to be solvable. By getting closer to a "one path through it" puzzle type, the puzzle is necessarily improved in challenge and likely fun. Dell never did remove those ~10 extra digits in Number Place or put in symmetry and look how that worked for them!

3. By allowing digit repetition, larger sets of sums/products/subtractions/divisions are available than in similar puzzles like Kakuro or Killer Sudoku which offers a new challenge to those both new to math puzzles like KenKen and those who have been doing Cross-Sums since approximately birth. Remember: 7+ = 1+2+4 normally, but with a bendy region, 1+1+5, 2+2+3, and 3+3+1 as well.

4. Commercial KenKen feels most natural at 5x5 or 6x6 in size. You can go larger, and appeal more to the "puzzle audience" of my blog/me, but I think the set of digits and operations in 1 to 5 and 1 to 6 is the right size with sufficient complexity for the majority of the puzzles. Better puzzle construction in this space should be highly commercial, with hard books venturing to 7 or 8 or even (gasp) 9.

You (I hope):

1. KenKen can be a good puzzle, but maybe not in the hands of those who control KenKen(TM).

2. Logic puzzles should be themed and written by human constructors. Unless you think all logic puzzles should be like the word search and lack character. Yes, some puzzles are diversions, but I like my diversions to be more artful than not.

3. Thomas Snyder is a very mean person - not only did he pick on Rex Parker, but he also introduced us to a fun "KenKen" puzzle for a week that we can't find anywhere else now, even if we wanted to continue solving them next week. I can't say I'm done with these - I have a couple "novelties" for another day or so - but I can't say this is going to be a daily thing for much longer.

Closing thoughts:
I answered a lot of my own questions through this exercise, and at least now have a section of my blog I can point to when people want the opinion of the 1st Greatest Solver of Sudoku Puzzles in the World! on this new thing called KenKen. I cannot really publish KenKen, as I'm not tied to the company with the rights, and fighting for space on the bookshelves with the Will Shortz name is very challenging. I would probably need a magic label like Mensa as in "Mensa Presents KenCubed: The Next Dimension of Puzzles" to even bother trying. Maybe there is room at Sterling, where this mark of "quality" exists, but I have other book promises to keep, and experiments to run before I sleep, and experiments to run before I sleep.