Xan Vongsathorn author page

This puzzle has a long history. I originally submitted it in 2011, and it was accepted shortly after. It got lost in the stack, and a few years passed. It almost ran in late 2014, but at the last minute, a test solver realized that Andrew Reynolds had written a similar puzzle which ran on April 1, 2014. Rather than running two similar puzzles in the same year, mine was pushed back to now.

Of course, I was less than happy about this chain of events at the time. But 2018-Xan is a whole different person than 2011-Xan, and 2018-Xan is thrilled he gets to enjoy the publication of a crossword puzzle he didn't even have to write. Plus, we get a nice opportunity to compare two independent versions of the same idea. Andrew's puzzle is fantastic.

For me, coming up with a crossword theme is a systematic exercise. My basic goal is to surprise people in some way, especially seasoned solvers accustomed to seeing the same tropes over and over. One approach is to consider an assumption that solvers make, enumerate a few ways that assumption might be violated, and then see if you can think of a thematic justification for each violation.

In this case, I was brainstorming ways to sneak an extra layer of meaning into the idea of a circled letter. (Another puzzle from the same brainstorming session can be found here). For no particular reason, circles are commonly used to draw attention to letters in crosswords. So common that we normally don't think about the circle itself as carrying any meaning; it could just as well be a diamond, or gray shading. But here, the circles form the edges of coins, and what do coins do? They flip. It was a nice idea that also gave me the opportunity to push harder on the boundary of double-cluing.

Seven years ago, after this somehow possible puzzle, I had the bold feeling that if you gave me two words, and I thought long and hard enough, more often than not I could come up with a single clue that worked for both of them. Many theme answers in the present puzzle make this task deliberately hard. I was motivated by the idea that it is more surprising and satisfying to discover two answers that differ by only one letter but have hugely different meanings, covered by a single clue parsed in two very different ways.

Overall, I would say the resulting double clues are "better than you'd expect, but not quite good enough." With the benefit of seven years of hindsight, I would dial the ambition of the theme answers back a bit!

Because the anagramming makes this more challenging than a typical rebus puzzle, I decided to help solvers out by placing the rebus squares symmetrically. That turned out to be pretty constraining for both the theme answers and the resulting grid.

I spend a lot of time on grid design, and it's worth noting that this grid is definitely an odd one, even for a rebus puzzle. I tried some of the more obvious designs and ended up rejecting them on the basis of fill, before settling on those odd crosses in the center. For instance, the most obvious tweak to the current design is to replace that left cross ...

. * .
* * *
. * .

with this instead:

. * .
* . .
. . *

(where * = black square, . = white square)

This is a much more efficient use of black squares which increases interlock substantially without altering the layout of the words. But it came at the cost of boring or lousy fill and I wasn't willing to pay the high price. That's a judgment call. Better interlock makes a puzzle more fun to solve, and so does better fill.