New York Times, Friday, December 1, 2017

This is my sophomore effort in the Times--whether it's sophomoric, I'll leave to the solvers.

I was thrilled to kick this puzzle off with the entry GENDER ROLES. I also loved the first clue I thought of: "Synthetic divisions?" Unfortunately, I quickly realized what you've already realized: a pun based on an obscure algebra algorithm is not going to land. I found it incredibly difficult to decide on a replacement. It was important to me to make my politics clear: traditional gender roles are artificial, divisive and destructive. Simultaneously, I wanted something clever and light-hearted for such a central entry in the puzzle. All credit for the published clue goes to my friend Gabe Chandler, who found a balance by referencing classic sitcoms: familiar and fun, but also based on tired stereotypes that have aged so poorly. I often ask Gabe to "beta-test" my crossword submissions and he's made many helpful suggestions. Thank you!

I enjoyed reading Kevin Der's Constructor's Note last week, where he classified some ways to make a clue end-of-week difficult: misdirection, vagueness, open-endedness, and trivia. I confess to a weakness for trivia. Personally, I love it when a recognizable answer has an impossible clue, and needs to be solved from the crosses. For instance, I wanted to clue ORION'S BELT as "Home of Alnitak, Alnilam and Mintaka," the names of its three stars. Of course, I didn't expect solvers to know these names--I didn't, either, until I looked them up--but I was hoping for bewilderment, followed by an epiphany after getting something like _ R I _ _ S B _ L T. It might look like "Jeopardy!" at first, but it's really "Wheel of Fortune." The same impulse led me to clue PEACEMAKER as "Nickname of the B-36 bomber, ironically," knowing full well that the nicknames of aircraft are typically known only to those who fly them. I was glad to see this one make it in.

As for misdirection, I was hoping to trick some people with a four-letter answer for "One in a story with an apple," especially since the three-letter answer was the obvious one. But since my first puzzle was heavy on misdirection, I tried not to overuse it here.

I'm glad I could sneak in the math term LEMMAS, with the help of easy crosses. This seems to be the first time that the clue for the plural matches modern usage. Logically, there's no difference between a lemma and a theorem: they're both words for proven mathematical results. But if you think your new result is interesting or important, you call it a theorem; if it's a result you feel obligated to prove, in order to get to the interesting or important stuff, you call it a lemma. Somewhat complicating this distinction, mathematicians are often wrong about what other people will find interesting or important (example: this paragraph). Years later, their so-called lemmas may be pillars of "mathematical canon," while their theorems lie forgotten.

I'm a huge fan of this "stair stack" arrangement, featuring three long headline entries in the middle. Stair stacks live and die on three regions. Listed from most important to least, it's critical to get zinginess from and cleanliness around:

1.) the middle stair stack

2.) the long entries in the NW / SE

3.) the medium to longish entries in the SW / NE

Oddly enough, the difficulty level doesn't correspond to this at all! Life is funny like that.

1.) The middle stair stack is of moderate difficulty, sort of the baby bear of the grid. Seems like it should be the hardest, but this is usually where constructors start, the blank slate giving maximum flexibility.

I thought Lily shined here. CONTROL FREAK is dynamite, GENDER ROLES is nice, with its clue making fun of classic sitcoms ("Three's Company," anyone?). And ZILLIONAIRES is awesome, throwing me off when I put in BILLIONAIRES.

2.) The long entries in the NW / SE, those are the papa bear, too easy to make sing. Even after fixing the middle into place, there's usually enough flexibility remaining that these sections aren't so bad. Lily even sections them off a bit with a black square in between ADS and LANDO, to make them even easier.

Here, I thought there was room for improvement. Although ORIONS BELT, PEACEMAKER, and ERROR PRONE are all fantastic, when you only leave yourself four long slots like this, it's imperative that all four sing. When you factor in the Maleskan SERE plus SSE (and SMS) in the SE, it's not a great corner overall. I would have liked another attempt there, maybe even with the black square between LEMMAS / SSE removed.

3.) The medium to longish entries in the SW / NE, those are the mama bear, too tough to fill with much color usually. It's so hard to convert 7-letter slots into great material. My expectations are never high for these regions, so I was pleasantly surprised by ZIP IT UP! (although, is "zip it!" more common?), WANNABES, even BAFFLE.

But overall, solid work and Lily shined in that critical central featured region. With more practice and ITERATEion, I see future POW!s for her.