Update, 4/14/15 @ 4:30 pm ET: If you’re clicking over from one of Rex Parker’s links and visiting Devil Cross for the first time…..welcome! I normally write puzzles instead of blog posts. You can access them by clicking on the links in the Archives sidebar, or in reverse chronological order by clicking on the links at the bottom of each post.

***NOTE: This post contains spoilers about the New York Times puzzle of Tuesday, April 14.***

Today’s New York Times puzzle by Bruce Haight sets an unusual crossword record of sorts: fewest number of different letters in a single New York Times grid (EIGHT, to be exact). The only letters that appear in the puzzle are A, E, G, H, I, R, S, and T. It’s an enormously difficult constraint to place on a puzzle given the limited number of available words that can fit together with only those eight letters, and as Wordplay blogger Deb Amlen notes, the theme produced “significant sacrifices in the fill.” Yet in the same post, Deb continues:

“Does that mean that the puzzle should not have seen the light of day, as some people may feel? I don’t know. Here’s the thing about creativity: In order to come up with new ideas, people have to be given the room to push the envelope, no matter where it takes them. Shutting someone down just because they tried something that forces a bunch of celebrity name plurals and words like RAREE is not good for the long game of crossword puzzle construction, in my opinion.”

She acknowledges that it’s critical to maintain high standards, but then asks:

“… is it possible that those who criticize every word in every puzzle are doing so because they have just fallen into a habit? And what does that do for constructors who might have a new idea for a puzzle, but are afraid to try it for fear of being criticized just because the idea is still embryonic?”

While I am not a daily crossword blogger like Deb or Rex Parker or Amy Reynaldo or Jeff Chen — and I believe it’s fair to assume that Deb is referring primarily to them when she talks about people who “criticize every word in every puzzle” — I am nonetheless someone who is just as keenly interested in the creativity and artistry of puzzles. Though I respect Deb a lot as one of those bloggers and appreciate very much that she has helped spread the word about some of my own puzzle projects, I felt compelled to respond to her points in greater detail.

First, I don’t think anyone would disagree that creativity means allowing people to think outside of the box and push the boundaries from time to time. New ideas require trying new things. But the question isn’t simply whether “pushing the envelope” (as vague a concept in puzzles as in all art) should be allowed; it’s about what purpose it serves. What’s the payoff, basically? Most importantly, is the final product enjoyable for the solver?

That last question is inherently difficult to answer since it’s impossible to account for everyone’s differences in taste, but that doesn’t mean one can’t identify problems in a puzzle’s execution. In Bruce’s puzzle, the gimmick asks solvers to see that it is possible to fill a grid with only eight different letters of the alphabet. For many solvers, that may be enough — it’s no easy feat, after all. At the same time, it’s not unreasonable to critique the fact that the same grid required many, many compromises to pull it off, with answers like HIES, TASTER, SIEG, SIGHER, SETHS, AGRI-, RAREE, SSGTS., TITI, ERIS, ATRA, STETS, TERA-, AST, SETI, RAES, IRREG., REES, HAIG, TARSI, GIRTHS, ATHS., TRISTE, ASTR., and ERTE. A couple of these in a grid might be okay; any regular crossword solver understands that there may be a few trade-offs to get an ambitious theme to work. But take them all together, and that’s a lot of trade-offs to swallow. For me, the trade-offs weren’t worth it.

Given the many compromises in the fill, it is fair to question whether a puzzle could be more enjoyable if it didn’t labor under such heavy theme constraints. Deb asks if people criticize poor fill like the aforementioned just out of habit, but there’s a simple reason that Rex Parker and Amy Reynaldo in particular critique filler words like RAREE and SIGHER: because those words are not in common use and they believe that crosswords would be better off without them. However habitual it may be for them to harp on poor fill, that doesn’t invalidate their complaints; RAREE and SIGHER and the like aren’t stellar words to use in any puzzle. Too many of those types of words in a single puzzle, and you’ll likely hear about it.

That brings me to Deb’s other question: what about those harsh critiques from puzzle blogs? How much does the possibility of a tough review because of some less-than-ideal words limit a constructor’s creativity or willingness to take risks with a potentially ambitious and clever theme? I don’t know, and I don’t want to presume to know how other constructors feel about that. As someone who cut his teeth in crossword construction by reading Rex Parker and Diary of a Crossword Fiend for years, I can say that their critiques have helped me immensely in understanding what makes a puzzle better, but obviously, that’s not everyone’s experience. Still, I’m not sure that Deb’s question represents the most useful hypothetical. Of course Rex and Amy and others play a key role in shaping opinion of a puzzle’s merits, but it’s the constructors who create the puzzles in the first place and the editors who publish them — they are ultimately responsible for the creative choices in puzzles. If an influential blog helps constructors steer clear of poor fill, then I don’t see the downside. Better fill is better, right?

(I’ll concede that getting a tough critique on something you’ve worked hard on isn’t fun — I’ve gotten them, too, and it’s fair to note that Rex’s review today is pretty scathing. But if a constructor enjoys making puzzles enough, they’ll keep making them anyway, regardless of what crossword bloggers might say about them.)

I would also submit: let’s not conflate “creativity” with willingness to take risks with subpar fill. A constructor can have many creative ideas and still prioritize smoothness of fill over a theme that may be proving unwieldy. If something isn’t working, the constructor can try starting over with the same idea, or try a new idea entirely. Besides, if a puzzlemaker decides that a certain crossword theme is proving unfeasible because the fill becomes too compromised, is that necessarily a bad thing? It gives a constructor the opportunity to assess their own work frankly. Look at this Greek letter-themed rebus grid here:

SACA, A BANE, ALALA, UPL, SHI, SIR SLAM (?!), THE TALE, DEATH SPIRAL TRAPS (?!!)….. it’s dreadful, right? You bet it is. I know it’s dreadful because I made it. I built it more than six years ago. It was the first full 15×15 crossword grid I ever completed. I never submitted it anywhere, but I still keep it around to laugh at it and remind myself that I’ve come a long way since then. Nowadays, I approach crossword construction with a lot more confidence. I developed that confidence not despite having much greater concern for fill, but because of it.

When asked about what makes a successful crossword, Trip Payne once said the following:

“My #1 rule has always been: It’s All About The Fill. Of course you want a great theme and clever clues, but the second you resort to weak entries, you’ve lost my interest. A lot of people are willing to’justify’ weak entries because they’re ‘necessary’ to pull off a wide-open grid or an ambitious theme; I don’t agree with that. With enough work, and perhaps a willingness to pull back a little from the original concept, you can pretty much always avoid poor fill.

Look at Patrick Berry: his themes are great, and you’d have to inspect his puzzles with a microscope to find a weak entry anywhere. That’s not magic — he just has high standards and a willingness to put in the work to make his puzzles as good as they can be.”

That captures my own sentiments well. It’s for good reason that many crossword constructors (and, I would guess, far more crossword-solving enthusiasts) hold Patrick’s work in high esteem: they’re clever and smooth, with hardly any compromises in the fill at all. In terms of early-week puzzles, I’d put Lynn Lempel‘s work in the same category: smooth, smart, and well-executed. Yes, you want a theme to deliver a satisfying a-ha moment, as Trip says, but when it comes at the expense of fill (which comprises most of the puzzle anyhow), then the puzzle suffers.

I’m in agreement with Deb that people should encourage puzzle constructors to think outside of the box and try out new ideas to see where they lead. So here’s what I’d recommend to constructors: don’t be afraid to experiment with themes, but don’t be afraid to start a puzzle over from scratch if the fill isn’t cooperating either. In the end, constructors don’t need to choose between holding the fill to high standards and maintaining a creative puzzle — making a puzzle as smooth as it can be is creative.