New York Times, Wednesday, October 8, 2014

MARY LOU: I enjoyed working with and learning from Jeff. He was very kind and patient mentoring me during the long evolution of this puzzle. He designed a grid that highlighted the five theme answers, and we tried to 1.) minimize the unsavory short fill and 2.) make sure no long fill would detract from or be confused with the themers. Thanks to Will Shortz for accepting and editing the puzzle.

I hope you enjoyed solving the puzzle. It was my first acceptance by the NYT, so I was quite elated to receive that "Yes!" email.

Working with Mary Lou is a pleasure. She exemplifies one quality I think all good constructors must have: the ability to generate and sort through a ton of ideas. I find it takes maybe 10-20 ideas to discover one worthy of publication. Many times a person will give up after two or three theme ideas, but not her — I admire her determination and drive. When people talk about hard work leading to success, that's ML they're talking about.

As for this puzzle, I honestly was only lukewarm on the general idea at first. But when ML found five strong themers that worked, I thought twice. I'm a sucker for Muhammad Ali — "When We Were Kings," a documentary about the Rumble in the Jungle, cracked my vaunted Tier 2 list of favorite movies. (I only have 28 movies listed in my Top Tier; the most recent addition was "The Fighter.") The final hook was when she pointed out the challenge of executing this grid in a way that was both smooth and fun for solvers. I'm also a sucker for a challenge.

Five 15-letter themers is so rough. Four of them is hard enough because there's so many areas that get constrained, and five adds a real kick in the pants (not the good kind). There are so few places that aren't affected by two (or more) themers. The NE and SW were especially painful — there are five parallel answers from STATS to LEE which run through the same two themers; usually a situation you want to avoid like the plague. I don't much care for A TUNE and the visual of those lone cheater squares in the corners, but all the alternatives seemed worse. Grid-making is often an exercise in iteration, trying to sort through hundreds of possibilities in an attempt to figure out what would be most interesting / least grump-inducing for solvers.

It took us a lot of iterating, plus helpful feedback from Will, in order to get it to a place where we felt like it would be a fun experience for solvers. I wish we could have worked in some long fill besides ASHANTI, OH THAT? and IN ON IT, but every arrangement incorporating 8+ letter entries caused too much stress on the grid. As always, construction is rife with trade-offs.