Kevin G. Der author page

This grid is actually a revision. The top right was the same, as CHEAT DAY / ROGUE ONE / ALAN ALDA was too good to let go. But ZOE KAZAN used to be in the bottom left, which created a corner filled with so many proper nouns and difficult words (e.g. ENFILADE) that I was asked to redo it. So ZOE KAZAN moved up to the middle area - I simply could not pass on those Z's - and finding a quad stack that connected with it was very time consuming. Discovering that delicious combination of LIMEADES / MAGRITTE / PLAYTEST, the last of which could pass as a seed entry, was pure luck. Overall, my notes grew to over 250 partial fills that were good enough to consider submitting, an indicator of how stubbornly I clung to my seed entries.

For the clues, I decided to forgo a self-imposed deadline for finishing them, and instead simply selected a few unclued entries at random each day to clue and only finalized something if it was really good. Only about a third of them ended up being changed whereas in the past it's been half or more, which is encouraging. For instance, I'm pleased with 26-A, 1-D, and 32-D. One category of clues that tended to be changed were tricky clues for entries that are less well-known. For example, I had "Fast break?" for CHEAT DAY, which would be quite difficult for a person who's never heard of the term; the same is true of "They're related to hobbies" for KESTRELS. To me these are fair for, say, a tournament final, but it seems that they're too hard even for a Saturday.

It's unclear to me how difficult it will be for solvers to grasp the theme, but I think two things will help: the clue for 19-A pins the answer fairly directly, and each section has answers going down or across in the normal direction which serve as toeholds. To make these toeholds more gettable, the clues are quite straightforward for a Saturday puzzle, and otherwise I suspect the puzzle would be near impossible to complete. Some of the more oblique clues were changed in the interest of solvability — for example, "Carnival offerings" for STATEROOMS and "One imagining a better future for herself" for DREAMER. Once the theme is revealed, keeping track of which rows and columns are reversed and entering those words backwards adds a twist to the typical themeless experience.

According to my notes, this puzzle sets my personal record for length in the queue — it was accepted in the summer of 2014 at the same time as the puzzle that ran on October 15, 2015. I think it's typical for creators of all types to look back on old work and wish for a lot of changes. That's the case here. In terms of grid design, I would have preferred having four theme entries going in each of the possible directions, but the pair of length 13 makes this difficult because they necessarily intersect the other pair. The 13's also necessitate a quad stack in the corners which can be mitigated with three sections, but which results in a high word count. I'm also not sure why I chose stacks of seven in each corner rather than eight and six, which I suspect could have resulted in a livelier grid.

I like ultra-tough Saturdays, so I tend to overshoot the difficulty of the clues. This grid has many difficult entries, so some clues should be slightly easier than average. Together, these resulted in several clue changes. There's always room for clue improvement, and analyzing editorial changes can be instructive.

There are many ways to increase clue difficulty including:

• misdirection (the clue appears to have one specific meaning, but actually means something else)
• vagueness (the clue could mean a lot of different things)
• open-endedness (the clue reads one way, but there are several possible answers)
• trivia (the clue involves obscure knowledge)

Of those, Will and his team seem to favor misdirection while minimizing vagueness. Trivia is minimized to an extent if it's not intrinsic to the entry, though I enjoy trivia-based clues if they're interesting.

SPILL BLOOD: "Act the gladiator, perhaps" was changed to "Engage in warfare." I was hoping to avoid outright mention of killing. But I guess that original was a bit open-ended and too difficult.

ANTI-CRIME: "Like some urban initiatives" was changed to "Describing the efforts of Batman and Spider-man" which is much better. It's more interesting, and the new clue is less open-ended.

ACCREDIT: "Recognize, in a way" was changed to "Give a stamp of approval." The original was very open-ended, though I feel like that's the main context where I hear the word, as in an accredited university. I understand the change in the context of easing up on the difficulty.

LARIAT: "What a catcher might throw" was changed to "Rodeo ring?" This is an example of a clue that was too vague; it's too big a mental leap from "catcher" to "a person with a lariat."

EL MARIACHI: "Titular film character who carries a guitar case full of guns" was changed to "1992 western with a Spanish title". The original is interesting (and recalls an iconic film scene), but since this is a difficult entry, reducing the trivia helps solvers.

HOT CEREAL: "One getting mushy in the morning" was changed to "Oatmeal, e.g.," The original has misdirection and seems reasonable, but maybe seems forced. Another possible ding is that it uses the "it" style of cluing. Nothing pains me more than definition-based clues for long entries that could have been interesting. (My runner-up pet peeve is reusing clever clues, which I'll check aren't in the database beforehand.) But it's not reasonable to expect the editors to replace all my duds with zingers, so I only have myself to blame.

Analyzing clues that made it through is a helpful counterpoint: e.g., CALDER CUP, EMPANELS, B-MOVIE, TEA TASTER.

I submitted this puzzle about a year and half ago, and fortunately it ended up fitting into this special week. I tend to like the rebus puzzles that have a reason for the rebus squares to exist, and creating dense shapes in the grid with rebus squares was something I hadn't seen before. A spiral seemed like a visually interesting choice that contrasted with typical crossword symmetry, and it would require a rebus density that was much higher than usual.

The original submission didn't shade the rebus squares, only the black squares that were part of the spiral's path, with a note telling the solver to look for a shape indicated by the theme entries. I slightly doubted that the puzzle was solvable that way because the gimmick is so unexpected, but I just figured Will would make it easier if he thought the same. I think moving the puzzle from its original Saturday spot to a Thursday, and shading the spiral, makes it tractable. Hopefully there's still a challenge in figuring out the theme and that the disruption in pattern matching makes for a unique solving experience. (I'm not afraid that the puzzle is too easy because a test solver already let me know that it was "sadistic." My goal isn't sadism, it's merely to provide a challenge!)

It seemed necessary to include theme entries representing spiral shapes, and I chose ones that spanned a variety of physical scales, as well ones manmade versus found in nature. Unusually, the theme entries are symmetric but their lengths are not, due to the spiral shape. Overall, there were dozens of partially filled grids before arriving at this final result.

KEVIN:

Working with collaborators lately has been great fun. Ian's puzzle sense and high standards with skill to match are helpful assets when it comes to co-writing a themeless. We had originally guessed this grid would be run on a Friday.

Will's notes about the clue editing are fascinating. Receiving feedback this detailed is a somewhat uncommon opportunity, so for at least constructors it's just gold. When it comes to writing difficult clues, I've found it can be hard to grasp the difference between misdirection and vagueness, and a number of the comments indeed relate to clues deemed overly vague.

IAN:

Kevin is a rare triple threat in the game: fill, clues, themes. He's equally skilled with all of them. That's his awesome 1A answer/clue combo. So naturally it was a blast to collaborate.

Several marquee entries got scooped before this was published, so apologies if you've already seen entries like PALEO DIET and IPAD MINI in grids. Also, glad to see my clues for 30D (GEL) and 56A (HATE MAIL) made the cut!

Hope solvers like it!

I am a perfectionist. My mind fixates on anything suboptimal. When constructing a puzzle, this habit can be helpful as hundreds of judgements are made — whether a partial is worth it, are there too many plurals in the grid, etc. And since a submission can be rejected for a single flaw — e.g. one entry is deemed too obscure — diligence is needed. The trap is that the mind can go too far in this direction, dwelling on the smallest imperfection while forgetting the merits. I once spent days agonizing over a themeless with a single entry I disliked — FDR — before I came to my senses.

The company I work at is researching methods for peer feedback. We learned the greatest factor in whether a team is thriving and successful is the ratio of positive to negative feedback. The best teams used 5x as many positive comments to negative — while average teams only used 2x. Sometimes the human brain works in unintuitive ways. Since then, I've tried to apply the 5x to rule to myself when constructing. It's helped me remain objective about imperfections while still remembering the merits. I've found myself happier and more motivated to construct. As is commonly said, perfection is impossible to achieve.

Onto the puzzle. I became interested in 4-stacks while observing Patrick Berry's puzzles (even his 5-stacks are smooth) and wondered if mere mortals could do it. Began at the top right corner with 6-A and 16-A as fresh entries. 18-A is a plural not ending in S, which adds some nice misdirection. 11-D was a serendipitous discovery for BLACKO??. And I liked some other lively entries in the grid. But I wish I could change KIEL. While not totally obscure — it hosted Olympics events as recently as 1972 — many people won't know it. I personally didn't. For a painful time I considered scrapping the entire corner because of it. Now I feel the grid as a whole is good enough to accept it. Hopefully, solvers will get the crossings and be satisfied by learning a word. Finally, I noticed that Will kept many of my favorite clues, but also changed some to make the puzzle easier. It's helpful to know my instinct is writing clues that are a bit too difficult. Overall, I'd like to improve several entries, but the result is okay and I hope the next is better, even though it will be imperfect.

ADDED NOTE: This puzzle was used in the 2014 MIT Mystery Hunt.

So many good constructors are creating puzzles today that it's hard to find an original idea. Whenever I think of a theme that might work, I add it to a list on my phone, which I'll check later when I feel like constructing. In this case, I think I saw a bumper sticker on the street that began with "HONK IF YOU LOVE...", which had clear Monday possibilities.

In the grid, the theme entry lengths (12's and 13's) forced a certain pattern because I prefer to avoid two cheater squares for 13's. The result is two long vertical entries that pass through three theme entries, which drastically reduces the possibilities. (I now see that I could've just flipped the top and bottom pairs instead. No idea why I didn't.) The two vertical entries I found are consequently ok but not amazing. I tried to compensate by increasing the number of 7- and 8-letter entries. Finally, I try to avoid using 3 letter words that cross the grid's center, because it feels suffocating.