Jeff Chen is a writer and professional crossword constructor living in Seattle. He currently operates the XWord Info website, where he writes commentary on every daily NYT crossword.
Years ago, the NYT began to modernize its crossword evaluation process, morphing from snail mail to various incarnations of email-based submissions, and finally to its current web form. It still has some wrinkles, but new constructors today have no idea how lucky they are. Today's puzzle was accepted back in 2018 but unfortunately fell through some infinite loops into the Phantom Zone. Oh well, what can you do?
What you can do is take the opportunity to reflect on yourself. I printed the puzzle, confident that I'd be able to solve it in speedy fashion. I didn't remember all the themers, but when you know the central entry is OXIDATION, it gives you a head start.
Breathe, Jeff, breathe …
To my chagrin, I couldn't solve the lower right corner on my first pass. Or my second. Or my third. That seems to bode poorly for solvers who didn't create this puzzle.
I've taken flak for making puzzles too hard, and I got a firsthand opportunity to see why that's incredibly annoying. If I were to redo this crossword today, I wouldn't place so much emphasis on jam-packing the corners with "strong" entries, instead prioritizing making the grid more conquerable.
The constructor's first and foremost job is to set up solvers for a win. After doing this for over a decade, I'm still learning that I still have much more to learn.
Dan is a pleasure to work with. Not only can he more than hold his own technically, but he's not afraid to ask the tough questions and pause or restart a build if something feels off. Most importantly, he tries to keep the question "what would make solvers happiest?" above his own needs.
In the spirit of Lincoln's ability to love his enemies and work with opponents, I wanted to highlight something Nate Cardin said a few weeks ago that both impressed and inspired me. So many people today so vehemently berate the other side, convinced that this will force the damn morons to come to their senses. It often does the opposite, causing further and deeper entrenchment.
I'm hopeful that his message might help a few people — myself included — to move toward a more open-minded and generous-of-spirit mentality, where thoughtful discussion and polite debate might produce positive change.
"Hidden words" puzzles are a tough sell these days, Will Shortz is getting a ton of these types of submissions. When John approached me about doing one using reversed words, I didn't have much hope it would go anywhere. But why not take a look?
As a solver, I'm most impressed by long hidden words, so I took the brute force approach of sifting through all possible medium-length words (5+ letters) to see if they were hidden inside any long (8+ letters) entries from the XWord Info Word List. After my computer nearly passed out from heatstroke, we had a long list to sift through.
There was a lot of dull stuff, but one line item caught my eye: ADELE, hidden backward in CHISELED ABS. Even this pop culture idiot knows ADELE! Once John suggested BACKUP SINGERS, it felt like we were onto something.
Thank goodness John is more well-versed in pop music since I wouldn't have recognized LORDE and DRAKE if they knighted or quacked at me.
A person who can do everything from theme-storm to code at a high level to put solvers' needs above his own? All the while making the process fun? Dan is a dream to work with.
Plus, he pushes through my obsessiveness with a smile. Let's say tolerates. Call it a non-frown. Okay, he doesn't toss me out on my bum.
I used to be a huge fan of "The Simpsons," so when he proposed this general concept, I enjoyed the throwback to Bette Midler's It's time to TAKE OUT THE TRASH.
Usually, I'm the one doing all the coding in a collaboration, so it was a great change of pace that Dan has programming chops. With two nerds on the case, algorithmically testing out different synonyms for trash didn't at all make me feel down in the dumps.
I know, that was rotten.
It's not only a Big EZ working with Lisa, but an immense pleasure. Two of the most important qualities in new constructors are listening to and processing feedback, and the dedication to grind through hard work. Lisa is exemplary in both areas, as far from a BZZT from me as possible.
It's such a delight to help someone so deserving make her debut today.
Lewis and I are both musicians, so I loved his idea of presenting these repeated letter strings as if they were within repeated measures of a score. Choosing ones that are four letters long — to mirror four quarter notes in 4/4 time — seemed perfect!
Then I waffled. Would non-musicians figure out what those dots were all about? I imagined non-musicians trying to erase what they thought were printing errors, flecks of ink that had dripped onto their crossword. I talked Lewis out of it, seeing if we could come up with a different, non-musical revealer.
After a few days, Lewis convinced me back to the repeat signs. That is, until I thought about the poor electronic solvers, forced to read a note to the effect of [The grayed squares are meant to have pairs of dots on their ends as if repeat signs in a musical score]. Blech!
Thus, I persuaded Lewis away from the concept again. It took some doing, but we agreed it was for the best.
Until Lewis convinced me back!
Shows who's the brains in this operation.
Finally, we agreed to a compromise: incorporating a revealer that would make things clear, even if you didn't see the pairs of dots or had never picked up an instrument. REPEAT isn't the most cunning revealer, but it at least gets the point — or points, get it?* — across for non-musicians.
Jim Horne probably thinks that Jeff, the former trombone player, should be sackbutted for that dotty pun.
Appropriate that Chase and I spent an eternity filling this one.
I hope you noticed the infinity grid art in the center of the puzzle. If not, we've highlighted it in green below.
Why green, you ask? Well, duh! Because green represents the Time Stone from Marvel's Avengers movies — the stone that has the ability to rewind events while simultaneously splitting the multiverse's timelines into particulate tachyon matter that both exists and does not exist within the minds of millions — and also the minds of none.
(It's because green is shiny. I like shiny.)
I didn't add one word to Chase's commentary.
Okay, it was more like seventeen.
What, you saw a mini-theme in KAMALA HARRIS to YOU'VE GOT THIS to NOT AFRAID?
Seeing as I try to keep politics out of crosswords, you must have consumed one too many SAKE BOMBs.
While working with Mary Lou on this grid, I wondered if A LA MAISON would be interesting enough to include in the bottom-left corner. I thought it was funny as a foreign equivalent to something like Ali G IN DA HOUSE!
My French friends said, "Mais oui!"
While test solving, Jim Horne said, "Hmm … I SEE." (Meaning, not so much.)
Curious how that corner treats the Saturday solving population.
I've been fascinated by Patrick Berry's SPLITS AND MERGERS puzzle for years. It speaks to both my finance and my crossword technician leanings. How the hell did he figure those crazy things out?
I finally got a chance to meet the introverted Patrick at the ACPT a few years ago, and it was great to get some insight into his process. Not surprisingly, there was some programming involved, but also some clever holistic problem-solving.
It was so interesting to hear him say that sometimes he simply looks for cool letter sequences and patterns to build themes around. I figured, if he can do it, why can't I?
Bwa ha ha ha ha!
Given that I had no idea what letter overlaps might be interesting, or even what kind of overlaps to consider, or where, this entire project would have been DOA if it hadn't been for my wife, Jill. After weeks of brainstorming off and on, the process littered by stupid ideas followed by yet stupider ones, she picked out one of my grid skeletons and said, "What about COMMON CORE for that one?"
Every day, I thank the stars for our merger.
Jim Horne curates our Grid Art page, which is one of my favorites, so we figured, why not try to add something cool to it?
What, though? Plenty of crosswords have used black squares as grid art, even some that employed rare diagonal symmetry. My favorite was one depicting a kite, but there's been a host of animals, too, including fish, birds, panda, dog, and even a pinata. Lot of these are Bruce Haight's … Bruce ...
Hey, wait a second! Jim. Jim! Let me tell you all about Bruce Wayne's journey to becoming Batman. We'll start off with the comic book lore, move onto the new era where Batman fought Superman, continue into the early movie days, discuss the pros and cons of each of the four people who have been Robin, although there is controversy about the exact number—
Jim? Hello? Did your line accidentally get disconnected?
Originally this comic book nerd wanted to do a giant Bat Signal in the sky, but that felt a bit too obvious. It had nothing to do with the fact that all my Bat Signal attempts looked like Bane smashed his fists into the grid.
After noticing that WAYNE MANOR and THE BATCAVE were matching lengths, that felt like something! Until we realized that regular crossword symmetry would place THE BATCAVE in the opposite corner, not directly underneath. Mirror symmetry wouldn't work and Will Shortz dislikes up-down symmetry. Once we settled in on diagonal symmetry, the bats sort of emerged on their own.
Hopefully, solvers will notice the bats after uncovering THE BATCAVE. Or at least, that effort won't drive them batty.
The Joker made me say that.
Who would have thought a YouTube channel about putting together jigsaws would be interesting? To my surprise — especially since I don't do many jigsaws — I've been gripped by Karen Puzzles. Whether it's her beat up on the Try Guys or getting her take on Jigsaw Treasure Hunts, I can't stop watching. Super fun to see those PICTURE PERFECT MOMENTS at the end.
I liked Christina's general idea of making a jigsaw crossword, especially in a time when a ton of people are stuck at home, doing jigsaws. That would be amusing!
Or would the reminder of pandemic-related activities be annoying? Hopefully not.
Speaking of annoying, some of my rough ideas made Christina say something to the effect of "I'D RATHER BE STUCK AT HOME FOR ANOTHER YEAR THAN DO SOMETHING THAT TERRIBLE!" In a kind way, of course. One concept involved sections that you'd have to solve, then mentally move those pieces onto chunks of similarly-shaped black squares, and that would complete a bunch of surrounding answers.
It made even less sense than that explanation I just gave.
Thankfully, Christina nudged us back into a more productive (less idiotic) direction, and after many iterations, we landed on PICTURE PERFECT MOMENTS in a 3x7 grid. Making sure that:
There were some picture imperfect moments during the process.
When I work with a new constructor, they often give up after a few back-and-forths after realizing that die-hard persistence is needed to spark a crossword idea. If they do stick around through my first rounds of feedback, most of them lose hope after we find a previous puzzle that's too close, or if we decide we need to regroup and shift direction to generate a stronger a-ha. I try to help every polite person who comes to me, but few people make it through to a published product.
Celeste is not only one of the few that gutted her way through the process (across dozens of different ideas), but we clocked in at two months of theme brainstorming for this one.
We had started with Celeste's proposed LONG TIME NO SEE and a few others, which like the seed of something interesting. A few more like TEE OFF TIME and UNEASY FEELING … but I had an uneasy feeling that this felt too similar to something I'd seen recently. Research ... alas.
We were on the road to scrapping it, but Celeste gave it second life with a few new possibilities, including one phrase with two missing letters. That smelled interesting, but we couldn't come up with enough in-the-language phrases that worked similarly.
One morning a few weeks later, I was in charge of a morning that involved rejected waffles, much crying, and a stray poop (don't ask), making me wonder how I got into such an unenviable position. Unenviable position ... wait a second! Silver linings of having young kids!
I appreciate how willing Celeste was to roll up her sleeves and dig into the grid filling. A 140-word Sunday is daunting, but once I broke it down into smaller, more tractable sections, she tackled each with gusto, taking feedback with open ears.
I often tell constructors that it takes about 20 ideas to generate one solid seed concept, and even that one doesn't always gain traction. I love it when someone doesn't quit when things get tough. Congrats to Celeste for pushing through to the finish line!
I often tell new constructors that it takes maybe 20 ideas before you arrive at one that's a worthy crossword seed. Even then, it might take several iterations to turn it into something interesting. Most newbs give up well before this point, but not Amy. I enjoyed going back and forth with her on several concepts, and even after we landed on this one, the finished product looked nothing like what we first envisioned.
Hopefully, no one will fall out of CLOUD 9 because of the total lack of symmetry — that peskily asymmetric number 9. We experimented with different arrangements that allowed for some grid symmetry (around the perimeter), but this one lent itself to best fill. We figured if you're going to go asymmetric, why not go whole hog (heaven)?
The first time filling a grid can be overwhelming, the learning curve steep. I like that Amy jumped right in, trying to fill it by hand, on graph paper. With giant regions, that's nearly impossible to do, so I segmented the grid into smaller, more doable regions, and she sharpened her pencil and got to work. I was impressed by how much iterating she performed, as evidenced by the amount of eraser dust I saw in her pictures!
I appreciate working with open-minded people who listen to feedback and think about solvers over themselves. Amy hit all those marks. The theme felt so appropriate for such an enjoyable collaboration.
SET THE BAR LOW? UNDERACHIEVE? Wow, this sounded like a terrible, or at least terribly pessimistic, puzzle. Glad that I didn't skim over Owen's email. The concept tickled me, Owen coming up with in-the-language phrases that perfectly described the LIMBO steps.
He proposed STAND TALL as a final themer, but that's not how you do the LIMBO. The correct answer is FALL FLAT. At least, the way I do it. Followed by CALL THE DOCTOR.
It would have been great to make a LIMBO stick out of black squares toward the bottom of the puzzle, with a person LIMBOing underneath, but it was not to be. I did enjoy working with Owen to weave in some subconscious LIMBOesque fill, like TONED IT DOWN and SAGGIER. And of course, BEMOAN the fact that attempting to do the LIMBO makes this old chap's GLUTEI AROAR.
Why does the proximity of GLUTEI and RECTOR make me laugh?
Talk about setting the bar low.
P.S. I worked in MATTHEW and RHEA intentionally. Now, I expect their firstborn ... unless they can guess my name! Bwa ha ha! What, they guessed it's Jeff? Darn it.
I can't resist programming problems, and John challenged me with a doozy. How hard could this be?
It didn't completely break me, but it took some trickery to figure out a combination of manual sifting and automated processing to rake through the search space.
Matching results: 0.
Well, except for the trivial closed chains of AIR MAIL to MALE HEIR and TOW PLANE to PLAIN TOE. Even then, I had to think whether I'd use PLAIN TOE as a theme entry — it's definitely a term used in shoes, but it's rather … plain.
Maybe if we could figure out a third pair, we could create something interesting? Off to the computer to generate lists of phrases based on switched homophones! Thankfully, it didn't take long before THYME TEA emerged. I didn't have it in my list because I'm not much of a tea drinker. Reading up, it felt solid enough to include — hundreds of thousands of hits with all sorts of reputed benefits.
Maybe I should become a tea drinker.
It's not easy to work with a bunch of mid-length theme answers, but we eventually figured out an arrangement that would let us work in some sparkly long downs. Hopefully, things like BALLGOWNS crossing GLITZ will keep you interested, if the theme isn't your cup of (thyme) TEA.
(I happen to know a certain six-year-old princess-loving daughter who would squee for the combination of those three things)
I love programming challenges — some of the most ecstatic moments in my life have come from working on difficult coding problems. One puzzle hunt challenge stands out, where the organizers left a tiny loophole in one of the instructions, and every team achieved the optimal 100 — except ours. Nothing can compare to that smug feeling you get when people yell how the #$%#! did you score 124?!
Oh yeah. Except for getting married. And having kids.
Now, if I could just score 124 on any parenting test.
Today's problem — discovering long words and phrases where all the letters are in reverse alphabetical order — is fairly trivial. The main challenge is figuring out how to smoothly translate letters into numbers in a data structure that allows easy—
Oh, right, you don't care. Better if I say it's magic and leave it at that.
It was much more of a challenge to figure out how to present as many interesting finds as possible. I was disappointed there weren't more long ones besides TOOK HEED and SPOON-FED. That meant featuring a bunch of 7s and even some 6s. It's notoriously difficult to create a smooth grid around a slew of shorties, but the theme might have felt too thin if I had packed in less.
I struggled mightily in the NW corner, unfortunately, the most important corner in most crosswords, since it sets the tone for the rest of the puzzle. I was already at a high number of three-letter words (editors balk at more than 22), so I couldn't break up PORSCHES.
I switched the theme answers and/or shifted their placements umpteen-thousand times, but nothing worked better than what you see. It pained me to use a partial as ugly as US DO, but at least I could use a great original quote from "Star Trek VI"! Ka'plagh, Starfleet scum!
Huh? "If you prick us, do we not bleed" is Shakespeare?
(insert WOOKIEE roar here)
I thoroughly enjoyed working with Francesca, appreciating her nose to the grindstone approach on both themer exploration as well as gridding. Our original submission got positive feedback from Will Shortz and Joel Fagliano, but they expressed concern with some of our themer finds. They were skeptical if it would be possible to find solid replacements, and frankly, I was too.
Thank goodness for Francesca and her optimism! Many people would have taken the feedback as "thanks but no thanks" and dropped it, but she chose to dig in — from Spain, where she had been quarantined at the start of the pandemic! It took us several weeks of back and forth, but we eventually uncovered enough alternatives to get the go-ahead.
Gridwork is often troublesome for newer constructors, so I was impressed by how quickly Francesca took to it. I showed her a couple of alternatives for grid skeletons — I hope a few people noticed the dollar-sign(ish) S in the middle of the grid? Anybody? Bueller? She jumped right in, working her way up the learning curve, asking all the right questions about what makes for solid or undesirable fill, doggedly filling chunk after chunk, iterating time after time, never despairing over all the rounds of feedback.
A delightful experience. Hopefully, the finished product is worth some of your spare cash. Er, spare time.
(In case you're still baffled by the theme, TIME has been replaced by different synonyms for MONEY, forming different but also in-the-language phrases.)
I met Derek a few years ago at an ACPT and enjoyed talking with him. He's got an easy demeanor and is so relaxing to be around. His reviews over at Crossword Fiend are some of my favorites since he always manages to keep things positive, making his essays a joy to read.
When he expressed interest in doing something together, it was a no-brainer. It took a while to find something we both thought was interesting — some of my first ideas were terrible, especially the concept of appropriately mixing yellow and black.
I enjoy being irreverent, but I had a feeling my ancestors would rise up and kung fu my ass for that.
Alas, a 108th anniversary puzzle doesn't quite catch people's attention. Seeing as Lee didn't want to wait 92 more years, perhaps it was better to shelve this one.
But wait! You all know I love me some grid art, so as I was reading up on this explosion — expelling 30 times more ash than the St. Helens eruption, for 60 hours? — I began to wonder if we could make something resembling a volcano out of black squares. Maybe with ASH blowing out the top? And LAVA flowing down the sides?
The grid turned out to be a red hot gusher, spewing crossword goo everywhere we tried to fill. 60 bleary-eyed hours later …
I wish we could have opened up the mouth of the volcano for better solving flow, but nothing worked. Ironic problem for a puzzle about a MOUNT blowing its top.
For those of you who don't know SHIBUYA, put it on your bucket list. I've had the fortune to travel to Japan for work a dozen times, and seeing all the crazy outfits and styles was always a highlight.
As with all highly constrained constructions, we had to figure out some trade-offs, and if a bit of HOO, ISA, ETD, VOL inelegance bugged you, that's on me. I give you free rein to kick my ASH.
But Crucivera is cruel. To make that happen, I'd have to intersect a pair of theme answers through the revealer, and none of the letter pairs within STRAIGHTEDGES both had straight edges. Starting from the outside, heading inward: S/S, T/E, R/G, A/D, I/E, G/T. Bah!
I briefly considered opening up my "straight edge" criterion to the letter I so I could make mirror symmetry work (crossing themers through that I/E pair), but that felt like a cheat. Solvers tend to write in the middle of boxes when possible, so the letter I wouldn't mesh with all the pre-printed "straight edges" justified left. People might write an I down the middle of a box and wonder why there was an extra line pre-printed to its left.
Huh? You're asking why I couldn't have placed STRAIGHT EDGES in row 12, and then placed my four themers at the top of the grid, in columns 3, 6, 10, 13? No revealer intersections necessary, since the longest vertical themer would have ended in row 10?
I didn't do that, because ... technical reasons ... feasibility studies ...
Shut up, you.
I enjoy having a crunchy problem to wrap my brain around. This one wasn't as hard as others — it's a simple matter to search for palindromic letter patterns buried within phrases — but it was still a fun diversion. I doubted we were going to find anything of interest, and I didn't think the strings would be long enough to be worthwhile discoveries, but I was pleasantly surprised.
As usual, shows what I know!
We even had the option of using only long(ish) strings that happened to be words in their own right (like PLAN and OMEN). We debated the merits of that approach, but eventually decided to pick the longest strings that were buried within the most colorful base phrases.
The grid was a bear, as crossing pairs of theme entries are notoriously difficult to grid around. Some of the palindromic strings weren't friendly, either. I struggled getting the skeleton together — all those crossing themers taking up so much real estate — until finally landing on a fortuitous overlap, COURSE CREDIT enabling a workable grid backbone.
I usually advise sticking to 140 words to achieve a strong balance of color and cleanliness, but taking out a black square at the first S of CLASS SIZE bizarrely seemed to help. It made sense after studying the area — four letters ending in O gives more choices than three letters ending in O — but I wasn't wild about ENIAC. We kept banging away at that section, and couldn't come up with anything better at 140 words, so it was what it was. Ah well, hopefully nerds everywhere are okay with it. Beep boop.
We originally had GOAT YOGA in the GOATHERD slot. Will Shortz balked, wondering if it was well-known enough, and if it has staying power. It was probably the right call, although if you're looking for staying power, maybe a couple of baby goats on your plank is the answer. Hard not to squee at that.
Ever since Michael Bennett's introduction in which he declared his true allegiance, I've wanted to get WAKANDA into the crossworld.
It was a pleasure helping Tracy achieve her goal of "hitting for the cycle." She's hard-working, responsive, and a great listener.
As is often the case with Sunday 140 word-puzzles, the middle is the hardest to get right. We tested out so many arrangements of black squares, so many options where GADSDEN now sits, but nothing was working. You start off with your most desired entry at a critical spot, test the heck out of it, then if it doesn't work out, move to the next highest. Suffice it to say, GADSDEN was not at the top of our list. Still, I'm hopeful that it at least sounds familiar from history class.
We had a good back and forth with Will Shortz on several entries. IRL (in real life) was tough for him to accept, since it's not uber-well-known yet. I use it all the time, but I can see how some might scratch their heads.
We made concessions to Will in the NE, where we had MURICA where PALACE now is. What is MURICA? What are you, a commie pinko? ‘Murica is the good ol' US of A, damn straight! Okay, maybe Will made the right call there.
Jack and I went back and forth a dozen times on the southwest corner. I tend to err on the side of cleanliness, but the trio of DITHER + ORNERY + TOUPEE made for such an evocative image. Calling Donald Trump!
Our first draft had too many names for Will's taste, and he had a point. If you don't know SCARJO (player of Black Widow!), GEENA DAVIS (Thelma! Or was it Louise?), LEILA, LESLIE Jones (hilarious on SNL), Julian ASSANGE, and of course James GADSDEN (you didn't know him? infidel!), our first draft would have irked you even more.
I enjoy working with constructors who keep an open mind to feedback, and who are willing to put in the time and hard work required to get a theme and a grid to an appropriate level of quality. Jack's one of the good ones.
It felt fitting to end this theme with THE NSA. They're everywhere — especially in (A)THENS
John-Clark and I brainstorm upon tons of ideas. For this one, the (H)OLDS OUT to OLD SOUT(H) pair stood out, but there weren't many other long ones. We tried combining pairs into longer phrases for kooky results, but EMANATE MANATEE! wouldn't exactly wow anyone.
Not only that, but would solvers even notice what was going on? Would they groan, thinking it's yet another anagrams puzzle? And would it even be possible to create a grid with enough pairs to be interesting?
After some consideration, I sadly moved it into my "not worthy of a full crossword" graveyard.
But John-Clark is persistent. He thought it would make for an engaging solve, and picked out a few pairs as an example set. I was doubtful that it was possible, given that it would need to feature HEAD TO TOE somewhere, and also integrate at least three pairs of themers. Seven themers in a mirror symmetry arrangement? Impossible!
Well, not impossible. Improbable. Improbable, in terms of making a high quality grid, that is.
Long story short, Will Shortz liked the idea, but not the cramped grid I laid out. High-quality, it was not.
Improbable, indeed. Back to the graveyard, evil undead!
But one night, as I wondered if expanding to a 16x15 grid would help, the zombie rose. Fast forward past several sleepness nights, and I sent John-Clark a new grid skeleton that passed my testing.
See? Making a crossword is yeas! Er, easy.
I'm often full of piss and vinegar when I see low word-count themed puzzles. Frickin' constructors, bigger than their britches, showing off all lah-di-dah! And look at today's bleedin' grid, 74 words, requiring LTD and CRIT? Why not follow your own advice, blankety-blank big-head Jeff, and go to a 76-word design?
I would have preferred that.
And another thing, you blimpity-bloopity … wait. What?
To Neil's point, I couldn't figure out a way to arrange the skeleton at 76 or 78 words such that it didn't feel like two half-puzzles. The black square separating FLYING and COLORS created more problems than I anticipated, causing a diagonal from SW to NE. I tried so many ways of creating a passageway through it, but no dice.
Putting black squares at the AR of ARREAR or at the S of UNIONS would have made for an easier, more newb-friendly Tuesday solve, no doubt. Both options would chunk apart the grid, though. Is that better or worse than a newer solver having to figure out IBERIA and RIYADH? After 10+ years of constructing, I still am not sure.
I'd love to have some feedback mechanism, where the NYT's app would allow for direct evaluation and commenting. Was the puzzle solvable? Fun? If not, why not? That type of feedback loop would be so valuable. Will Shortz has done a great job of expanding the NYT's crossword business, but it could be a true empire with some steps such as this.
P.S. If you're not sure what a GLITTER BOMB is, why aren't you watching Mark Rober's YouTube channel? Dude does some amazing stuff.
Tough grid to figure out, with awkward themer lengths. Interesting finds, but of length 10, 8, 9, 9, 7, 7? Bleh! Good thing I happen to enjoy mirror symmetry ...
I did the skeleton work, and Erik and I went back and forth, shifting things around, testing, filling it piece by piece. That we have different tastes should come as no surprise — I loved CLOWN NOSES at 11-D but he wasn't a fan, and when I proposed the IRON DUKE (awesome nickname for Arthur Wellesley, the 1st Duke of Wellington) I felt his side-eye straight through my dial-up modem.
We were having some issues with the SE, and then Erik magically solved it by incorporating INUKTITUT. My first reaction started with a "what" and ended with "the F is that?" But when I looked it up, I decided I loved it. All educated solvers should learn that word. (Including me.)
Great pleasure working with Erik on this one. And by "working," I mean "sitting back and enjoying being in the presence of jeeus."
I enjoyed going back and forth with Christina on grid filling once Will approved the current set. She pointed out that it'd be elegant to not have any other Ys in the grid, so I wrote some quick code to eliminate any entries containing Y.
But it wasn't working. It should work! I told the computer to do everything right! Why the frick would Crossword Compiler tell me that I asked it to accomplish an impossible task?
Ah. The Y in Y2K was already fixed in place. I instructed my computer to use anything that fit a ?Y??? pattern (at 31-D) … using a word list containing no Ys.
It's a good thing I don't work in computer science.
This was the third or fourth puzzle Christina and I developed together, and her speed up the learning curve was amazing. Usually, by this point, I'm still leading the gridding efforts with a newer constructor, but it was Christina's suggestion to insert CALVINBALL, with the surrounding areas tested to make sure they were fillable. So much of the color in the grid is all her.
A-HA, GOTCHA JEFF! You always snootily points out the flaws in others' work, and look at today's grid! AMTS, BSA, HST, MAI, MER, SASE, SYS, XIS. That's not clean! I THOUGHT YOU WERE A STICKLER FOR CLEANLINESS IN GRIDS, WISEGUY!
It's a fine point, sir. (I say "sir" with confidence, as 99.72% of my outspoken haters are mansplainers. Also, "sir" is best said using air quotes.) Even though this count of gluey bits is lower than the NYT Sunday average, eight is usually too high for me, even if they are all minor.
It doesn't even seem like the grid would be that hard to build, right? Three long entries, plus six pairs of shorter ones? Any doofus could build that grid! However, having six pairs of short entries means that you have to spend a ton of your black squares separating them. We had to deploy so many precious black squares separating LEVITES and MOUSSES (on both sides), for example.
Well, fine. Couldn't you have gone to a higher word count, like you often chide constructors about?
Will and Joel were kind to let us push up to 143 words, three higher than the max. It is true that we could have gone higher, but that would have meant taking out some goodies, like NO PEEKING, REDDITORS, GEE WHIZ, PINE LOG, etc. It's only rare constructors who can make a 144 word puzzle feel interesting.
Hopefully Christina and I landed on a decent balance of color and cleanliness, and that the theme makes it all worth it. It'll be interesting to see how many people write in with comments like LET ME EXPLAIN EXACTLY WHY PITCHY DOESN'T MEAN EITHER "CONCISE AND TO THE POINT" OR "DESIROUS" YOU FRICKIN MORONIC "SIR"!
(PITCHY without the C of CASTOR = PITHY, and PITCHY without the P of POLLUX = ITCHY)
We constructors talk about "seed ideas" (for themed puzzles) or "seed entries" (for themelesses). Why not actual seeds — or S E E Ds? It seemed like a natural way to seed a puzzle based on ornamental POTs made of black squares. Easy peasy!
Over two hundred e-mails back and forth later …
As a constructor, I try to steer clear of long "bonus fill" across entries that aren't part of the theme. These muddy up what is thematic and what is not. To our dismay, this grid layout — largely fixed by the four POT placements — required a long across at the bottom of the puzzle. Bah! No matter how hard we tried, we couldn't break it into pieces.
Wait. How about making it thematic? Of course! A perfect solution, let's just do that!
Several thousand searches on our Finder and onelook.com later …
We wanted to have each herb break across multiple words of a theme entry, but we eventually had to make a compromise. GROWTH POTENTIAL was a strong enough bottom revealer that we figured we could let MESSAGE BOARD slide by.
Many thanks to Will Shortz and the team for sticking with Matt and me as we struggled our way through this one!
You don't know ze difference between MACAROONS and MACARONS? Pah! (Said in a snooty French accent.)
Fun experience working with Ezra. He approached me with the idea of using DOUBLE BLIND as a revealer, and it immediately struck me as having huge potential for something interesting. Ezra's original concept for how to play on it didn't grab me, though, so thus began the long brainstorming process.
Once we hit upon the idea of having two sets of doubles within a single themer, it started picking up steam. But that didn't seem like enough. Skipping random pairs of doubled letters is okay, but it would seem much more deliberate if the letters were chosen for some purpose, not just whatever worked.
Off we went down internet rabbit holes, thesaurus.com and onelook.com my best friends (and worst enemies). But once I happened upon CLONES as a possibility, things began to click. Several (dozen) hours of coding, scouring, and struggling to match crossword lengths finally turned up our set.
Then it was just — "just," ha — a matter of filling the grid so that it looked like a perfectly normal crossword, except that those 12 special squares would be omitted. That might sound daunting, but it's not so bad. Just a lot of trial and error. Except for that pesky UNDONE / UDON middle region. That was an enormous pain.
And wait! Ezra pointed out that we shouldn't have any other doubled letters in the grid. Good point — er, god point. That seems like it should be easy enough a constraint, right?
Sigh. That eliminated some of the most promising ways to fill each little section. Who knew, an innocent EE here, an overlooked OO there, and JEFF AND EZRA SMASH.
And the one crossing entry I thought would be no prob? The MACAROONS slot. Surely there were plenty of words that became other words when one O was removed.
Overall, a lot of pain, a ton of rework, and dozens of hours sunk in. I'm hopeful our focus on the solving experience translates into smiles today. Sometimes I wonder if tricky puzzles like this are fantastic for a niche group of solvers, but leave everyone else confused / annoyed / calling me a macaroon.
It's so cruel when you come up with an exciting seed idea but can't find other examples. In my mania, I ran through at least three dozen X OR Y phrases. Plus or minus. Give or take. At first, I did all my sifting and sorting manually, but that felt like writing a book by typewriter. Why not tell my computer to do the work for me?
Hello, computer, find phrases using the X OR Y pattern, and then find partial strings where both X and Y can form valid words. Easy peasy!
Stupid computer and its inability to do what I want. Grumble grumble.
It took a couple of sleepless nights to realize that I could write a program that would find phrases containing the X string of letters, then substitute in the Y string, and finally check to see if the result was in my word list. Hooray!
One of the first findings the program spit out was HOT OR NOT and LIKEAS___.
Aargh, idiotic computer, I'll kill you, die die die! LIKE A SNOT isn't a thing!
Oh. LIKE AS NOT. Right.
Then I stared at the next result, FRIEND or FOE and DE___. WTF is DEFOE, the opposite of de friend in Brooklynese? Ha ha ha dumb computer, you made another mistake!
Ah. Daniel Defoe. Ahem.
You're only as smart as the weakest link, which is not my computer.
My original submission used:
CAN YOU BELIEVE IT, IF YOU FILL IN A SINGLE LETTER FOR EACH BLANK, YOU CAN SPELL E R A S E!!! THAT'S SOMETHING AWESOME, RIGHT?
Well, it should be.
Things happen in threes. First, it was the AVXC, then the WSJ, now the Universal crossword. I wrote this puzzle in late 2017, and Will accepted it in mid-2018. I realize that the Venn diagram overlap of Universal and NYT solvers is likely small, maybe minuscule. That didn't make it any less painful when I opened up Mary Ellen's puzzle and got that sinking feeling.
I'd love to see Will improve the lag time between acceptance and publication. Over a year is tough to swallow. It's easy to see why some top constructors have shifted to other venues, notably the WSJ, American Values Crossword, and the New Yorker. It's a tremendous loss to the NYT crossword that Liz Gorski has gone off on her own, with a fantastic subscription service.
It's a tough ask, though. Reducing that queue time would require a systems overhaul and a much more nimble operation. It would mean going from a paper-based system that Will prefers (he likes to mark up snail mail submissions) to an electronic one, where he could better organize and keep track of how long any particular accepted puzzle has been sitting in limbo.
And for what? To massage the fragile egos of constructors? Yeah, but he's already built an incredibly successful operation, with hundreds of thousands of digital subscribers, getting 125 submissions a week for seven precious slots. What works for him is doing just fine. So why mess with success?
(Although a retooling of Sunday puzzle querying is badly needed, since of those 125 submissions, only a tiny fraction are Sundays.)
Sometimes it's best to eat one's frustration and focus on the good. It's illy say — er, silly — to not concentrate on one's appreciation and thankfulness. Even after 100+ publications, I still get a lot of joy from giving a few people out there 10 minutes of (hopefully enjoyable) diversion in their day.
Historically, I haven't been a fan of space exploration, considering it a waste of resources that could be redirected to the real-life problems all over the globe: infectious disease, poverty, inequalities of all sorts, you name it. Why on Earth (ha) would we spend money on non-Earth purposes?
But when Jason pinged me, I read up on the events leading up to the landing and was reminded of how it brought a nation together. So many folks worked so resiliently through hundreds of failures, never giving up, always regrouping to emerge stronger than before. And during that historic moment, all eyes were glued on the team of brave souls, people from coast to coast witnessing the fact that there was no such thing as impossible.
I'd love to see America united like that once again.
Such a tough grid to construct. It's usually not THAT hard to work with stacked themers, if some of them are shorties. But we had such little flexibility — the partial word had to start directly under the H in each HAND, and we only had a few to choose from.
Besides IWORK, ICAPS, YMEN, there was only the less interesting IER, ILY, LER, EDIN. I would have loved a fourth that was more interesting, but of the less ideal candidates, EDIN seemed at least interestingly kooky — something you'd never see in another (reasonably good) crossword.
Took me a couple dozen tries to arrive at a skeleton that tested out reasonably well, and I counted myself very lucky that BADPR existed to satisfy that tough -DP- letter combination, as well as DNA LAB. What else is going to fill a DN???? pattern?
I don't like using "corner blacks" (in the very SW / NE), as they are visually unappealing to my eye (Rich Norris at the LAT likes them even less than I do). But sometimes you gotta do what you gotta do. With an already biggish SW corner, and the ultra-constraining DNA LAB, we could barely escape with a SABE / OCULI.
Not the most elegant corner in the world, but sometimes you have to accept compromises in the service of a theme. Hope it didn't pull down people's solving experiences.
Every puzzle's a struggle. So many aspects to balance out. Hope we arrived at the right one for you.
Seems appropriate that MACHU PICCHU has UP in the middle of it. The Incas and their clever wordplay!
I don't often build crosswords around short themers, because these shorties tend to get lost amongst all the long fill that (generally) has to go in a puzzle. But I liked the concept and theme set well enough, that I decided to give it a try. How hard could it be?
Hard, as it turns out!
Just placing the symmetrical pairs and laying out black squares to separate them was tough enough. It took a couple of dozen tries to come up with an arrangement that seemed fillable.
But I had to deploy so many black squares in the center of the grid — four sets of three diagonals takes up a lot of real estate — that I ended up needing at least a few long across answers (where STAGE ACTOR and STATIONARY are now). That felt problematic, because wouldn't solvers wonder how these two were related to the theme?
Why not go whole hog, I reasoned, and throw even more long across fill in, that it felt more purposeful as fill? Surely that would fix the issue.
What? You wondered how SUPER SLO MO was part of the theme? Darn you!
Finally, I rarely use cheater squares in the way that I did just before KRAFT. It just looks fugly to me; seems cheap. But darned if I couldn't figure out a reasonable fill in the SW corner without it. I tried every usual trick in the book, but every single one dead-ended at some bizarre piece of fill.
If you haven't gotten the chance to go up to MACHU PICCHU — and more importantly, even higher to the peak at Huayna Picchu — that's one to put on the bucket list. Not for the faint of heart, faint of breath, or faint from fear of falling to one's death off of towering heights that have virtually no safety measures, though.
And if there were, there surely couldn't be a way to cross them in pairs, through a shared letter.
And if we could somehow make that work, there'd be zero chance that all of the synonyms would be containable in solid phrases.
And if by holy hell all THAT managed to be doable — fat chance, I said — there was NO FRICKIN WAY that it could be laid out in a crossword grid.
As it turns out, we found more than enough synonyms, BUCK the only one we weren't able to use because it shared no common letters with any other synonym. (Sorry, BUCK ROGERS!)
They managed to cross into pairs.
What? All six had solid phrases like FAWN OVER?
BUT. But but BUT! There was no way we'd possibly lay out the crossing pairs — plus DEER XING — in a grid.
The great goddess Crucivera is cruel! Gnashing of teeth!
I hate giving up on an idea when it's that close to working, so I tried dozens of arrangements using normal symmetry, and then dozens more with mirror symmetry. Finally, I hit upon this layout, which seemed to be the only one possible one.
Forgive my blasphemy, oh great and glorious Crucivera! I will sacrifice many pencils to you.
Tracy did a wonderful job of filling, especially in the SW corner. We had pegged this as a Monday theme, accessible to most anyone, so we wanted to make the fill as newb-friendly as possible. In some iterations though, I had put awesome (to me) material in the big bottom corners, proud of all the juice. But looking back on it, much of it was too tough for newbs. Tracy helped us dial it back, favoring smoothness over going gonzo. I learned a lot from her Monday approach.
Great fun to help Sophia develop this idea. When she came to me with a couple of themers, I wasn't sure there would be enough material. Or whether it'd be more fun to have VENUS DE MILO in the clue or as the grid entry! A lot of unknowns.
Our spreadsheet for book titles and possible "biography" subjects extended into the hundreds. After narrowing it down, it was unfortunate to discover that the winners didn't pair up at all for crossword symmetry. Fie on thee, crossword gods!
But I've done enough mirror symmetry puzzles that it felt like things had a chance to work out that way. After realizing that THE ONCE AND / FUTURE KING could split out 10/10, I decided I might have been too hasty shaking my fist at the crossgods. Then, finding the fortuitous intersections of A GAME OF THRONES / GONE GIRL and LORD OF THE FLIES / LIFE OF PI made me stand up and say I'm a believer!
Word lists aren't nearly as important as grid-building experience, but (shameless plug alert) the XWord Info lists sure helped with this one. It's incredibly hard to create a clean and colorful 140-word puzzle, what with all the huge white spaces you have to work with. Sophia got stuck several times, and twice, a single entry helped me get her unstuck: ETOUFFEE in the NE, and BODY SURFS in the SW. Building this puzzle with only entries that come with a standard dictionary-based list would have been possible, but only with severe compromises.
Sophia's in college now, so doesn't have that much time for crossword construction. But after she graduates, I think she could rise to the top, if she decides to put more time into it.
Jeb and I meet up a couple of times a year, and it's fun to brainstorm with him. He showed me one built around OUT OF ORDER SIGNS, but he had flipped pairs of sign words, i.e., WANTED HELP and WAY WRONG. I didn't find that interesting, but there was something about the revealer that tickled my fancy.
Several weeks and several hundred anagrams later, voila! I liked the image of a "Fawlty Towers" imp or Bart Simpson scrambling road signs. It amuses me to think about people confusedly driving past a sign reading DECAL ODORS.
Well, amusing, until they ran off the ROAD that was CLOSED. Ahem.
I almost always do the grid skeleton work in collaborations, but this one was all Jeb. I hesitated at his layout at first, especially considering how many down answers had to work through the first two themers (in the north region of the grid), but he made it work pretty well. Always interesting to see another constructor's process — I would have shoved SPEED LIMIT all the way to the right to create better spacing. Pleasantly surprised to see that it all worked out. Just took a few back and forths to figure out best options for each region after that.
We were hoping that this would run on a Wednesday, as I want more out of a Thursday theme. If you're gonna toil away at a tough solve, the payoff better be good! Hopefully at least a few of the anagrams were amusing enough to be worth your time.
P.S. Jeb and I have a much more juvenile taste than Will. Our submitted anagrams:
Jeb asked me to come on board after Will gave him "I like the concept but not the grid" feedback. My first impression: I thought it was fun to have the snakes snaking around the puzzle, but I wasn't so sure about WATCH YOUR STEP. I so badly wanted something more … snaky.
Then it hit me! Indiana Jones's IT HAD TO BE SNAKES! Counting … 15 letters = perfect!
Only, not perfect, since that's not the exact quote. Shows what I know.
The first set of revisions were painful. Eventually, I came up with a grid skeleton that seemed to work well enough. Some testing showed it was likely doable. Fast forward eleventy-thousand hours, and we got something that at least was clean. Good enough!
Except, not good enough for Will. He pointed out that it was smooth, but there was little to no juice in the bonuses. We had to use a bunch of dry stuff for the long slots, and where was the fun in that? Excellent points.
Back to the drawing board.
I noodled around with some different snake shapes, including one where a single snake started from the middle, but that led nowhere, quick.
Thankfully, Jeb came up with the idea of intersecting ALL of the snakes to the middle! That curiously made things easier, as all the theme material took up less net real estate in the grid. I worked up some possible skeletons, focusing on squeezing as much pizzazz out of the long slots as possible, and voila! It still took time, but only maybe eleventy-hundred hours this time.
(But this huge Indy fan still wants the revealer to be IT HAD TO BE SNAKES. Accuracy, schmaccuracy.)
I love working with Jim. He's thoughtful, brilliant, talented in so many arenas (check out his Hamilton cover band!), and best of all, funny. While we were working on this puzzle, I was endlessly amused by one of his clue suggestions for ARCHIMEDES: [One of his inventions really screwed things up].
His more risqué version: [The guy who invented screwing].
This particular themeless was borne out of a curiosity: what new 10-letter entries from our XWord Info Word List would make for great themeless seeds? (When it comes to themelesses, 8-10 letter entries are the easiest to build around, and there's a larger quantity of 10-letter entries compared to 8s and 9s.)
For several weeks, we culled out a sub-list, adding only entries that
I thought we'd be able to incorporate at least three seeds into this standard-ish sort of themeless grid, but only MARS LANDER and MBA STUDENT (go UW Huskies, class of 2002!) made it in. As we flowed the fill from right to left, it became apparent that we'd get more colorful / cleaner fill if we didn't fix anything further in place. So we didn't push it.
Hopefully, more seed entries from our list will show up soon! Jim and I are always working on something fun.
Today I'm thankful for ML's perseverance. She comes up with some of the toughest concepts of any collaborator, and she doesn't give up when they prove to require dozens of hours to iterate. Never say die!
This grid was one of the top ten hardest 15x15s I've ever worked on — maybe even top five — and I nearly threw in the towel on four separate occasions. Rebus squares are easy to work with, but two in close proximity is asking for trouble.
I almost said no right from the get-go after I heard that number!
It's so satisfying to have worked through the grinding process and come out with a grid that we're both happy with. I'm so glad that ML included me in this concept — it's a great feeling to have all that hard work lead to a publication.
I've been interested in MICROLOANS for a while now, and I still think that it's an interesting concept even though the research doesn't show great economic results. The jury's still out on whether the MICROLOAN industry will eventually provide wide-ranging financial lifts to the poor, but it was a fun enough crossword conceit that I put aside my rebus fatigue (I rarely work on them these days, considering how few editors take them) and agreed to hop on board.
Such a tall order, to spread out the five rebus squares. But I understood where Will was coming from — not a lot of fun to work through a rebus puzzle and then realize an entire region wasn't thematic.
Getting a fifth square around the MICROLOANS revealer was so tricky. Tucking it into VARIOUS / ENVIOUS didn't make for fantastic rebus entries (I prefer jazzy multi-word phrases like GRACIOUS ME or broken across words like WHY I OUGHTA), but it did end up working out.
Wish we could have worked in just a bit more spice in addition to STEINEM, ZINGERS, ALADDIN, IT HELP, but you can only do so much when working with a ton of constraints. Made for such a tough set of trade-offs. Hope it was enough to keep solvers happy!
The pleasure was all mine! Johanna is a delight to work with.
It's too bad Johanna's original concept didn't work out: kooky phrases resulting from BRA removals. I mean, BRAHMS LULLABY to the HMS LULLABY = a ship of peace? That's awesome!
Unfortunately, there just weren't enough others. BRAIN SURGEON to IN SURGEON, BRAIDED ROPE to IDED ROPE … hmm. Flat to non-sensical. Johanna came up with ZEN HUSSY (from BRAZEN HUSSY) … I decided it was best not to say anything on that one.
The grid skeleton was a pain to figure out, what with all the themers taking up so much real estate. I didn't like leaving those big NE / SW corners, but I couldn't figure out a way around them. Props to Johanna for tackling them with gusto! I thought we'd have to add a cheater square, maybe at the C of CLEVER. Not so, for CLEVER Johanna — CREDIT where CREDIT is due!
Truthfully, I wasn't hot on the idea itself at first, as it felt like an unspecific hodgepodge of Yiddish words that have made their way into English usage. There are too many of those for me to feel like was a tight enough theme.
So initially, I politely declined. But Scot is persuasive, plus I like him a lot and wanted to help. We went back and forth, changing theme material from his original submission, i.e. deleting the oddly-spelled SHLEMIEL, adding SCHMALTZ, etc.
Finally, we added YIDDISH as a revealer — I felt strongly that at least this would tie things together a bit. It still wasn't a theme that delighted me, but at least it worked well enough for my taste with YIDDISH smack-dab in the center.
It's not the most innovative theme ever, but I hope we worked in enough fun YIDDISH words and bonus fill to keep solvers entertained. Who doesn't like saying OY GEVALT!
As I was filling the grid, I realized 38-Down could be YINGLISH, which seemed so fun! But ultimately, it would have been duplicative of YIDDISH. I also wasn't sure if that term was precise enough for this theme. Ah well, perhaps YINGLISH will find its way into another crossword.
Trickiest part of this construction was figuring out to get enough of the seven "theme" entries to be 1.) long and 2.) interesting. Wouldn't be too fun for I be hidden in IMAC, for example, since IMAC and MAC are so close that you'd never notice that something was amiss.
The fact that all seven themers had to start at the top of the puzzle took away a lot of flexibility, so it wasn't easy to lay out the grid skeleton. I wish I could have made all seven as long and interesting as RENUNCIATION to ENUNCIATION, but it just couldn't happen.
Hopefully, some of the shorter ones played out in a fun way. I liked ISLAM to SLAM pretty well. And it was funny to think about the Greek god ARES streaking as he BARES his bottom.
I wanted to work in more long bonus fill, but it was tricky — I didn't want to confuse the issue of what was theme and what was fill. As it was, ESPERANTO looked a bit too themey to me. But breaking it up would have pushed the grid over the 78-word limit, and more importantly, the puzzle would have been swimming with oceans of boring short fill. Bleh!
Hopefully, no one spent hours wondering how ESPERANTO fits into the theme. Sorry if you did!
Anyone notice the black squares in the middle of the puzzle look like an iceberg? No? Not even if you squint?
WELL SQUINT HARDER DAMMIT.
Building around an 11-letter entry can be difficult, so I did a ton of grid testing before settling on the upper left corner you see today. It took a lot of shifting black squares around before I was finally happy with the combination of AU NATUREL, SPARE TIRE, TEE SHOTS, which all seemed to have great potential for clever clues.
Some solvers may not be familiar with SHROOMS, but one of my college roommates used to love that word (he hated SHROOMS on pizza). It's more commonly referring to the hallucinogenic drug, but SHROOMS will always make me think of Chris Wand.
I almost gave up on the grid pattern when flowing toward the SE, but thankfully KIBITZED looked like it could connect things. I enjoy kibitzing the world's best bridge players, trying to figure out how they engineered some brilliant endplay or squeeze.
Thankfully, most everything else came relatively easily from there. I was unsure if I could get Z IS FOR ZEBRA to work, but the pieces dropped into place. Not only did I like the two Zs in the entry, but I had thought of a fun clue for it, misdirecting from an (elementary school) primer to (paint) primer.
I had a ton of flexibility in the upper right — I love me some dinosaurs, so the IGUANODON made things easy — but the lower right was rough. I was hesitant to use STATE CAR, as it felt a little dull to me … until I read up on the Wikipedia page for Official STATE CARs. I spent way too much time reading about the various weapons on some state cars. Whoa! And I love Will's clue, so fun.
I was a huge Paddington Bear fan in my youth, and now my daughter is hearing about the clumsy fella's exploits. I'll always have a soft spot for MARMALADE.
I would have loved to drop this from 68 to 66 words — more of a personal challenge — but I couldn't get rid of the black square between GREG and SCAB without sacrificing some snazziness or cleanliness. Ah well — the quality of the solving experience has to be the priority, not the constructor's intellectual stimulation.
Fun to work with Sande on this one! When I collaborate with a newer constructor, I often end up doing the lion's share of the grid layout and filling, but not this time. I constructed the skeleton, but it was Sande hacking away at the grid, showing me versions that numbered into the dozens. Most of my efforts were just in pointing out problematic spots and giving suggestions on some piece of long fill that might work better than others.
I love seeing that kind of work ethic — many other constructors throw up their hands at my overly critical eye toward grid design, but Sande fully embraced it.
There's a good amount of theme material, what with the 9 — 14 — 7 — 14 — 9 lengths, but it's usually not that hard to work with. The middle 7 is especially friendly, compared to a middle 9, 11, 13, or 15. So typically, I'd be loath to finish a grid like this without at least four snazzy long bonus entries. But the no-I constraint turned out to be tougher than I first thought.
Well, it would have been easy to work in even six pieces of long bonus fill, if we had been okay with accepting globfuls of ELL, RUR kind of stuff. But I had imagined this would run on a Tuesday, so I pushed Sande to keep that glue count down to a bare minimum. I thought he did well in that regard.
Always the trade-offs, though. Maybe we could have made the long slots sing a little better? I like BYE WEEKS a lot, and ELOQUENT is pretty good. But I sure would have liked to get something more out of EQUALLY and SEAWEED (we ended up having to put in cheater squares to facilitate better fill, so these turned into seven-letter entries, which tend to be harder to convert into colorful stuff).
Turns out there are a lot of words and phrases that use the letter I! No wonder it turned out to be so difficult to fill with color and cleanliness. Learn something new with every puzzle.
Executing on this grid was such a damn pain. Stacking pairs of themers is rarely easy, and I felt like I needed a bunch of pairs, since they were all so short-looking. Stick in ENTER at the middle right — where it is on a keyboard — and put five theme pairs in. Couldn't be too hard, right?
Why do I torture myself?
Even without real themer symmetry (gasp, I know!), I had so many problems in so many places in the grid. Take the SE, for example. I tried shifting CHICK / IYAKI up, adding cheaters, shifting black squares around, but what you see was the best I could do. It bugged me to cross SHEEHY with AIMEE — should educated solvers know at least one of these two? I think so, but I'm not 100% positive.
I started by filling around each themer pair, then tried to knit everything together, using snazzy long answers. That was painful. I do like OPERA STARS, FREE SPIRITS, STILETTO, even CREDENTIALS. But ATTRACTANT felt so boring. With only six long slots, I so badly wanted to make them all count.
And although ATTRACTANT forced me to do things like CEES (ick!) and THANES (boo!), it was the best of all the dozens of attempts I made in that SW section. Working around a long answer in ATTRACTANT plus THE plus TAINER was so rough.
As a sanity check before submitting this grid to Will, I asked Jill if she knew who Chrissy TEIGEN was. She nearly hit me upside the head, because OF COURSE SHE'S CROSSWORTHY! It was a huge relief, as whenever I ask Jill if something/someone is crossworthy, the answer is usually a silent stink-eye.
Sande comes up with some of the most creative ideas of any of my collaborators. I liked his initial concept of doing something Scrabble-related in a crossword. Having four "racks" around a scrabble board, plus a "bag" made out of black squares (look at the center of the grid again!)? That was pretty good.
Then he came up with the idea of TILES inside the bag, mixed up into STILE and ISLET. Even better!
But I felt like we had to have Scrabble-esque words in the four racks, not what he proposed: any four seven-letter words that would help us produce good fill. I worried that this would kill the idea. Ah well, lots of ideas need to die a natural death.
Then he came back with PLAYERS ANAGRAM JUMBLED LETTERS. Oh so perfect! Now it was just a matter of working up a grid skeleton.
Nothing good by revision 10. Tricky lengths to work with, considering we had to build around that central "bag."
Revision 30? Still flailing.
Somewhere around 50, things started to look up.
Back down by rev 60.
Finally, I thought, what if we punted on a 15x15 grid? 16x15 or 14x16 is no problem for the NYT, but those didn't help. Finally, I tested out a weird 13x17 on a lark and bingo! It all fell into place.
After 40 more revisions.
Thankfully, Will thought he could bend some rules to the get the odd shape to work with his syndication partners. What a relief! Revision number 101, clued and filed as final.
The sight of a single Scrabble TILE will fill me with Pavlovian shudders from now on.
This crossword was an enormous pain to put together. Like, trying several hundred times to fill it, with several dozen different black square skeletons, with countless hours lost to dead ends and restarts.
I'd be able to work in a few good long entries, but then I'd run into some area where I'd need a gluey bit to hold it all together.
Or I'd figure out how to turn one of the many corners, but then be unable to turn the next one successfully.
And then it must have been 20 or more times that I got 95% of the puzzle finished, only to figure out that I couldn't get the ends to mesh.
I love puzzles with wide-open grid flow, where you can solve without feeling like you run into dead ends or segmented corners. But man oh man does that make filling a grid so painfully difficult.
It should come to no surprise to people who know me that I originally clued POWER UP as something like [Awesome ability gained by a video game avatar, such as the embiggerator or the super-immortality mushroom for Mario].
My gridding philosophy has always been to pack in as many strong bonuses while minimizing gluey short fill. Sometimes I wonder if I overdo it, though, making a puzzle too tough for solvers. I've recently had a string of people tell me how hard my puzzles are.
I love me THE FLASH, PHENOM, EXPONENT, all of which I think everyone ought to know.
Okay, I only expect people to know at least two of the four people who have been the Flash, not all four. I'm not a monster.
But did I overdo it with FLIP TOP, TRI TIP, LSD TABS? Not sure. Even BILGE RATS, which I love (I take full advantage of "Talk like a pirate day") might make a solver's life challenging.
As a constructor, I want to produce puzzles that first and foremost are solvable, and that hopefully entertain. I'm never sure if I accomplished that. So I continue to seek feedback and try to learn from it. You can reach me through the "Contact us" link on the home page if you have constructive criticism.
Final note: the grid would have come together a lot easier if I didn't feel like I had to avoid any non-thematic doubled letters in the fill (aside from the necessary EE in SEEING DOUBLE, of course). Turns out it's tough to eschew entries with EE, SS, even OO. All those entries went POOF! I have a feeling no one even noticed this, but man, my compulsions are strong.
I so badly wanted the mixed-up ingredients to go vertically, each one intersecting MARINARA SAUCE (which would run horizontally across the center of the puzzle). How cool would that have been, like each ingredient being dropped in! But it was not to be — I had a limited pool of options for each ingredient (ONION, TOMATO, GARLIC, HERBS, in case you didn't figure them out), and I couldn't get the symmetry to work out.
Gol durn crossword symmetry!
I even tried different "recipes" to get such a fortuitous interlock — PESTO SAUCE with BASIL, PINE NUTS, etc. was the closest to maybe possibly sorta kinda iffily working — but nothing would fully cooperate. Dratted PEN I STUN. Er, PINE NUTS.
To those who have never played ZORK, you have not lived. Okay, maybe you have, but still, ZORK was such an integral part of my childhood, a sort of augmented Choose Your Own Adventure that got me obsessed with computers and programming.
BTW, apologies to those who scratched their heads at the PROSODIC / ZORK crossing. I figured whether you're a techie or a fuzzy (engr/sci vs. humanities), you'd likely know one of them. (PROSODIC is related to "prose," so you should be able to suss it out, yeah?)
I always try to work in a bunch of bonus entries, pushing and pushing, hopefully stopping before having to hold the puzzle together with a bunch of ugly crossword glue. And although my efforts to draw from different areas so that there's something for everyone — MAKES BANK, MON CHERI, INKBLOTS, ARGONAUT cover a lot of ground — I couldn't resist putting in MC HAMMER, even though I had already worked in AEROSMITH.
What can I say, I love me some Hammer pants.
Letter(s)-addition puzzles usually have all sorts of themer possibilities, and OM is a pretty common doublet. So it came as a surprise to me that there were so few possibilities. I thought three themers (in addition to PREMEDITATION) would feel thin, so thankfully, there were just enough that worked.
No constructor is dumb enough to use themers of length 12 / 11 / 9 / 11 / 12 (PREMEDITATED had to come last). It's completely inflexible, forces themers to be squished up and causes all sorts of headaches in the middle of your puzzle. No matter how hard we tried though, we couldn't come up with any shorter themers as replacements. Drat!
There were so few possibilities in the layout that I thought the concept might be dead (or necessitate a larger grid). Thankfully, the one you see seemed to work out … although it had a ridiculous number of long slots everywhere in the grid. It felt like there was no way it could all work out, especially given the pesky J within OMITS NO JOKE.
But somehow, it did! There was precisely one possible skeleton — I like to have at least 5-10 to consider — so it was a minor miracle. NET JUDGES isn't something I strive to incorporate usually, but man, did it save our butts.
I didn't intend to make it a low word-count puzzle, but that's what it ended up having to be.
Josh helped figure out what the best options for each corner might be. Fun to do the optometry test — which is more clean and colorful, A, B, or C? Since Josh is doing plays these days, we prioritized getting play-related material in. Glad to work in OFFSTAGE right at 1-A.
When I ask a collaborator to explore the solution space for a theme, I usually get a couple of ideas. Maybe a dozen. Rarely, a Word document.
Pris created a spreadsheet. Awesome!
The spreadsheet grew and grew … doubly awesome!
… to over 300 entries. HEE HEE HEE SQUEE!
I love it when a collaborator is more than willing to put in the work. There are so many *pretty* good movie titles on that spreadsheet, but Pris was never happy until we found one that made us both laugh out loud. Took a long time to hit on LICENCE TO KILT, but man, do I love that one.
Tough grid to put together. I wanted the circled P L O T T W I S T letters (hopefully you noticed that hidden phrase?) to look like they were twisting, like a DNA helix. Crossword symmetry didn't make that perfectly possible, but I hope this layout came close.
My stubbornness to stick with this twisty layout meant we didn't have much flexibility in themer placement. I ended up using more cheater squares than I like, but hopefully it doesn't look too overrun by black squares.
Nine themers is never easy to work with in a Sunday 140-word puzzle, but we tried to include a little bonus fill for everyone. We debated whether TERMINATORS was too close to the theme? But both of us enjoyed that one so much. And then BLACK OPS, EAR CANDY, SPEED DATE, MUSCLE CAR, CELTIC HARPS, ERIC IDLE seemed like they were from different enough walks of life.
Who knows, maybe there's even some SYNERGY there …
For years and years, Liz Gorski had been doing wonderful Xmas dot-to-dot Sunday puzzles for the NYT. It had been a while, so Mary Lou and I decided to try to carry on the tradition.
But what shape to make? We threw out a bunch of ideas, and the one that stuck was RUDOLPH prancing through the skies, with his RED nose lit up. Easy-peasy, right?
Um … no. We needed to flesh out the idea with some themers, and ML found a nice split of THE MOST FAMOUS / REINDEER OF ALL. Along with some other assorted RUDOLPH-related answers, it felt like the grid wouldn't be so bad to construct.
So very, very wrong.
Like, miserably wrong.
Just testing out different arrangements of where to start the path took forever, as we kept having to adjust the position of every single circled letter. And then sometimes one felt like it might be amenable … except for some stupid little corner. I probably restarted the process 50 times before hitting on the current arrangement.
But it just got harder from there. The biggest problem was that it's really hard to make a 140-word Sunday puzzle, period, and it's much harder to do it when every region of your grid is constrained. Sure, we could shift around letters a little bit here and there, but then they'd get in the way of the themers, or they'd make RUDOLPH look like he was limping, etc.
I should have studied Liz's work more carefully — in most of her dot-to-dots, she did well to spread out her letters through the grid in such a way that they didn't interact with the themers, or she used much fewer than 26 dots.
Live and learn. Given how much frustration and sleeplessness it caused me, I doubt I'll ever make another dot-to-dot. (Unless it just has two dots. Anyone have ideas for a LINE theme?)
Oh right, the puzzle. After coming up with this idea, I wrote a Python script to help me figure out possibilities. That helped, but it still took several hours to sift through hundreds of possibilities. Some were cluable in the "X in Y?" format I wanted, some were snappy phrases, but few were both.
Then it was just a matter of figuring out matching lengths, grid creation, roughly eleventy hundred grid iterations, and most of my remaining hair plucked painfully out one by one with red-hot tweezers.
You know, the usual Sunday construction experience.
Aside from POSSE CUT, sometimes I wonder if my assumptions about what educated solvers ought to know are getting skewed. KARPOV to me is a no-brainer — you have to know at least Garry Kasparov and Anatoly KARPOV. HATHA YOGA? No matter where you are within Seattle, you can throw a yoga ball 50 feet in any direction and hit a room full of downward facing doggers. DULCINEA? Come on, you ought to at least recognize the name from "Don Quixote," which you were supposed to have read in HS/college!
Okay, I only pretended to read it, too. Fine!
Overall, I hope there was enough from different walks of life to make everyone a little happy, and not too much to make people feel like it was a trivia quiz rather than an entertaining word game. Probably will be some of the latter folks ... ah well. POSSE CUT!
Jerry got in touch with me after several submissions and resubmission to Will and Joel on this idea. I was lukewarm on the triply-doubled letters concept, as I had seen it once or twice before, but Jerry hooked me onto the project with the magic words: "Will thinks it might just be too hard to fill a grid like this with high quality."
I'm so darn predictable!
Jerry's original grids definitely had too much crossword glue for my taste, but more importantly, there weren't enough themers for my taste (four, plus TRIPLE DOUBLE), and some of them didn't hit my ear well. HEEL LOOPS I wasn't familiar with, but that did appear to be a real thing. WOOD DEER was the one that threw me. It took a lot of Googling to find enough examples of famous WOOD DEER carvings to be able to shrug it off as passable.
Even then, I felt like the puzzle needed more meatiness. A quick Python script helped uncover a few more, including one that I wish we could have incorporated: QUEEN NOOR. Ah well.
Putting the grid together took roughly eight hours over two sessions. I kept on ending up unhappy with some little corner, some piece of long fill, or both. For a while, I wondered if Will was right — even more so now that I had added two more themers to the original set.
I've had some luck with stacking themers for high-density puzzles, so I experimented with that. When I hit upon the arrangement you see, everything seemed to test out well. Plus, three themers in the top and three themers in the bottom makes for a double of triples … or TRIPLE DOUBLE!
(Sort of. More or less. If you squint a little.)
The rest of the grid was just a matter of trying out a few hundred possibilities in each corner. Not so bad.
I'm still not 100% sold on WOOD DEER, but if you aren't either, the other themers will hopefully help you forget it!
Locking CON MEN and MARK ANTONY into place made for some teeth-gnashing. After my first pass, I ended up with CON MEN crossing ... NEOCON. Is that a dupe? Not really, since one CON stands for CONFIDENCE and the other CONSERVATIVE, two words with different etymologies.
But every time I tried to convince myself that it was perfectly fine, it bothered me mightily. Maybe having something that looks like a DUPE is appropriate for this puzzle (groan), but I just couldn't live with it. There went another few hours, into the garbage.
Apologies to anyone whose sensibilities were offended by PORKY'S, but talk about a cult classic. It's hard to believe they got away with naming a character "Beulah Balbricker."
I had fun working on the lower right, trying over and over to maximize my impact with those long slots. I'm a fan of Latin phrases, physics, and throwbacks to the 80s, so QUO VADIS, BETA RAY, UNDEROOS made me happy. Talk about a disparate trio of terms.
I also spent a ridiculous amount of time debating the middle section. I was in finance for years, so I've always found the term I BANKER (investment banker) colorful. (Not always ethical, but colorful.) Was it worth ORANG? I've heard people say ORANG before and I'm fine with it, but I realize it's not everyone's cup of tea. I could have gone with something like STINKER instead, but really, isn't an I BANKER the same thing? (*rim shot*)
ADDED NOTE: Jim didn't know what an I BANKER was, and had a hard time convincing himself that it could possibly be correct. Maybe I've spent too much time in finance ... dang it, maybe I should have gone with STINKER!
ML asked me to come on board with this one after an encouraging rejection from Will. It's so tempting to try to fix a few little spots here and there on a themeless … but often, the skeleton and basic structure make that near impossible.
We decided to do a total teardown, starting with ML's (and Will's) favorite entries: I MEAN, REALLY! and RIJKSMUSEUM. I love the former as a colloquial phrase, and the latter as a place of incredible beauty (and weird consonant patterns!). If you ever get the chance to travel internationally and can't figure out where, the Netherlands is amazing.
As I mentioned yesterday, 11-letter entries are problematic for themelesses. ML originally had stacked sets of 11s in the corners, but that's a rough way to start — you force a bunch of 3-letter slots right off the bat, and too many of those can make for a choppy solve. So I repositioned those 11s more toward the middle of the puzzle.
It quickly became apparent that a few cheater squares would help out, and I'm pretty liberal in that aspect of constructing. I didn't like nibbling away a precious 8-letter slot into a 7-letter one, but thankfully COSPLAY seemed to work there. (I'd totally dress up as a Klingon at ComicCon, BTW.)
I like trying multiple dozens of versions of every region within a themeless. Luckily, ML graciously puts up with my OCD about exploring the entire solution space. We saved so many versions that I didn't remember which one we finally submitted. Thankfully, ML has a good eye for what's the best trade-off between strong entries and gluey short ones.
Looking back on it, I would have tried to open up the NW and SE corners, which feel more disconnected from the rest of the puzzle than I like. It did let us easily try many dozens of versions in each area, but with so few entries connecting everything up, hopefully solvers don't get stumped on LAKSHMI or RIJKSMUSEUM.
I enjoy working with new constructors, and I really enjoy it when they're like Michelle. Open to feedback, never quitting, willing to work as hard as it takes to create a product that solvers will hopefully enjoy conquering. It was a pleasure to go back and forth with her.
She had pinged me with a very different idea, and it was neat to see her take my feedback about brainstorming … and actually do some brainstorming! A few ideas later, she hit on something that seemed interesting — hidden plays / musicals, i.e. HAIR in BRITISH AIRWAYS, with INTERMISSION as a revealer. It didn't seem quite right to me, but switching it so the play had an "intermission" in the middle felt like it could have legs.
Michelle started a list of possibilities, and we tried to find as many long musicals that could be broken up by short intermissions, like a real musical. HAWAIIAN AIR has a looooong intermission, while the GREEK VASE much more reflects the actual relative durations of a musical's first half, intermission, and second half.
We debated WISECRACKED vs. WINGBACKED — the latter felt very close to a real thing, but WINGBACK seemed much more in the language. I suggested WII WHACKED, i.e. what happens to a TV screen when a kid wildly flings a Wiimote at it. Michelle politely said that might be a little … ahem … esoteric. Good call, looking back on it!
It is incredible that anyone ever listens to me.
I laid out a grid skeleton for Michelle, testing to make sure it wouldn't cause problems, and Michelle took off from there. I coached her a little on what solvers might hate and love, and she did a fantastic job of iterating over 20 different versions to come up with what you see.
This puzzle started out when I noticed ECHO sitting in the middle of LIFESTYLE CHOICE. How neat to put a black square on top of the E, making that spot an ECHO LOCATION!
(insert groan here)
That didn't end up making the cut — stupid crossword gods and their symmetry requirements. It was so tough to find eight total themers (four pairs) that matched in length. So many times I'd get to seven ... and then crap out at the eighth.
After months of repeated failures, I happened to be watching "Sherlock" on BBC, and inspiration hit, thinking of the CANAL ZONE within FORENSIC ANALYST. Thanks, Sherlock!
The grid was an interesting problem in construction, working with all the constraints. I like challenging myself, so I wanted to see if I could do a 136-word grid that would be a fun (and smooth) solving experience. Earlier in my constructing career, I probably would have called some of my first versions good, but I just wasn't happy about some of the fill, no matter how hard I tried.
So I rebooted with a 138-word grid, and the result felt like it'd be a much more fun solving experience, with fewer odd words and gluey bits. Hard to turn off my stupid constructor's brain that's always looking for a challenge — sometimes I have to beat it into submission so I can remember to keep the solver as my first priority. Maybe I should have stuck with a 140-word grid to eliminate things like RATINE? I think it's an interesting word, but it's a tough call.
I was such a fan of "Spy Hunter" (the arcade game) as a kid. Pressing the button for that OIL SLICK was so fun! That's how I clued it in my submission, but big ups to Will and Joel for their brilliant clue. [Disaster film?] is one of those devilishly clever plays on words that makes me mad I didn't think of it myself!
Jeb and I live close by, so we get together and talk shop. During one session, he asked me to review some of his ideas, and I thought this one had the most potential. It seemed a little thin, but I suggested he ought to run with it.
Months later, he got back in touch, saying that Will and Joel liked the idea, but not his rounds of gridwork. After several backs and forths, they thought he might be biting off more than he could chew; that perhaps the grid was too ambitious to construct to their standards.
Those are the magic words for me, so Jeb didn't have to do much convincing to get me on board.
The first thing I did was to switch the order of the themers — Jeb had BY HOOK OR BY CROOK first, and I felt like it'd give away the game too quickly. Easy to swap them.
Then, I tried to rebuild around more theme — how cool would it be if there was something thematic running through the two themers? It took a lot of searching, but I finally came up with something I liked: the GREAT OUTDOORS. Perfect! Well, maybe not perfect, but a nice, descriptive phrase that tied LAMB and FISH together — could be clued with respect to a meadow and a fishing hole. So I spent maybe eight hours coming up with a grid around that.
Jeb's response: "Neat grid! But how is GREAT OUTDOORS related to the theme?"
Not perfect indeed.
So, back to the drawing board with a completely different grid. Ten more hours later, I finally decided that I'd have to accept a trade-off, and SST allowed for a good amount of snazzy fill. I was pretty sure Will and Joel would favor the first (GREAT OUTDOORS) grid, anyway.
Yet again, shows what I know!
16-wide grids can be surprisingly difficult. Just one extra column creates such challenges. One of the bigger issues is that I hate going above 78 words (the usual max), even though one could argue that it ought to be 81 words for a wider-than-normal grid.
Why? As a solver, I'm used to 15x15, and if my solve feels slow, that makes me like the puzzle a little less. So as a constructor, I try to keep a 16x15 grid feeling like it's a 15x15. That usually means using longer words than normal on the whole, all the while keeping things smooth. Not easy.
In this grid, I felt like we needed at least two pairs of long bonus answers. It's easy to work in the first pair of long downs (BEDHEAD / DOODLES), but it gets exponentially harder after that. Putting in RUB NOSES and IN THE WAY meant using fairly big corners, piling on the challenge.
Finding more long slots was tricky — long acrosses like INFIDELS and MAIL IT IN are usually tricky to incorporate, given that it means stacking them with theme answers, but some testing made this arrangement feel like it'd be doable.
Finally, we debated long and hard over ETERNE — was it worth the price of admission for AMIRITE, RUB NOSES, MUFFLER? Ultimately, it seemed favorable, especially since we didn't have that much other crossword glue in the puzzle. The eternal constructor's dilemma of snazzy vs. smooth rarely has a clear answer. Hope you agree with our decision.
I've always favored the generalist's approach. It's hurt me in some areas, like in business, where 90%+ of C-suite (CEO, CFO, CTO, etc.) want specialists, and deep specialists at that. But I feel strongly that expanding one's boundaries is important to personal development and leads to better deliverables, even in the areas in which you might specialize.
My annoying philosophical discussion aside, I enjoy a challenge. It's tough to get three great 15-letter answers stacked atop each other while not compromising in short fill. I loved RARE STAMP DEALER, and all those common letters felt like they'd make for easier construction. And it was a stroke of luck that I happened upon the combination of I BELIEVE I CAN FLY, which is often played for CINDERELLA TEAMS (I'm a sucker for basketball stories, especially underdogs).
It was more a stroke of relief, given how many other 15-letter entries I tried below RARE STAMP DEALER. I saved several dozen versions and discarded probably hundreds more.
The one hesitation I had was in BOOBOISIE, which was the only possibility given the rough ISIE ending I had backed myself into. I had heard the term before and thought it was hilarious (a take-off on "bourgeoisie"), but I wondered if it might come off as condescending. Thankfully, I could attribute it to Mencken (as I giggled about it to myself).
Huge props to Will and Joel for the USED CARS clue, which I didn't get at first. "… old and tired" as in tired = having tires. Ha!
I originally wanted TRICK / KNEE to be even trickier, with KNEE getting no clue. Probably would have been too evil to get an unexpected Thursday-ish trick in a themeless puzzle though. Ah well.
I probably won't do any more triple-stacks now — what a great learning experience, but there are so many other themeless layouts I'd like to try out and learn from. Hope you enjoyed the solve!
I love a challenge, and another 62-worder seemed like it might occupy me for a while. After deciding on WWIIVET as the seed, it shouldn't be that hard — after all, I've used a similar wide-open pattern before. Heck, two times even! Surely I learned enough about this overall pattern that I'd be able to finish a solid grid in a week or so.
Fast forward to five months later ...
Huge props go to Will and Joel for the clue of the puzzle, [Head scratcher?]. I struggled mightily with that upper right corner, only settling on what I thought was a dull WOOL CAP because it allowed me to work in some fun stuff like ANTIHERO, TEASER AD, SWATCHES, and DUELIST. But I wasn't happy about letting a precious 7-letter slot go to waste. It's a good lesson for me to learn — I shouldn't have dismissed WOOL CAP so quickly as irredeemable. My (scratchy wool) cap is tipped to you both, good sirs.
I debated over the inclusion of XKCD for the longest time as well. As much as I love the strip — Randall Munroe, if you're reading this, YOU'RE MY HERO! — if a solver doesn't know it, he or she has no way of piecing it together. Ultimately, I figured inclusion might be a way to get even more people reading it, and I'd just have to make 100% certain that the crossing answers were easily gettable. Hope that worked out for people.
So glad Will kept my "Rabbit of Seville" clue!
Oof, this grid was a bear. The skeleton alone nearly didn't come together, what with four pairs of long crossing answers needed. And we had to lay them out such that they didn't interfere with each other! Eight long themers is tough enough in a 21x21, but try intersecting pairs of them and cramming them all in. Not easy.
And then there was stupid old crossword symmetry to worry about. Tracy put together a good set of theme answer pairs to start, but we quickly realized that having to intersect answers at the E of RED each time was going to give us a symmetry nightmare. Specifically, the placement of the letter E within BLUE vs. GREEN was not good.
A few days later, it occurred to me that if we pluralized GREENS, we could make everything work. Tracy disliked the pluralization, though, and rightly so — a lone plural themer is inelegant. My a-ha moment might have come when eating HASH GREENS — er, (BROWNS) — which is much more natural in the plural.
Once in a while, the crossword gods throw you a bone.
The grid came together fairly quickly after that. I laid out about 20 skeletons before landing on one I thought would work. It also looked like we might be able to go down to 136 words so we could add in a whole bunch of snazzy material. But after a lot of back and forth, we came to the conclusion that it just wasn't going to be worth the crossword glue needed to hold those (not-so-snazzy) long fill together. I hate giving up on a tough construction challenge, but we decided it was the right thing to do. Going up to 138 words allowed for much, much cleaner fill, and still a good amount of KOHLRABI, TEND BAR, STOPGAP, RAZOR WIT, CATSUIT, I DOUBT IT, etc. sort of bonuses.
Such fun when a collaboration comes together like this, where we have to work like the dickens and struggle through so many problems. Finally getting across the finish line hasn't felt so satisfying in ages.
It would have been easy to use a set of random numbers for the redo — FRESHMAN FIFTEEN leading to FRESHMAN SENATOR and SWEET SIXTEEN leading to SWEET CAROLINE were two of my favorites — but that felt inelegant. I wanted some set or progression. It took a long time to figure out a set that worked with 10, 20, 30, 40.
If you've never tried to create a grid in which themers have to be at specific numbers, don't. It's like hitting yourself in the head with a hammer. A white-hot hammer. A white-hot hammer with an angry shark duct-taped to it.
I didn't want to use so many black squares, and I hated going up to 78 words, but every sub-78 word grid I completed (about two dozen of them) had some small entry that rubbed me the wrong way. Or didn't have enough colorful material. Or sticky shark bite prints.
Big relief to finally get the "we're good to go" email from Joel!
Matt! I love this guy. We've only worked on a few puzzles together, but I've enjoyed his sense of humor and his work ethic. Always striving to improve fill, even if the improvements are just incremental.
Really fun to brainstorm better examples of friends, Romans, and countrymen. Took a lot of research to turn up good options for "Romans" and "countrymen" especially, and we debated mightily the pros and cons of PRAETORIAN GUARD vs. VESTAL VIRGINS. We ended up proposing a set with each.
Then Matt and I went back and forth on grid design, taking roughly 30 variations/revisions to finish. It's not easy to work with 14-letter entries as the first and last themers — they mess up grid spacing — so this can make it very tough to weave in long bonus fill.
But it's important to me to incorporate as many strong extras as possible, as these fun entries can help to keep solvers' attention. So we kept on trying different themer placements and black square layouts, always testing to make sure the entire grid could be filled smoothly. Hopefully, the final product has a little something for everyone, things like ST PETER for joke fans, WENT YARD for baseball fans, OPEN BARS for--
(running off to the OPEN BAR)
I had queried Brad Wilber at the Chronicle of Higher Education with a list of made-up compounds in fantasy: UNOBTAINIUM, KRYPTONITE, DILITHIUM, ICE NINE, MITHRIL. He said it was too much of a list puzzle (imagine that!) for him, but UNOBTAINIUM would sure make a great themeless seed.
I don't make many themelesses these days since everybody and their brother submits a ton of them to Will. Plus, I think it's a much more interesting challenge to come up with a great theme. But when I heard NERD CULTURE in an audio book, I quickly counted it out (stupid compulsive constructor's brain, just let me listen to the book already!) — 11 letters, the same as UNOBTAINIUM! Felt like a mini-theme I could get behind.
Unfortunately, 11 letters is a terrible length to work with in a themeless. It's so stubborn, forcing placement of certain black squares right off the bat, draining flexibility. I tested dozens of possibilities, all of them leading to some dead-end or requiring 70+ words (that's not a very interesting construction challenge to me), before finally arriving at what you see.
I wasn't thrilled about three middle entries — PERSIANS, CADENCE, SANCTION, but each felt like there was possibility for good cluing. SANCTION, for example, is a word that can mean one thing … or its opposite. (I clued it as [Approve … or prohibit].)
The upper right took hundreds of tries before I finally felt that the result was sparkly and clean enough ... with one hesitation, anyway: I love ASCII ART, and it fits great with the NERD CULTURE mini-theme, but would solvers struggle with the SIRENIA (it's the order containing manatees) crossing? I love that SIRENIA term, derived from the Sirens of Greek myth. Hopefully, people who don't know ASCII can think about the term "Animalia" and get the -IA ending that way.
My original clue for ANACONDA was [Sir Mix-a-Lot's "don't want none / Unless you got buns, hon"]. I was pretty sure Will wouldn't be able to use it for the NYT audience. But I had to try.
Shout at to BRAD at 26-Across! Thanks so much for the idea.
I love working with Seth — we've had 10ish puzzles accepted in various media so far, and we continue to work on more.
Sneaky devil switched the order of our byline just before sending this one in, though. This puzzle was his idea, so he deserves to be listed first.
I grew up playing cello and trombone, so this classical music-lover didn't even hesitate when dropping in Die MOLDAU, Smetana's famous composition. Seth gently nudged me, asking me if that might be replaced?
Nonsense! I said. Surely, Die Moldau is in everyone's consciousness, or should be. Plus, it was critical in holding that corner together, allowing for nice goodies like MAGNUM, APROPOS, and the DEATHLY Hallows.
Will and Joel thought we did a very nice job on the grid … except for the odd entry, MOLDAU, which "they had trouble swallowing."
Yet again, shows you just how much I know. It's a mystery why anyone ever listens to me.
ADDED NOTE (from blog reader Mark): "Glad MOLDAU remained. Instant answer for me. One of the first musical compositions I adored all because of a marvelous high school Music teacher who taught … wait for it … Music Appreciation!! FYI: I'll be 78 in January. The memory still lingers."
I enjoy working on grids with ML. She sent me this one after a rejection — there had been too many unsavory short gluey words — and asked if I could help her redo it. I liked the skeleton of 15s, NO DAY AT THE BEACH feeling particularly relevant, so off we went.
I've commented before on how tough themelesses can be when you fix a large skeleton of interlocking answers into place, and this one was no different. We actually had to tear the entire thing down to the bare studs, move a whole bunch of black squares around, and try dozens of arrangements to find one that gave us even a modicum of flexibility. It was very important to me to get more sparkly answers into the grid than just those 15s, and that's tough to do — there's virtually no place in the grid that allows much freedom to breathe.
We settled upon a few long slots in the NW / SE corners, and although I do like RAN A RISK, ANOMALIES, NO-GO AREAS, and NOBLESSE pretty well, they didn't sizzle as much as I wanted. But working more colorful answers into all four slots would have meant using big globs of crossword glue, and there already was some present due to the skeleton's trickiness. Always the trade-offs.
A similar story for the other two corners, with END IN A TIE and TRANSLATE not as stellar as I would have liked, but they allowed for clean(-ish) corners. Also, they seemed like they could be ripe for really clever clues, which is a trick Patrick Berry often uses to spice up the less colorful long entries in his ultraclean grids. If we could have come up with a sizzling clue for TRANSLATE for example, I think it could have flipped the neutral entry into something memorable. My offering — [Turn chat into cat, say] — is an attempt to mislead away from the fact that "chat" is the French word for "cat." Looking back on it, I don't think it really works though. Drat.
What a bear of a construction — all those Js + the constraint of every black square chunk having to be J-shaped = BLAAARGH! Thankfully, the engineer in me loved how well the construction process lent itself to a methodical approach:
From there, it was a matter of systematically testing each of 3x2x2x2 =24 possibilities. There was something so satisfying about keeping a master list of 24 possibilities, putting checks or Xs by each one as I drilled down to find potential problems with each.
(Man, I'm weird.)
I actually got very far — a full grid — down one path, and I thought it could be fine. But after letting it breathe, I took another look and felt like it just wasn't NYT-worthy — too many ugly bits, and not enough colorful fill.
It took some convincing to really try the layout that you see in the finished puzzle, because I was sure that isolating the first and last Js in the NW / SE corners would be the way to go. (Shows what I know!) With this final layout, I happened to get lucky by figuring out a good option rather quickly in the NE, filling acceptably around that J.
The SW … I constructed something I liked, but I did wonder if Will was going to like it as much. It contained HAPTIC, I FROZE, and AB TONER, all entries I dug. Will, though, wondered if any were common enough to be acceptable, and all three in that one region felt like too much. Even though I did a lot of HAPTICs in engineering and my dad has a (dusty) AB TONER and I love I FROZE as a stage fright line, I could see where he was coming from.
Redoing that little corner was rough. I churned out four options, each with some trade-offs, before Jon had the clever idea of paring the grid back even further than I had considered. I wasn't a fan of AS A SET — seemed like a partial to me — but I agreed that it was better than using something like OENONE, the woman Paris left for Helen.
I love mini-themes in themeless puzzles — they can be such a nice little treat. The crossing ALL HALLOWS EVE and LORD VOLDEMORT did make for a constructing challenge, though, sort of breaking up the grid into four quadrants. We tried to equalize the quadrants, leaving mostly 8-letter slots all around, for ease of filling.
The SW and NE corners weren't so bad, especially after ML suggested the lovely LE FIGARO to fill the tough ??????RO slot we were working with. I put in LIFE HACK in the symmetrical slot, and both of those corners fell pretty quickly. We considered a few different options for the SW, since LENGTHS is a bit dry, but we both loved one of our first tries at the NE, with PIANO BAR / IDIOLECT / TED TALK and even ARMPIT / AD REP colorful pieces of language.
The opposite corners were much harder. They wouldn't have been, if we had broken up LASER BEAM and I WONT DO IT into two words apiece (with a black square at the B of LASER BEAM and the N of I WONT DO IT), but I just hated to lose that long slot.
My stubbornness led to all sorts of issues in terms of clean and colorful fill, and I debated greatly whether all those cheater squares (the black square in the very lower left, the one under ALL HALLOWS EVE, and the one after MINETA) was just too much. In the end, we decided that the visual effect wasn't too badly hurt, and it allowed us to retain LASER BEAM and I WONT DO IT. (Talk about WHOA WHOA and HARD SELL!)
As we always do, Ellen and I brainstormed for weeks for this one, eventually landing at a spot very different than where we started. The concept of OVER and UNDER phrases, sitting literally over or under their target thing, is a little straightforward for my taste, but Will reminded me last year that a huge chunk of his solvers tend to prefer fairly straightforward themes (read: ones that don't mystify them).
That comment caught me completely off-guard back then. Personally, the NYT crossword first grabbed my attention when someone showed me all the crazy, creative ideas constructors incorporated. But as I thought more and more about Will's comment, I started recalling some of the feedback I've gotten over the years. I've heard from a good number of solvers who finished some of my puzzles — without ever understanding what was going on!
It serves as a healthy reminder that I sometimes try to be too clever for my own good. When more than just a handful of solvers don't figure out what the theme is, that's on me, not on them.
Here's hoping that this one walks the fine line between cleverness and allowing the solver to emerge victorious, with a warm glow of satisfaction.
Everything came together so nicely for this low word-count grid ... except for an unfortunate ??N?V pattern at the left side of the grid. ANNIV came immediately to mind, but I didn't care at all for that abbreviation. I did some brainstorming and wondered if there had been a famous queen, Anne the Fifth (ANNE V)? Came as a complete surprise to me that there's a very popular model named ANNE V.
Should I have gone with ANNIV? I debated this one for days. ANNIV felt so ugly, and the more I read up on ANNE V, the more crossworthy she seemed, having risen to the highest levels of the modeling world. And it's kind of cool to have that bizarre ANNEV string, not knowing how to parse it. I had a feeling she would be a toughie for many, so I tried my best to make the crossings very easy, so she wouldn't hold up people's ability to achieve a correct solve.
Still hard to figure out if I made the right decision. I'm sure there will be some who disagree, but I'm equally sure that there would have been grumbling (probably more so) about ANNIV. I could have also gone with just four themers, but that seemed too thin. Given that it's a pretty straightforward theme type, I felt that having five themers would help give the puzzle a meatier feel.
Always the trade-offs!
Jim came up with a theme idea Will liked, but he had trouble with the construction, going through three grid submissions with none of them accepted. Producing a basic Sunday 140-worder is hard enough, and to spice it up with a few strong entries, while keeping it relatively free of ugly gluey bits, is a formidable task.
So Will put Jim and me in touch. I liked the theme pretty well — GOLDING DIGGER tickled me — and I have a hard time resisting a grid-building challenge. With Jim's offer of a collaboration, I dove in.
Even after many years of making Sunday puzzles now, they rarely come together easily. Nothing different with this one. It took about 20 iterations before finally sensing that this fortuitously interlocking arrangement (SQUARE ROOTING crossing FASTING FORWARD) might be a good direction forward.
Even though I've worked with Will for years now, I'm always learning more about his preferences. I have picked up that one of his least favorite type of entry is the little-known acronym, because these are impossible for some solvers to figure out. For example, there's no way to convey the real meaning of MMPORG, is there? (Massively multi-player online role-playing game.) So it shouldn't have come as a surprise when Will kicked back the grid we finally submitted, which depended on OBO ("or best offer," in classified ads). My spidey sense actually did tingle a little when I put it in — I should have listened.
OBO was unfortunately right in the middle of the grid, so Jim tasked himself with redoing the middle section. It was a very tough region, so he ended up with a few things like RELOST, which I wasn't fond of. The final product you see today was the result of yet another redo.
I still like OBO better than OSE (which replaced it), but I think Will has a point — OSE is something people can figure out based on sucrose, fructose, etc.
Never easy, these Sunday 140-worders.
Okay, so this one was backbreaking. I was really glad to have Kathy as a collaborator speaking her mind, as I was ready to make pretty big compromises at many points during the construction process. Coming up with the various individual elements wasn't that hard — we both really liked MIND CANDY hiding the GRAND CANYON, and OLD FAITHFUL going up through YELLOWSTONE, etc. — but putting all the pieces together in a 140-word grid made for a rough go. It was even tougher when we decided to locate all the parks roughly where they are on a US MAP.
Okay, ARCHES is a little off. Ahem. We had ACADIA there for a while, but we decided having both ACADIA and DENALI was duplicative. (Plus, ACADIA isn't actually the name of the mountain.)
Some of the odd stuff I tried to pass off as acceptable, before Kathy reined me in:
I really enjoyed working with Kathy on this. My threshold of what's an acceptable trade-off tends to dip after I get 50 or more grid skeletons in, so it was great to have her kindly telling me at many a point that what I proposed as reasonable trade-offs were actually (cough cough) not.
I admit it, sometimes on these really complicated projects, I stick my head in the sand and go into Denali. Er, denial.
It's a lot of fun living near puzzle people. The ideas Parker and I pass back and forth don't always go anywhere, but when he showed me how beautifully TEMPERAMENTALLY broke into TEMPE RAMEN TALLY, it felt like the seed of a nice idea. However, after a quick scan through my list of 15-letter entries, I couldn't find anything that worked nearly as nicely.
Both of us are recreational programmers (me much hackier than Parker, who actually works in the field). It occurred to me that his TEMPE/RAMEN/TALLY seed idea might be a fun project to expand my coding skills, and after a few attempts, I figured out a way to pare down my list into a subset that might be reasonably human-scannable. Et voila, the other two popped out! (After about 10 hours of line-by-line searching.) HYPER CRITIC ALLY was also a possibility, but it wasn't nearly as nice, as it didn't change the "critic" meaning at all.
To the credit of Parker's programming skills, he came up with these ... and a bunch of others (mentioned above). Shows you what I know!
I thought THE SPLITS was so clever, such a neat way to sum up this idea! Parker wasn't convinced, but he reluctantly agreed to go along. We went back and forth with the tough construction, a mirror symmetry 15/15/15/9, and came up with something we really liked. It had so many great pieces of fill like I, ASIMOV, CAT TOYS, SO SORRY, BARTEND, etc.!
Will and Joel liked the general idea ... but felt like THE SPLITS wasn't apt. I wrote to Parker, asking what sort of idiotic ninny would come up with such a ridiculous revealer. (Again, shows you what I know.)
So, it was a complete redo, but we saw it as an opportunity to create something we liked even more — a 15/15/15 with a themeless word count. Not easy to do, but it was a fun challenge to see how many colorful long answers we could work into the grid while keeping the gluey short answers to a minimum. I always like to feature long fill, so hopefully things like Superman getting a SPRAY TAN in a BLIND ALLEY behind the DAILY PLANET while drinking a WET MARTINI enhanced the quality of your solve.
It's been an amusing game, waiting to see which July 4th this would be published on. We intended it to be hard, for July 4th, 2013 (a Thursday). Then July 4th, 2014 was a Friday. And July 4th, 2015 was a Saturday. Stupid calendar!
When we pinged Will about possibly running it this year, he confessed he was struggling with how to word the revealer. In classic wordplay, FOURTH OF JULY hints only at the letter Y, since it's the fourth letter of JULY, so many solvers might get baffled by the theme concept. If it were to be run on a Monday, we'd have to make sure a bigger chunk of solvers figured it out.
We ran by several ideas — our favorite for a crystal-clear clue was [The starts of 20-, 26-, 36-, and 42-Across are all one-___], but that admittedly felt inelegant. So I like the compromise with the [… literally] hint. Hope that makes it clear to everyone that J U L Y are each literally 1/4 (one fourth) of JULY!
Tough grid to assemble, with much less flexibility than usual. In working up the skeleton, I pulled out every trick in the book, including shifting each themer back and forth, changing the spacing, trying stacking / interlock, etc. I'm not a fan of the big L blocks of black squares on the sides, but it felt like the least of all evils.
Then it was a matter of choosing some good long fill — Jill liked MOM JEANS over LEE JEANS and DIONYSUS / TS ELIOT over many other options, even though the latter pair required an ORU to make it work.
I did sneak in a NYUK — the two of us have radically different opinions when it comes to the Three Stooges. I suppose she had to have some flaw.
Will sends some newbie constructors my way, usually when they have a decent idea but can't provide him with a satisfactory grid. I like going back and forth with the person as we trade ideas on how to build a nice final product. And sometimes I learn something, which I absolutely love. Nothing like unexpectedly picking up a little tidbit.
That was the case in working with Pris. Not only did I really enjoy working with her on this fun theme, but she challenged me on a couple of pieces of long fill I suggested, giving me rationale on why something else might be better. The best example: HAVE PITY. I had HATERADE in that slot at first, feeling good about introducing a newish piece of slang (haters drink the haterade, a play on Gatorade). Pris tactfully mentioned that she worried about it being a negative term, and she preferred to keep her crosswords upbeat and positive.
What a great reminder! Sometimes I get too focused on incorporating entries that are fresh and new, at the expense of the main goal: providing people with an enjoyable and uplifting solving experience. Many thanks to Pris for helping make a course adjustment as we filled this grid.
For any Sunday 140-word puzzle, I'm not happy unless I work in at least four long bonus entries. And for a wide-open grid like today's, I really want at least 10 snazzy bonus entries — given that there are only five theme entries, it's important to me to give solvers a lot of pick-me-ups to keep up their interest. There was something amusing about balancing a UNICYCLE over BAD DATES, and I liked the variety of OLD MASTER to LAWYERS UP to HOVERCAR to TSA AGENT to HOT PANTS. Something for everyone.
We tried a couple of fresh entries in the upper right, my favorite being OTTER PUPS. But we couldn't get that to work to our liking, and although OUT OF TUNE wasn't as squee-worthy, it seemed like it would lend itself to a fun clue. It also allowed us to incorporate CREED. If you haven't seen it, Michael B. Jordan is amazing ... and surprisingly, so is Stallone!
I liked working with this grid pattern last time, so I tweaked and tweaked and tweaked it to figure out how I might be able to wring something different out of it. Even though I used quite a few cheater squares (those stairsteps of black squares), it's still rough to work with a low word count (62).
Every time I got the middle to work, I'd launch into the four corners, only to find that one or more (sometimes all four!) were just not tenable. Took me a few dozen starts and stops before I finally hit on this middle region with JAZZ AGE / SAWZALL. I got three of the corners to work out, but then I had a sinking feeling the NE corner was going to bonk — that W in TWEEDLES was giving me fits. The crossing of SHINOLA/ALEK is my least favorite part of the grid, but ultimately, I thought solvers who didn't know ALEK Wek would most likely guess ALEK instead of ELEK, ILEK, OLEK, etc.
Other random notes:
I have one more themeless in the queue using this pattern, but I retired it after that. It's good to move on to something different.
Maybe something that doesn't cause my wife to ask me what I'm doing up at 4 a.m. for the tenth night in a row.
Took us a while to find enough good examples of the [Double ___?] style — we continued brainstorming for a long, long time until we finally had enough to cull a symmetrical set from. I really like how thorough Ellen is in all of her processes, and I've learned from her rigor.
It's fun as a constructor to push myself, so we tried to work all five themers into a low word-count grid, focusing on getting as much colorful long material in as possible. Entries like TRES BIEN and PEN NAMES and BAD DREAM are fun for me as a solver to uncover, so we placed high value on working a lot of those in.
I made a boo-boo in our first pass of the grid skeleton, Ellen asking me who HENLE was at 7-D. I wondered, who in the world hasn't heard of Don Henle? Turns out that's Don Henley. Drat! So we ended up with a little more crossword glue than I like in that north region — DE LA, BRYN (can only be clued one way), and IS SO — but we felt that overall, the assets in fill outweighed the liabilities.
PEN NAMES was fun to clue, as I love the story of J.K. Rowling writing a new series under a new pen name, Robert Galbraith. "The Cuckoo's Calling" is a well-written hard-boiled detective novel with compelling characters, yet it only sold something like a thousand copies. She eventually got outed, and now it's a major series, each new chapter eagerly anticipated by millions.
Okay, as an aspiring and unknown writer, maybe I hate that story.
What you see is a complete redo of the grid Will accepted a few years ago. Alex and I could only make the original work with the awkward A KNEE ("take a knee" to end a football game), and although it bugged the heck out of me, it was the best we could do in that region.
Several years passed, and I got to thinking about different grid structures, new ways of laying out theme-dense puzzles. This one came to mind, and it seemed like an interesting challenge — could we redo it for better color and cleanliness? After working through a few possible skeletons, we both liked what was going on here. It unfortunately required OPE to hold it together, and it's rarely good to start out the filling process with something gluey right from the get-go. But it looked so promising otherwise that we decided to go ahead with it.
We did end up with one region, the west, where we couldn't get around something like A HIT or A HAT. But we felt that overall, the price of those minor gluey bits was well worth an open, 72-word grid, with bonus material like LOVE HOTEL, TATTOO ART, TURN TAIL, etc.
Some clue notes:
These days, I've been spending a lot of time figuring out ways to work in more and more bonus long fill into puzzles. It's important to me to always give the solver more than they expect, and with the bar being raised higher and higher by strong constructors, this is not an easy task.
It used to be enough for me to shoot for six good bonus entries, but I've been experimenting with eight or even ten. This can really strain a grid, and trade-offs crop up often. As with today's grid, I find that using a bunch of long downs staggered through the grid, each one intersecting no more than two themers (but ideally, just one), can often work wonders.
Once I fixed R O O T in place — Will and Joel's second choice for a revealer — I filled around that corner as cleanly as possible, which led me to BLACKLIST as one of the only good possibilities at 34-Down. That reduced my flexibility at 4-Down, but after trying a bunch of things, THE MAFIA seemed to work well. Basically, I repeated this process from left to right in the grid, stopping and restarting dozens of times when I was forced into trade-offs I thought solvers would groan at.
I wasn't happy with SLEEPS IN, as it feels more neutral than an asset, but having something great in that slot meant a couple of gluey bits holding it in the grid. Always the trade-offs — I pity the fool who has to make them.
(I so badly wanted the full A TEAM intro, but my original clue might have been too long for Will: ["In 1972, a crack commando unit was sent to prison by a military court for a crime they didn't commit. These men promptly escaped from a maximum security stockade to the Los Angeles underground. Today, still wanted by the government, they survive as soldiers of fortune. If you have a problem, if no one else can help, and if you can find them, maybe you can hire... the ___."] Ah, well.)
This puzzle construction gave me fits. It was hard enough to identify eight themers that would work in the way I wanted — circling around a single black square like a storm, before continuing on its merry way — but working with the max 140-word constraint made things tough. Each of the eight themers took up very short slots, meaning that I had to work in more long fill than usual. And having those single letters stick out — the H and S of IS THIS THING ON, for example — constrained the grid in tough ways. Throw in CORIOLIS FORCE for a revealer and the fact that I wanted the central black squares to look like a storm ...
The bottom left corner alone took me maybe 25-30 reboots. I was hesitant to use MUTTONY, which I needed to make that corner work, but I thought of a fun clue for it: [A little sheepish?]. Sadly, it was left on the cutting room floor.
Will worried that BANH MI wasn't going to be familiar to many solvers, and the crossing with DEBI might feel unfair. (In Seattle, you can throw a loaf of crusty bread 100 yards in any direction and hit a BANH MI shop, but his reasoning was sound.) I tried a few other options, but in the end, Joel helped us settle on the original, given the BANH MI's rising prominence in the food world and its utterly delicious combination of roasted pork and julienned vegetables.
All of a sudden, I'm hungry.
Will also felt that the puzzle as I had it — no arrows — would be just too hard. I drew up some storm-looking things, but Will decided to use arrows instead. I still like the idea of "storms" within the puzzle, spinning in opposite directions in the Northern hemisphere vs. the Southern ... I wish my meager art skills had better conveyed that. Maybe then, my original title of "SPIN CONTROL" would have flown. Ah well.
I enjoy working with mirror symmetry, which we needed for this theme (lengths of 9, 13, and 11 make regular symmetry impossible). Not only is it an interesting challenge given its idiosyncrasies, but you usually have to place a horizontal line of three (or more) blocks across the middle column. Put in a few stairstep squares, and it's almost hard to avoid giving your puzzle a smile.
We spent a lot of time trying to figure out how to give the solver bonus fill, adding to their fun. Usually, I like to work in at least six longer entries, stuff like HEALTHY GLOW, BOB AND WEAVE, and MINUS SIGN. The AMERICANO is my drink of choice at Starbucks, so that was fun too. I sure would have liked for us to work in one more pair of long answers, but that just strained the grid too much with little gluey bits everywhere. And what's most important — by far — is how much fun the solver has. Ah well, ICE BOX, AIRBOAT, PC CLONE, and (my childhood favorite) ASTERIX will have to do.
ML's great idea, making candy cane shapes out of black squares. And then when she paired CANDY CANE with ST NICHOLAS (who's tied to CANDY CANEs by legend), it seemed nearly perfect. If we could only figure out what might fit in those "unchecked" squares in the crooks of the candy canes …
The engineer in me laid out the X M A S in left to right, top to bottom sequential order (the M to the west, the A to the east, and S in the south). It should be logical and orderly, dammit! ML very tactfully pointed out that this looked weird. Being weird, I took this as a compliment.
Thankfully, ML managed to convince me that a clockwise ordering of X M A S made much more sense.
We started with a 64-word grid (the block between AL HIRT and RCA removed). I like a challenge, and going down that low — along with the puzzle already being constrained by the candy cane blocks and the pair of mini-themers — seemed like a fun adventure.
After laying out many skeletons and testing to see which would give us the most flexibility, we barreled into the east section. We had the OPERA???? pattern in place pretty early, and we tried many different things (BUFF, COAT, DIVA, FANS, GOER, HATS, SOLO, STAR, etc.) before OPERA CAPE lent itself to snappy fill. MSRP / OCT / SEP weren't great, but getting SMART CAR / NO-LOOK PASS (I was a huge Magic Johnson fan) / IT'S A KEEPER / HALF-COCKED fill seemed well worth it.
The other three corners proved more challenging. It was only after weeks of work that we finally gave in and added the black squares to turn the 64-worder into a 66-worder. This allowed us to get in a lot of nice SEA MONSTER / KGB SPIES / AD AGENCY / TELENOVELA ("Jane the Virgin" fans, anyone?) / FREE RIDE stuff without much glue, but man oh man do I hate giving up. Anyway, who cares how many words there are — what's most important is how much entertainment the solver has.
Merry X A S M! Or whatever. Weirdo.
Sam Donaldson is the man. Excuse me — The Man. (Whichever is the good one.) We finalized the last revisions for this puzzle during the 2015 ACPT, where Sam provided no end of entertainment with his crazy bets. 70-1 odds that he could pick seven names, one of which would win the ACPT? Sure, he'll put $10 on that! (He won.) 300-1 odds that Lynn Lempel didn't write Puzzle #5? Heck yeah, he put $10 on that too! (He won that as well.)
He was able to go back home and tell his wife that he won an easy $20. But man oh man, was he sweating it when 1.) Kiran Kedlaya showed up, finishing fourth and 2.) Puzzles 1-4 were NOT written by Lynn Lempel. I would have paid Will pretty much any sum of money for him to announce that Lynn wrote Puzzle #5.
Writing a puzzle with Sam is almost as fun as watching him squirm. Good times, good times.
P.S. Grid was pretty hard — trying to figure out how to squeeze in all those long, turning answers was a bear. But once I hit on SHOW ME THE MONEY intersecting PAID THROUGH THE NOSE, it was only a matter of (many dozens of) hours after that.
P.P.S. I've fixed up the answers so they appear in the database correctly — SNOOKERC to SNOOKERCUE and CUE to RIGHTONCUE, e.g. Hope everyone figured out that the "RIGHT ON, RIGHT ON!" title was a good hint!
ML brought me in on this one late in the game. After I declined to work on it (I've sworn off word ladders for the most part), she showed me Will's message to her (see above). She knows me too well — there's nothing I like more than trying to pull off something that might not be possible.
So off I went, trying out skeleton arrangement after arrangement. Working with four long themers plus another six short words proved to be quite a challenge, but I just hate giving up.
I finally came up with something I liked, something that looked like the word ladder "flowed" from FOOL to SAGE. I wasn't a fan of the "jump" from the right to the left side, but I did like how TALL and TALE naturally served to link the two halves. I sent the rough skeleton back to ML, and she filled it in with a lot of nice stuff.
It's important to me to include at least a little bonus fill in every puzzle I work on, so I'm glad MONOLITH (huge "2001: A Space Odyssey" fan), SO SUE ME, and SEX SELLS made their way in. Even though I've become a bit snobby about candy these days, I do like me a Mr. GOODBAR every once in a while.
Both of us are big "Despicable Me" fans, so I'm glad ML pushed for GRU to get his 15 minutes, despite my hesitations.
As always, a pleasure working with ML!
I pondered the question of single words (GRANDILOQUENTLY) vs. multi-word phrases (SOCIAL BUTTERFLY) for a long time. On one hand, it's sort of neat to see such long words with exactly one instance of each of the vowels (AND Y of course). On the other, I'm not super fond of single-word themers as a solver.
There's something so cool about multi-word entries. Not only do they help distinguish "good crosswords" from "bad crosswords" — or worse yet, from computer-generated ones — but there's so much potential for snazzy phrases. It took a while to come up with all-vowel-plus-Y ones such as THE DYNAMIC DUO and THING OF BEAUTY, but Dan and I agreed that these types of answers would make for a more fun experience. We were glad that Will agreed.
Dan and I also discussed packing in five themers, one for each of the five vowels, but that proved to be too restrictive, if we wanted an orderly presentation of the five circled letters. I didn't totally balk at the original layouts Dan sent over with random circle placements, but I always appreciate layers of elegance in a puzzle if they're possible. We tossed around putting the five vowels on a diagonal leading to the final Y, but this layout resulted in smoother fill.
This type of grid skeleton, featuring six long downs, is something I use all the time. Spacing out these long guys and separating them with adequate black squares often makes for a surprisingly easy filling process. I'm not a fan of APO, but to get ARM CANDY and AW GEE — along with locating the O in the proper spot — felt worth it. I encourage constructors to try out similar skeletons, shifting black squares liberally.
Dan is patient, hard-working, and he took constructive criticism extremely well, always thinking about what would make the best experience for solvers. A pleasure to work with.
It usually takes 20+ ideas to come up with a good one. To Jason's credit, he kept at it, never giving up as we bounced around concepts. Finally, his idea of famous people "wearing" their trademark hat amused both of us.
We wanted an even gender split, but we couldn't find many females with a trademark hat (Jackie Kennedy, couldn't you have worn a hat without an "X" in it?). Luckily Jason thought of CALAMITY JANE, so we kicked off the puzzle with her in the upper left. I usually try not to use REATA, but 1.) that corner was rough to fill, and 2.) it's amusing to have the STETSON atop her head and a REATA twirling above.
Speaking of grid-filling issues, yikes. We felt seven people would be minimum, otherwise solvers might find their experience thin. But it took 50ish skeletons, shifting people and hats around, before one finally looked feasible. (UNKEYED comes from my music background — apologies if it's not familiar to you.)
But in that grid, Will pointed out that LNG (liquid natural gas) would be rough for solvers. He was absolutely right. Even though we crossed those letters fairly, LNG was something I looked the other way for, because that single entry made the skeleton possible.
You'd think I'd learn after all this time.
So we rebooted. After dozens more iterations, I threw in the towel and suggested that we go up to 142 words. We were still able to incorporate a lot of long slots for snazzy fill, so thankfully Will gave the thumbs-up. Jason worked in a lot of good stuff like WHAT A GUY and TV DAD.
I'll fess up to OSA, an entry I really dislike. But it allowed for EMOJI, DATA SET, and LEFT JABS — three goodies for the price of an OSA. I still cringe at seeing it, but hopefully solvers will find the trade-off favorable.
Finally, we had to use much more glue than I like in a crossword, but with much more than average thematic material, plus the constraint of seven pairs of entries having to be right atop each other, having some was unavoidable.
Fun to work with Jason!
I owe Joe Krozel for all the good advice he's given me over the years. Hearing from him and reading commentary he's written for his older puzzles has been invaluable.
So much fun to work with Ellen. She came up with the idea to have a giant black hole in the middle of a grid; sort of a HOLE faux-rebus. I thought it was clever, so we went to work trying to find enough "___ HOLE" phrases for the north and west, and "HOLE ___" phrases for the east and south.
After banging my head for ages, I came to the conclusion that it just wasn't possible. There were relatively few "HOLE ___" phrases, and to group them into triplets such that all the crosses worked proved infuriating. Luckily, Ellen came up with another brilliant idea: all the answers adjacent to the black hole really ought to be "sucked inside," i.e. reversed. Not only was this a cool visual representation of what happens with a black hole, but it allowed us to take advantage of the huge arsenal of "___ HOLE" phrases we found.
Then it was a matter of adding theme density. I thought it would be amusing to have funny phrases around the perimeter like DISAPPEARING ACT, but I worried that there wouldn't be enough apt choices. Ellen came back a few days later with a huge list, most of them of the "wish I had thought of that" variety. (Sensing a pattern here?) Turns out that the hardest part was picking the best ones from the list! Ellen's a hard-working, creative person.
Our original title made us laugh: ONE SINGULAR SENSATION. To all the die-hard physicists out there, I realize SINGULAR isn't exactly the same thing as a SINGULARITY, but we thought the wordplay was too good to pass up.
Our original clue for BLACK (HOLE) was [Giant sucker in the middle of the puzzle].
We amuse ourselves; hope you were amused, too. And if you weren't, I have a few other "___ HOLE" phrases on the list ready for assignment. (WORMhole. Why, what were you thinking?)
I started searching for themers by hand, working with favorable letter patterns to turn out entries like SCHOOLED = SHOE + COLD. But that seemed like a silly way to approach the problem. Why not use all the modern tools at one's disposal?
Thanks to Bob Klahn, who spurred me on to pick up programming again, I was able to futz out a Python script which helped me to turn up many more options. I was delighted to find so many fine two-word phrases in particular. And when ALTERNATION turned up, it sealed the deal.
But how to work in ALTERNATION to the cool EVEN / ODDS revealer? Luckily I had CELLO SUITES as a symmetrical entry to ALTERNATION, but having those two as my first and last themers and EVEN / ODDS in the center only allowed for two additional themers, for a total of five.
Or did it? I normally dislike overstuffing of themage, since it tends to cause ugly or even egregious fill, but I figured I'd at least see what was possible with seven themers. I experimented with a few dozen combinations, and this stacking pattern with BLUETITS over FREE MEAL over ALTERNATION gave me enough of what I wanted: flexibility in fill plus the ability to work in some good medium-length fill. The grid doesn't allow for as much juicy long entries as I usually like, but I had fun monkeying with all the seven-letter slots.
And my favorite entry of the grid? It amuses me to no end to imagine speed solvers stalling out at the kooky-looking EINK [Noise from a scared pig?]. I'm a huge fan of Kindles and e-readers in general, so I was happy to work in E-INK.
Apologies to the non-physics ubergeeks out there if the ENTROPY / MESONS combination stumped you. One of my alternate grids also contained Z BOSONS — probably a good thing I (just barely) managed to restrain myself.
Seven theme answers which "turn up," i.e. they become complete by reversing back into themselves. GLUTEN FREE B becomes GLUTEN FREE B(EER) when you hit B and bounce back upward, thus completing the entry. NOW WHERE WER becomes NOW WHERE WER(E WE), DO EXACTLY AS I (SAY), etc.
A relatively easy construction (as far as 140-word Sundays go), but finding theme answers was tricky. That is, until I dove back into the world of computer programming, spurred on by Bob Klahn and the needs of our CrosSynergy syndicate. It had been a few years since my last experience with Pascal and C++, so getting started again was quite a struggle. Turns out technology is vastly different from 15 years ago, dang it!
Wrenching myself away from my beloved arrays in favor of "lists" and "dictionaries" proved a bit puzzling, but now that I'm climbing my way up the learning curve, I'm finding Python to be more intuitive, akin to working on a Mac vs. a PC. It's pretty cool to be able to write code to find "all words whose last three letters reversed match the previous three letters" or similar such query.
All in all though, programming is just another tool, in the same way that a Crossword Compiler, or Excel, or even a computer is a tool. Finding that FREE B(EER) and ONE TO (TEN) fit my criteria was a good start, but FREEB and ONETO were much too short to use as themers. Good thing a close friend of mine is on a quest to find decent GLUTEN FREE B(EER)! He and I have done many taste tests, and ON A SCALE FROM ONE TO (TEN), most of them rate about a negative four hundred and twenty nine.
Ellen approached me with this idea, and I thought it had a lot of promise. I wasn't at all sure that we could 1.) find enough "___ FIRE" and "FIRE ___" entries, and 2.) arrange them in pairs so that the surrounding fill wouldn't be compromised. Especially difficult given that we wanted the "holes" to not be connected to any other black squares.
If there's anything that drives me, it's a challenge. So I enjoyed working with Ellen to figure out if this was feasible. A key factor was using varying lengths — not just all ???? FIRE and FIRE ???? entries, for example — so that we had a large enough pool of words to choose from. Then it was "just" a matter of trying out hundreds of combinations and black square arrangements to make it work.
Spending so many of our black squares around those two "holes" meant that the NW and SE had to be giant, wide-open spaces if we were to keep to the 78 word max. Fun challenge, and not too difficult if you're okay with a few pieces of Thursday-ish toughies.
Speaking of that, we originally targeted a Thursday-ish puzzle, but Will thought it would be neat to run it as a Monday, replete with a fire visual in each hole. Nice idea, and well worth taking a jackhammer to the NW and SE sections to redo them completely so they'd be Monday-easy. We could have increased the word count past 78 to make it a whole lot easier on ourselves, but 1.) I like a challenge and 2.) I'm stubbornly moronic. Not easy to fill those huge chunks of space without relying on non-Monday bits of crossword glue.
I was glad that Will ended up changing it back to a Thursday, where I think it really belongs.
Ellen has such patience with me. I think it's important to brainstorm dozens, perhaps hundreds of ideas in order to come up with something I feel is really NYT-worthy. We went back and forth and back and forth, trying to come up with a few good Sunday ideas, and each time, a promising idea came up just shy, or we decided that it just wasn't good enough. Weeks, maybe months of headbanging.
But when Ellen proposed this little idea in a short email, I felt sure there was something to it. I worried we wouldn't be able to come up with enough instances that fit the pattern perfectly, but Ellen came back shortly with a promising list. Turns out a computer science background and some query skills comes in handy! I added a few more, we finalized the selection, and off we went.
The grid design was incredibly challenging, and it looked like it was going to be a bear to fill, given all the crossing constraints. We originally tried using standard crossword symmetry, but all the theme answers branching downward made that very difficult. (It might have been possible if we didn't have the theme answers in symmetrical locations, but that felt inelegant.) Mirror symmetry made the layout much more feasible, and also let us place the themers (at least their first halves) in spots of symmetry.
But after putting together a draft skeleton, we weren't sure it was going to fly, given all the crossing constraints. I took a tenuous pass at filling one tiny little section to see if the entire thing was feasible, and after maybe ten hours and many hundreds of hairs pulled out, I emerged victorious. Ka-ching!
Then Ellen politely mentioned that I had entered IS IT REAL, not IS IT LIVE. Into the circular file.
Several dozen iterations later, passing the file back and forth, going cross-eyed with frustration, we emerged from the long, dark tunnel. At revision 43j, it's not the most work I've put into a single puzzle, but it's pretty close.
A really fun time working with Ellen; hopefully the solving experience amuses people!
The grid to the right is what happened when I tried to do this one on my own. (I felt there had to be a progression from BOGEY to PAR to BIRDIE to EAGLE.) I didn't really like the themers. The fill. The layout. The solving experience. The author. But I proudly showed it to Jill.
"Hmm," was her reaction.
She's very considerate of my feelings. Canadian-esque, I might say.
Thankfully, I had the wisdom (dare I say, genius) to bring her on board so we could tear it down and create something a whole lot better. Sometimes I wonder how my wife's brain works. Normally when someone says "I need a seventeen-letter phrase incorporating the word EAGLE," you don't expect people to respond. Ever. (Then again, people often ignore me. For good reason, I'll admit.) But Jill blurted out THE EAGLE HAS LANDED within roughly 16 microseconds. I did some puffing out of my chest and blurting out things like "well, that one's kind of obvious, isn't it?" But I deigned to go with it. So generous am I.
The gridwork was a bear. Stacking PAR above or below each of the themers proved incredibly challenging. There was no intention to purposefully have POOL PARTY and WATER PARK be such long answers — short answers like PART and PARDON failed miserably in roughly a thousand-billion different ways, none of them quite working. Finding this particular arrangement was just a stroke of luck, one of those things where if you try a thousand different things, one might actually work.
Hope you enjoy! (And for those still confused by the theme, EAGLE, BIRDIE, and BOGEY are golf terms for two under par, one under par, and one over par.)
(Will is in D.C. for the North American Teams Table Tennis Championships, so I (Jeff) thought I'd relate Will's acceptance letter:)
"The MacArthur Foundation missed yet again this year in its 'genius grant' awards, overlooking your obvious once-in-a-generation — nay, mystical — abilities. It is a true wonder that people do not bow reverently to you as you walk down sidewalks — or may I more accurately say, veritably hover with your angelic beatitude. People will surely recount with misty-eyed fondness where they were the day they finished this puzzle; when they acquired a magical moment in time, surpassing even the births of their offspring."
(Or something like that.)
ML pointed out how the word COGNOSCENTI (or its singular COGNOSCENTE) was both interesting and appropriate to crossword people, and as usual, we were off to the races. After much iteration, we were able to come up with a center triple-stack all (loosely) related to "The Cask of Amontillado," one of my favorite Poe short stories. (Fortunato is a snooty COGNOSCENTE of Amontillado, and Montresor exploits that hubris in his plan for revenge.) When we realized Halloween would be on a Friday this year — ripe for an appropriate mini-theme — it felt like kismet.
We actually finished a first version with more blocks and more words (see left), which surprisingly turned out to be a little harder to fill well. Unusual for that to happen, but sometimes the letter patterns do tricky things. I was a little skeptical of SATE SAUCE, but a quick check with some of my Malaysian friends came back with the result that SATAY SAUCE in fact was the "incorrect" one (along with some appropriately snarky comments about how they wouldn't eat our Americanized Asian foods if they had been BURIED ALIVE and it was the only option). I was all ready to settle for the alternate grid, with its more choked-off grid (the NW flows into the rest of the puzzle so nicely in the original, and feels comparatively sectioned off in the alternate), so it was a pleasant surprise to be able to go with the more wide-open grid.
As an aside, Sanford and Son is one of my favorite shows of all time. If I had my druthers, that theme song would be playing in the background 24/7, and all grids would be filled with AUNT ESTHER swinging her purse at FRED SANFORD's face while spouting off her Bible verses, GRADY WILSON shuffling along with his wisecracks, and LAMONT's buddy ROLLO taking the brunt of Fred's ridiculous one-liners. Heck, even the horribly stereotypical AH CHEW makes me laugh. The entire show is so politically incorrect, it gives me an appreciation for how far things have come in 40 years.
I'm usually one for uplifting entries, grids and clues that make the solver finish with a sense of happiness, but sometimes exceptions keep things interesting. (Full disclosure, I had to sleep with the lights on after finishing "Harry Potter and the Goblet of Fire.") Now, who's up for a nice glass of Amontillado? Just follow me downstairs...
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.
A delight working with Tracy. She's creative, thoughtful about entry selection, skilled, and most importantly, willing to go the extra mile to create the best experience possible for the solver. It was her idea to place the individual N E W S squares in the grid, which I thought was a beautiful added touch to the center compass.
Filling this bad boy was no walk in the park, and I was impressed at how well Tracy's first attempts went in various quadrants. Sometimes people get lazy, calling a handful of glue-y bits good enough, so I was happy to see Tracy's willingness to go back and forth, trying again and again to improve the quality of long entries or to eliminate even just one ugly piece of fill.
It's too bad Across Lite can't handle these types of graphics, or handle odd rebus-like squares. (Sorry for those of you struggling to figure out how to enter that compass rose into each of the eight special squares.) I'm glad that the NYT tech team is taking on the challenge, improving their online solver app bit by bit. I met one person on their team (Scott Koenig) at the ACPT this year, and was impressed at his drive to to give the online solver the best possible experience, accommodating all the wacky things we constructors do. I'm looking forward to seeing how the app handles this rule-breaker.
I get a kick out of collaborating with people who are thoughtful, kind, and willing to work their butts off to create the best possible solver experience. Feel free to drop me a line if you're like Tracy! FYI, these days I focus mostly on Sunday-size puzzles, since they're a great need for Will, Rich (Norris) and Patti (Varol).
So sad that my original clue for HAIRLIKE didn't make it in: [Quality of a quality rug?].
Today on XWord Info: a colossal clash, a fiery fight, it's...
MAN VS. MACHINE!
In the red corner, we have one of the puzzle's constructors. The Brawn from Taiwan, the Absurd Nerd of Crossword, he'll punch in those squares (along with a bunch of swears). Representing the human race… Jeff "THE BIG BOSSMAN" Chen (if I say if often enough, maybe it'll stick). Since he hasn't seen this puzzle in over three years, it seemed only fair to allow him to study and memorize the answer grid for 24 hours before solving. Fair is fair, after all.
In the blue corner: the upstart challenger, an artificial intelligence program which has worked its way into the top 100 finishers at the American Crossword Puzzle Tournament. Armed with a database of answers, clues, and puzzles from throughout the ages, it employs an arsenal of pattern matching algorithms stemming from the mind of the brilliant but nefarious Matt Ginsberg. Introducing… Dr. Fill!
Who will win in this epic engagement? Will the human race come out swinging against its cunning creation? Or will the progeny punch out its pugnacious producer, making men mewl "Mercy!" to their mechanical masters?
Jeff's time: 3 minutes, 3 seconds.
Dr. Fill's time: 5 seconds flat.
Thankfully for humanity (collectively shaking their heads at my slowness), Dr. Fill came up with a few errors. Matt mentioned that his program was unable to see that SPACE SPACE SPACE SPACE SPACE SPACE SPACE was an actual word. I declare victory on the basis of nonsensicality! Cue the music:
Seriously though, I'm fascinated by Matt's efforts toward building crossword AI, and by man's attempts to keep one step ahead. I'm practicing communicating in binary code for the near future. Beep boop.
It's time for another edition of JEFF VS. DAN! For those of you who don't know of this magnificent contest, I race Dan Feyer, the reigning ACPT champ, at my own puzzles. In the past he's stomped me into dust, often by a factor of two (a factor of three, one time). So after a while, what could I do but up my level of cheating, going one step above knowing the answers? By studying the entire answer grid for five minutes and memorizing as many answers I could, I proceeded to trounce Dan (by two magnificent seconds).
So I decided to be the bigger man and give "the crossword titan" (note the theatrical air quotes) a handicap. I'm eventually going to beat Dan fair and square (by only knowing the answers — more than fair, really).
My time on today's puzzle: a blazing 3:56! My hand shall be victoriously raised!
Dan's time, listed at his blog... 1:49. Dagnabit! He beat me by a factor of two. The sun was in my eyes. And the clouds. And these grapes are awfully sour. Ahem.
BTW, fun to work with Dick on this one. We had a lot of back and forth, brainstorming and revising Dick's core idea. The grid came pretty quickly after finalizing the themers, as 15/14/14/15 is a pretty easy pattern to work with. And I enjoy the exercise of fitting in good short fill like ONE TWO, NO PROB, BAD RAP, R KELLY. BTW, I actually don't know who R KELLY is. I mean, I know the name. But if you put Eminem, Psy, Lil' Wayne, Yo-yo Ma and R Kelly together, you could easily convince me that they were The Village People. R Kelly is the construction worker, isn't he?
It's a real pleasure to work with Mary Lou. One aspect I really like about her is that she understands how many theme ideas it takes to yield a single half-decent one. Too often when people approach me with a concept, we find out it's been done before, or it's too loosey-goosey, whatever, and they don't continue to brainstorm. Not ML! I find it takes at least ten thoughts to produce one quasi-workable theme, and even then, development of said theme takes time (and may not work out in the end). I'm glad she sticks through the arduous process.
This puzzle was a bear to put together, and we went through many iterations. People might say left-right (or "mirror") symmetry is becoming my shtick, and I do admit to liking the visual appeal. Most often I need to incorporate blocks in the center of the puzzle, which by nature either take on a smile or a frown, and who doesn't like seeing a nice friendly face right in the middle of their grid? (Don't answer that.)
Theme development alone on this one was very tricky. ML came up with a long list of workable phrase, and we wanted to narrow it down to a consistent set of at least four. Even then, the puzzle didn't feel heavy enough until we came across the idea of adding THINK / THROUGH, a double hint to the themers.
And then the grid work. Tortuous! Every time you work with crossing answers, you heavily constrain your grid. One set of (fixed) crossing answers is easy-peasy. Two is no bother. Three gets a little tricky. Four... groan. You might think, why do those goofballs need so many black squares up there? You'll answer you own question if you try to position those four sets of crossing themers within a 15x grid (and tear your hair out in the process).
Then to the bottom section. I wasn't wild about sectioning the grid in two (the edges of the smile sort of split the grid) but we didn't have many options for layout in those side regions. Keeping the four themers as separate as possible helped us fill relatively smoothly, so it was a trade-off we were willing to make. We did consider splitting some of the entries like CHESS GAME and OUTHOUSES to get rid of uglies like ETE, ASA, EST (ick!) but it surprisingly turned out to only reduce the ick factor slightly. So we decided to accept a tiny bit more ugly stuff to incorporate those nice I AM SO DEAD type answers. Honestly, it's too many glue-type entries in general, but I thought the overall concept was neat enough to be okay with it.
Finally, two notes interesting to me. First, I'm sure there will be people who gripe about DUROC because it's a word they don't know. But why not look it up and learn something? There's a giant swath through the Midwest who would likely argue with you cotton-pickin' city-slickers (pretty sure that's what Midwesterners say). Second, check out the "cross" made out of black squares at the top center. Typically I don't like to do this as it makes the grid feel too "filled-in" with black squares, but in this case, I thought it was a nice echo to the shape of the themers (highlighted in blue/red below).
David was kind enough to put up with me, a slightly demented 18-year old freshman roommate interested in all sorts of bizarre experiments. As an example, I set up trials to determine how long I could go without sleeping. (Hallucinating began at hour 45 and I fell asleep in the dorm hallway at hour 50.) Needless to say, people drew mustaches on me.
This grid was challenging, since the high theme density made it near impossible to incorporate long downs. I used to be fine having only a little snazzy fill in a puzzle, but these days I hate letting a puzzle go unless there are at least four nice non-theme entries inside the grid. As a solver, I highly value sparkling fill, so I always keep that front and center in my mind when I construct.
Since using long downs wasn't possible, I had to incorporate long across fill. Typically that would bother me because those answers might be mistaken for theme material, but in this case I didn't mind since the themers were so long (15/11/15/11/15) that they stand out on their own.
Unfortunately, with all the constraints, the only option was to put long across fill in the NE and SW corners. I typically don't like splitting rows 1 and 15 into two words apiece, because three words apiece is so much easier to fill cleanly. But in this case, it was the only way I felt we could add some sparkle.
I tried several hundred alternatives for each corner, and I'm never happy to have a partial (I LAY), but I thought it was a reasonable trade-off to get such goodness as 'NUFF SAID and ASIAN FLU (I find the pan-Asian financial crisis of the late 1990's incredibly interesting and useful to study, in hopes of preventing future outbreaks). I was also happy to give LAILA ALI some props, as well as Darren SHAN, the author of the Cirque du Freak series. I can only hope to someday have one-tenth his success as an author.
Fun to collaborate. I'm particularly interested in increasing diversity within constructors, so please contact me (jeffchen1972 at gmail dot com) if you'd like to break into the most fascinating hobby in the world. (And you're willing to work dozens of hours with little pay!)
Genius, simply pure genius. What else can I say about my "JEFF vs. DAN" contest... or should I say, my "JEFF vs. DAN" con? (rubbing hands deviously together)
For those of you not familiar with JEFF vs. DAN, I speed-solve my own puzzles against five-time American Crossword Puzzle Tournament champion Dan Feyer. It's now been years of sandbagging, pretending like Dan's been beating me by a long shot, whining about how I'll never catch him, even though I know all the answers in advance. And it finally paid off! For I have pulled off the heist of the century.
Danny Ocean? A mental midget compared to myself when it comes to pure devious plotting. (Vizzini? Moron.) See what I have achieved: I finished today's puzzle in -0:43 (yes, that's negative forty-three seconds), clearly besting Dan for the first time ever, using my trade-secret combination of dual-handed solving, Asimovian psychohistory, and warp-induction time travel. Call the Guinness Book people, folks!
And the coup de grace: by beating Dan on this most important of days, I won the ACPT 2013 title from him, based on the contract clearly implied by the JEFF vs. DAN contest. So sayeth my esteemed barrister.
I WIN, FEYER, BWA HA HA!
P.S. My actual time, having studied all the answers before clearing the puzzle and solving it on computer as fast as possible: 1:59.
And now for our next installment of JEFF VS. DAN, where Jeff feebly attempts to solve his own puzzles faster than ACPT champion Dan Feyer. Seeing as how Dan spanked me silly in our last round, I decided to give myself a significant advantage: not only did I look at the answers beforehand, but I STUDIED them. I tried to memorize as many as I could so I'd be able to fill answers in just based on letter patterns. My time was a personal best: 2:25. Check back in to see what times Dan and the other speedsters clock in at.
ADDED NOTE: Drat it! David Plotkin, aka Bug Out, checked in at 1:57... ON PAPER. That's almost 30 seconds faster than me, using a slower solving medium (I solve on computer). Sigh.
Matthew contacted me with this idea, and I thought it was hilarious, and insider's nod to the crosswordese that often must be used to glue a puzzle together. And to have all common E words that were all "four-letter words" just tickled me. We went to work trying to figure out appropriate definitions, and given my distaste for entries that sound like they're from a dictionary (I just don't find them fun to uncover in a crossword), we batted around many phrases until we arrived at ones we felt like we'd be fine with seeing in a normal xw.
The fill was especially challenging, especially because if we incorporated a single piece of crosswordese (outside of the themers), the puzzle just wouldn't seem elegant. Not to mention, we'd leave ourselves open to all sorts of chop-busting from our crossword friends. So it took a long, long time figure out an arrangement of those five themers which would allow us to produce a relatively clean fill. I don't like that we had to leave in REE or ANAT or the singular TAPA, but almost every crossword with five themers is bound to have something.
Now, if we could have just fit in FOUR LETTER WORDS...
And now for another installment of "JEFF VS. DAN", where I speed-solve against ACPT champ Dan Feyer on my own puzzles. Because he's spanked me so badly recently, I studied the grid extensively just before solving this time, trying to memorize every single entry. My time: 3:00 even. And that's with me typing in an utter frenzy! Check back into see what times Dan and the other speedsters post at Dan's blog.
Very fun to work with Matthew on this one. He was so pleasant in writing, and equally pleasant when I gave him some feedback. Theme ideas are hard to come by, and often times it's a real strength to know when to let something go. Very few people can do that, moving on to brainstorm further (I typically go through 20ish theme ideas before settling in on something that moves me).
This one was a toughie! We wanted to incorporate WASH DRY PRESS FOLD in that order, use LAUNDRY LIST as a revealer, and have each of the four words in snazzy phrases where the word had a different meaning. Not easy at all. As Matthew mentioned, we originally had a different grid, and Will gave us thoughtful feedback with his rejection. After nodding my head (read: swearing up a storm and perhaps making a mustachioed voodoo doll; I neither confirm nor deny this), we went back to the drawing board. Good thing we did, because Will's intervention forced us to dig harder, and we eventually came up with these themers, which we liked much better than the original set.
One aspect I'll point out in the gridwork: incorporating five long themers is tough, and it becomes even tougher when your middle entry is a "weird" length (13, 11, or 9 letters), which sort of splits the grid in half. Solvers might have minor gripes today, saying we could have cleaned up NEH (yup, ugh!) and SWED (double ugh!) by breaking up OF SORTS and PREPARE with black squares. We tried that, but look where POLIO sits. Turns out that there were few entries that could have worked there, and I just despise OLIO. Personal tick of mine. So we deemed it better to include the nice OF SORTS and exclude the ugsome OLIO. Would you rather have NEH or OLIO? Matter of taste, methinks.
Matthew and I have another collaboration waiting in the wings — he's two for two on submissions to the NYT! A whole lot better than my overall acceptance rate, roughly 33%. And that's a whole lot better than my original acceptance rate, which was roughly 0% (0 for 22). That was a long two+ years (with a very patient editor)...
Thus continues the saga of JEFF VS. DAN, wherein I try my durndest to beat ACPT champ Dan Feyer on a puzzle I wrote. My solving time: 6:39. You'll find Dan's time (and that of the other speedsters) posted at Dan's blog. Curious to see how badly he beat me this time. It's been over 30 puzzles now, and I still haven't won. I've even started handicapping it by reviewing all the answers right before I solve. And sometimes I don't even come within a factor of two!
ADDED NOTE: David Plotkin's time (on paper!) = 5:54. Dagnabit! And Dan's time... 3:50. Whoa. Talk about superhuman powers. Someone ought to make him a cape.
Very fun collaborating with Dick on his big debut. He came up with the idea and I helped him refine it. I actually didn't do that much; the credit should go to him. It's tough to come up with clever ideas for a NYT Sunday puzzle.
Ah, the gridwork. Not as much of a challenge as I've encountered in some other constructions, but a toughie nonetheless. There are only six themers, but incorporating the dagger with the M A C B E T H letters made it tricky, forcing us to deploy our black squares quicker than we had desired. This left big open spaces in the NW and NE, and without those cheater squares (the two stair patterns at the top of the grid), it looked pretty hopeless. We debated whether EXERCISERS was legitimate, eventually deciding it was KOSHER, and were glad that it allowed the snazzy BANDOLERO and LITERATI to work.
The other tricky spot sort of snuck up on us. Because of the word count limitations (140 max), we had to incorporate a few long across answers, IONE SKYE notably. It's a nice entry in itself, but boy did it cause problems. Because it crossed three theme answers, there weren't many options there besides IONE SKYE, and once that was fixed into place, KIEL was the best we could do at that spot. Now, I'm a huge fan of James Bond villains and Jaws in particular, but even I don't like having to remember who played him. And having OHO, OOH, OH I, UH OH... that's my fault. Sometimes at a certain point (in this case, maybe 50 attempts), you've done your best.
Ah well, as with most constructions, there will be compromises. I had a great time working on this with Dick, and I welcome anyone who's interested in collaborating (read: allowing me to ride on their coattails as I politely shoot down all their ideas until something fun emerges) to contact me at jeffchen1972 (at) gmail (dot) com.
THE REAL STORY!
Wanting another Sunday crossword, I summoned nefarious powers from the darkened netherworld (also known as Canada). To my horror, the dirt and clay outside my house trembled, clods bursting toward the sky. A horrendous form pushed through the surface, two arms of congealed earth atop a trunk of elephantine proportion. I had called forth a golem, a powerful being from Hebrew legend!
Luckily, I had just brushed up on my golem lore (if you haven't read Helene Wecker's "The Golem and the Jinni" please do, it's fantastic). I rushed outside with a scrap of paper with which to animate the creature and wrote "Make me a memorable Sunday puzzle, something a cut above the rest." Jamming it into its mouth, I stepped back with trepidation as the clay monster creaked to life. Golems, after all, have a tendency to go mad and start crushing things.
I followed the great hulk-beast as it tromped down the street to the local playfield. Scratching my head, I watched as it clobbered its fists into the grass, the earth shaking with each of its strikes. After a few minutes, I yelled, "What in blazes are you doing? CUT! Cut? Cut ..." My jaw dropped as I realized it was pounding the word C U T at the top of a gigantic grid. Genius — a literal C U T above the rest! As it finished its task, its eyes flashed a terra cotta red and it ran into the darkness, roaring about how it would smash out all life on earth. That wasn't so good, but its puzzle was.
So there you have it; that's how a Sunday crossword is made. And if you see a golem rampaging in the streets, I don't know anything about it.
Loren was so much fun to meet last year at the ACPT. Such a nice person, fun to hang out with, and a great sense of humor. But she's too humble — it was another puzzle she put together which led to this one. I thought it was extremely well constructed, but I also felt like it was lacking something, a certain je ne sais quoi or a fuller raison d'etre. When we tossed around SNAKES ON A PLANE, it felt like everything snapped into place. I love it when a plan comes together (cue The A-Team theme song).
I always look at the good and the "could have been improved" in a puzzle, and there's a lot I really like about this puzzle. But there's also no doubt that the heavy constraints forced some compromises. Neither Loren nor I were fans of all the short stuff at the top (ISPS, RIAA, IRR etc.), but at the time it seemed like a reasonable price to pay for including SIOUX WAR and SANDBAG (and the symmetrically placed KLAXONS and SEMINARY). We looked at possibly breaking up some of the long stuff to get better short fill, but the compromises still were there so we ran with this. Tough trade-offs.
And to those grousing about the BIKO/ORIBI crossing, I raise my hand on that one. Loren was keen to get something more friendly there, but I used the Jedi mind trick to convince her otherwise. So if you faltered at that spot, please look closely into my eyes and accept my apologies ... (waving hand hypnotically in front of your face) and go get me a scotch.
Looking forward to more from Loren! Oh, one point I forgot: happy 65th bday to Samuel L. Jackson! We hoped this would run next month on his actual 65th birthday (Dec. 21), but what are you gonna do.
ADDED NOTE: Time for another installment of Can I Beat Dan At My Own Puzzle? My solving time, just having looked at the answers = 3:27. Will Dan Feyer and the other speedsters beat me?* We'll add Loren's time here too as soon as the puzzle publishes.
*Dan's time: 2:17. I need ideas for a bigger handicap. Blindfolding him?
This puzzle was two years in the making, with umpteen thousand versions and three back-and-forths with Will. My first version didn't have the perimeter words interlocked, which Will thought inelegant. I'm really glad he pointed this out. Even though it took me ages to figure out a set of words which did interlock in all four corners, I like the elegance of it.
Now that I've had some time to review the puzzle with fresh eyes, I would have loved to incorporate AROUND in a symmetrical way. See how it sticks out, no symmetrical theme answer balancing it? If I were to do it all over again, I might place the letters A R O U N D in a tight circle around the center of the puzzle, with each letter circled. A notepad note could explain things, or better yet, it would be devilish to make the solver connect the dots, figuring out that the circled letters were a subtle hint to those perimeter answers.
Fun to see how one can learn and take away lessons for the future.