That Puzzle Guy's Blog See that book on the right, there? You should buy it.


The War on Fill

We're coming up on the hundredth anniversary of the crossword puzzle, and it's amazing how it's changed in that time. These changes are largely seen as positive steps towards enhancing the enjoyment of crosswords in modern times. If two-letter words hadn't been banned, we'd likely be sick to death of closed-off corners and many, many bigrams that would appear therein. Rotational symmetry lends a welcome aesthetic quality to the puzzle. The list goes on. To an extent, we should expect this evolution to continue. However, I've noticed a trend in some of today's puzzles that I believe is detrimental to the fun of crosswords and the art of their construction. In keeping with the manufactured outrage displayed on 24-hour news channels, I've dubbed it the War on Fill.

The "fill" simply refers to the words in the grid, typically excluding theme answers unless it's one of those themes that applies to the entire puzzle. Because it's at least the vast majority of the answers in a crossword, the bulk of the solving time is spent on them. Without the fill, simply put, you don't have a crossword.

I understand that a theme is the most memorable part of a puzzle and frequently provides the biggest aha moment. I agree that it needs to be good to consider the puzzle good. But the fill, being a more pervasive element, can impact your solving experience a lot more. You can mentally shove aside a weak theme and focus on everything else, but there's no getting around writing in all those shorter answers.

Thus, I hold that the fill is the most important part of a quality crossword. So I must ask: Why does good fill seem to be decreasingly prioritized in today's puzzles?

All regular solvers have seen their share of unsavory fill. The cheap partials (OF IT, IN NO, etc.), the abbreviations nobody ever uses in real life, the tired crosswordese, the boring obscurities. Some are merely little warts that can actually be very easy to figure out, while others can frustratingly keep the solver from full success. Naturally, some of these poorer entries are inevitable; the rules of American-style crossword construction are challenging and sometimes require concessions. (This is another reason some prefer cryptic crosswords, in which only about half of the letters need be crossed.) Plus, one's mileage may vary with so-called obscurities. ("How can it be obscure? I know it!") Even taking these factors into account, the fill in some recent puzzles has been, in my view, lackluster.

I don't mean to suggest that every puzzle I've ever made has been resplendent, nor that fill standards are gone completely, but I do think they've eroded quite a bit. It's possible that I, as an inveterate solver and occasional constructor, have become overly sensitive to this matter. I find it hard to believe, however, that the solver wouldn't enjoy the experience more if some of the junk were replaced, even if it were with ordinary, everyday words that aren't particularly peppy. You can always bring the liveliness in the clues.

Is it laziness? It does feel that way sometimes; on occasion, a corner seems so blatantly refillable with superior entries that one has to wonder why the constructor settled for his/her version. Even so, I don't think indolence is a fair charge to throw around; there's no real evidence to suggest it. (If you ARE a lazy constructor, though: Stop it. Stop it right now. Try harder.)

I believe the War on Fill is, at least in part, the result of how the crossword-solving community has evolved. Nowadays, several puzzles a day are available with just a few mouse clicks, and, through the blogosphere, solvers and constructors alike are deluged with opinions about all of them. All this content drives up the pressure on puzzlemakers to create something innovative. When you've seen your hundredth add-a-letter theme, it feels, and very likely is, less desirable for constructors to make and for editors to accept. Plus, it's natural to want to knock a tough crowd's socks off with something new.

As I implied above, people don't really remember a puzzle solely based on its fill words. The fill might be appreciated and enjoyed, but it's not what gets Puzzle of the Year nominations. An ingenious theme gets far more attention, and it's the starting point for making a crossword, so trying to make that stand out makes sense. I certainly sympathize; struggling to come up with good themes is a big reason why I haven't been constructing too much lately. The catch with themes, however, is that they still have to be molded to fit the crossword, and it can be very difficult to let go of a good one, no matter how it resists a smooth execution. This can lead to some strained shoehorning in the form of weaker fill.

It was a highly regarded member of the crossword community who said, "It used to be that if a theme didn't work, you wouldn't do it." It is very easy to slacken one's grid standards if one feels the theme is worth it. In my view, though, it usually isn't worth it. I've seen some themes that are, make no mistake, utterly brilliant, but the concessions made in the fill kept me from enjoying the puzzle. You might disagree and suggest that this is my problem, and I see your point. Obviously, I'm stating my opinion here. In the end, though, we're solving the crossword because we enjoy crosswords, and thoughts like "That's a thing?" and "Well, I guess that's right" lessen the joy of solving, no matter what genius lies in the theme.

I have even stronger opinions when it comes to themeless puzzles, in which innovation comes in the form of stacking fifteen-letter answers or using as few black squares as possible or what have you. I'll be blunt here: If I never see a quadruple-stack of fifteens again, it'll be too soon. Sure, it's eye-popping, but it's no fun at all to muddle through the inevitable handful of weak answers crossing the stack. My favorite American-style crossword is one that's tough as nails, well-filled, and fair. That's it. I don't need or want the grid to show off. I'll take a solidly filled 66- or 68-word puzzle over a sub-60-worder every single time with no hesitation.

So that's my take, for whatever piddling amount it's worth. I know I'll continue to put a lot of effort into turning out the best grids I can, and I hope my fellow constructors do likewise. Viva fill!

Comments (138) Trackbacks (3)
  1. HEAR, HEAR. The NYT has been particularly guilty of this of late. I’d rather solve a puzzle with a pretty good theme and lively fill than a puzzle with a never-been-done-before theme where the fill has been compromised in order to accommodate the theme. And there is simply no excuse for shitty fill in a themeless. The whole point of solving a themeless is to enjoy the fill. I don’t give a fat rat’s ass whether the puzzle has OMG THE FEWEST BLACK SQUARES EVER and I definitely don’t care how many 15s the constructor managed to stack.

  2. I dunno – if it’s a cool or original theme, I would much rather see it published with compromised fill (if it’s compromised due to whatever constraints, rather than laziness or lack of skill) than never see it at all. Of course, I’m still entertained by quadruple stacks, as long as the 15s aren’t stale. And I don’t mind a few crappy entries gluing a themeless grid together — if their number is proportional to the number of sparkly entries. Everybody has their own “line” where one outweighs the other…

    I definitely agree that Shortz and crew could do a better job cleaning up crappy corners*, but not sure I agree with the thesis that fill has been getting worse overall. Could there be a touch of “things were better in the good old days” nostalgia here? Then again, I tend to forget puzzles immediately after I’ve done them, so a poorly-filled puzzle doesn’t bother me for more than two minutes. (And when I see the byline of a repeat “offender” before starting, my expectations head downward.)

    *But should they? What about authorial intent? If every jagged edge were sanded off, as it were, the NYT puzzle would have a lot less personality.

    • I think constructors can have a distinct style that doesn’t involve warts. A lot of personal flair can come across in the clues, for instance.

      It should be stressed that bad fill has degrees, too. A weird abbreviation that nobody ever uses is significantly worse than, say, a four-letter partial clued with a common phrase.

    • > *But should they?

      Yes, they should.

      > What about authorial intent?

      I see a lot of crap corners that come about just because the author is trying to get QIX or JAY-Z or something in there; then you could make a case for authorial intent, though I think that if an editor prioritizes clean fill over rare letters they have the right to enforce that. But a lot of crap corners are just “this is the first thing that autofilled from the cruciverb database”. Those are easy to recognize, and editors shouldn’t hide behind “authorial intent” to excuse themselves from doing the work of fixing them.

    • I have only published at the LAT, but my limited experience has been that Rich Norris respects authorial intent without neglecting his editorial responsibilities. I can say for myself that my authorial intent is never to publish a bad puzzle, and so I’m grateful for the editing he has provided for my puzzles. (I’d say I’m less happy when he rejects a puzzle of mine… though 75% of the time I can see the point and probably 50% of the time – once some time has passed – I’m grateful they didn’t get published with my name on it!) Of course, rejections are usually based on the theme – but acceptances have usually been accompanied by a few grid changes, and in a couple of cases saving a puzzle with words I would not have found myself.

    • Thanks, Dan F.

      I get lots of laughs from crabby criticism, but am happy to have one or two “Ah, ha!” moments a week.

  3. I’m on Team Dan. Give me a cool theme or quad stacks over boring “perfect” fill anytime. We need to keep innovating.

    Focus on the five best answers in a grid, not the five worst. Did a batter with 2 Home Runs and 2 strikeouts have a great day or a lousy day?

    And go check puzzles from 10 years ago. The fill is much better than it was overall.

    • i dispute the notion that “perfect” fill is boring. good fill is better than bad fill; that much should be tautological. if you are constructing a puzzle and think to yourself, “the only way i’ll get the solver to remember this corner is to fill it with ridiculous obscurities”, well, you’ve gone off the deep end. fortunately i don’t think anybody does that on purpose, but a lack of effort often leads to the same result.

      also, if we are focusing on the best answers in a grid, why should we like a puzzle with quad stacks? they always have fewer lively answers (and many more crappy answers) than a “normal” themeless grid. if you are putting 15-letter answers into your grid, it should be because they are awesome 15s, not just because you can cram in a lot of 15s.

      • Indeed, the fill we praise most is the exact opposite of boring.

      • No you shouldn’t like a puzzle just because it has quad stacks. However, one shouldn’t bluntly dislike it for the same reason. It seems the same less-optimal fill in quad stacks is judged more harshly than when it appears in other puzzles.

        • I agree to an extent, but I think context should not be ignored. Four or five crappy answers look a lot worse when they run side-by-side crossing a quad-stack than when they’re spread about the grid to hold it together.

        • I think this is because that even if bad fill makes a puzzle no fun to solve, the theme can still be appreciated, to an extent. With a themeless, the fill is all there is.

      • Wholehearted agreement on the 15s issue. I believe I have seen only one that was very well done and at the same time minimized the crappy fill (I think it was a Jonesin’ from 2009, but don’t quote me).

        • i thought tim croce’s recent triple-stacked 15s (NYT, 4/13/12) were awesome. the crossings weren’t the cleanest i’ve ever seen for triple-stacks, but they were worth the trade-off, i think. less recently, mike nothnagel’s double-stacked 15s (NYT, 4/23/11) were fantastic and the crosses were flawless. much easier to do with only double-stacks, though. (not that i’d know from experience, never having attempted any kind of 15-stacking myself.)

          i’ve yet to see any quad-stacked 15s that have had me at all excited about the answers themselves.

  4. Hey, we’re all on the same team here… :) I do think that bloggers/commenters are right to call out subpar fill when there’s too much (again, everyone has their own level of tolerance), but I hate the obsession with pointing out every bit of dreck in an otherwise fine puzzle. And then docking it “points”! (Anybody else think that the “star system” is the worst thing to happen to crossword blogging since Ryan and Brian hung it up?)

    • ugh, sorry for the smileyface. ;) :p

    • I think criticism can definitely go overboard. I think fill could stand to be better, on average, but sometimes I also think we crossword critics are too difficult to please.

  5. Great article, Tyler. Everything you’re saying about construction is right on (though I haven’t been doing enough puzzles to agree/disagree with there being a trend.)

    If there is a trend, though – I wonder, at least for edited outlets, whether it’s an editing issue as much as or more than a constructor issue. (again – not singling anyone out here.) If puzzles with creaky fill are seeing the light of day, it’s because editors are putting them there. Surely, with so many talented puzzlemakers out there, there’s plenty of puzzles being submitted for publication – it strikes me that we’re looking at a demand problem (the demand being the editor, not the solver) if sturdier puzzles are being blocked out by the whiz-bang themes sitting on unstable fill foundations.

    • I imagine an innovative theme might come as a breath of fresh air to an editor who’s just sifted through a bucketload of the same old stuff. That could give a big boost to the puzzle’s chances, even if some ickier fill has to be forgiven.

  6. Interesting post. I don’t disagree with your observation — your solving regimen certainly qualifies you to note such trends in crossword construction. I’m not convinced that poor fill is the result of blogosphere-fueled stress that encourages constructors to attempt elaborate themes or record-breaking grid feats. Are you seeing more new names in crossword bylines? The problem could be an influx of novice constructors who don’t have the experience/resources to produce satisfactory fills. I’d enjoy seeing some kind of empirical analysis that would identify the likely causes, though I don’t know who would be willing to do it.

    • Well, that’s why we have editors. My first few puzzles were awful; they would rightly have been rejected had I submitted them anywhere. If what you say is the case, new constructors should be encouraged by helpful, constructive notes from the editor and other puzzlemakers rather than quick publication. I think puzzles should be judged without regard for who wrote them.

  7. I’m with Tyler; I don’t think one should lower the bar for fill just because of a challenging theme and/or grid. I do agree with Todd’s point that some of the poorer filled puzzles have also been by new constructors though.

  8. I wonder how much of this is caused by constructors autofilling their grids. Very few people fill by hand any more, which forces you to really think about every word you’re putting in. I’ve seen so many small isolated corners with terrible entries that I suspect this is at least partially the problem.

    Also, as you suggest, it’s largely a case of poor prioritizing by the constructors. If you’re willing to sacrifice the fill to get a quad stack, or some idiotic subgoal like a freaking pangram, your priorities are off. Look at the best of the best, like Patrick Berry: do you see him putting in crap to tie his puzzles together? Then why is it ok for others to do it?

    My longtime puzzle motto (which I try to live up to, though of course I’m sure I fall short at times): It’s All About The Fill. Theme makes a few puzzles memorable in the long term, but fill makes every puzzle good or bad in the short term.

    • Agreed on all counts. I hadn’t considered the autofill angle. Computers are invaluable tools, but there is a limit. In large part, I consider construction an art, and the idea of just letting a computer do it with no guidance simply seems wrong.

    • I got called recently by Rex/Michael for that poor prioritizing (trying to cram too fresh entries at the expense of surrounding fill). It was a fair cop, so I definitely learned something as an, I guess, journeyman constructor there…

    • I’m a little late to the discussion, but I think Trip said it best. You won’t go wrong if you ask yourself: WWPBD? We all fall a little short at times, but clean, interesting fill should be at the top of a constructor’s list of priorities.

  9. This topic comes up regularly in conversations with a friend about the NYT in particular, and crosswords in general. Quoting myself from recent days: “Good enough is the enemy of good,” and (pertaining to one particular grid where themers intersected unnecessarily, forcing especially crappy fill) “Just because you can do something doesn’t mean you should.”

  10. I might (and often do) theme-climax within the first few minutes of solving, but dammit, you better fill-cuddle me until I fall asleep.

  11. Count me in as one who’s basically uninterested in themes (and totally uninterested in pangrams). Most of the time on Mon-Wed NYTs (and pretty much all the time on Newsdays, LA Times, etc.), I don’t pay any attention to the theme while solving, and often don’t even go back to see what it was once I’m done – just not interested, and those themes are so rarely unusual. (And like Dan, I usually can’t remember most of a puzzle moments after I’m done with it, although in my case it’s probably the onset of senility!) I’d rather have great fill – but that can be in the clues, not in the fill words themselves. My dream solving world is an unending stream of themelesses from people like Berry, Shenk, Walden, and Quarfoot… to name a few.

    A theme has to be REALLY innovative for me to notice or remember it, and that rarely happens. (And when I say really innovative, Merl, I’m talking to you about “HEY! A DECIMAL!”, my favorite entry in what’s probably one of my favorite themed crosswords ever…) This is one of the main reasons I don’t construct, btw… I can’t imagine ever being anything other than mediocre at it, at best.

  12. I’m willing to put up with a little stale fill as long as the theme is well executed. ALER isn’t pretty, but it’s consistent, recognizable, and filled and forgotten in a second. It’s the theme that I remember. I’m much less tolerant of obscure fill, and, obviously, obscure crossings, but “obscure” is of course a nonexact term, defined as “something I don’t know”. Entries that Joon will complain about as obscure are well known to me, and I’m sure the opposite applies (as evidenced by his regularly kicking my butt on Tausig/Onion/BEQ puzzles :-).

    Also, I’ve been working through a recently discovered stack of old puzzles from the late 90′s/early 2000′s and I do notice more fill ugliness in those, for example manufactured TLA’s. So I agree that overall fill has seemed to improve, at least in my perception.

    Great article, Tyler, this has been a very interesting discussion. I enjoy seeing everyone’s perspective on this issue.

    • I’ll agree with you and Jeffrey that fill overall is still better than it was in the “old days.” Looking at a smaller window, however, I think the idea of reaching too far for wacky themes and eye-popping themelesses is pretty new. For instance, the first quad-stack in the NYT was pretty recent, right? And since then, people have tried to duplicate the feat or even put two quad-stacks in one puzzle, and fill suffers. It’s not worth it.

  13. Do you believe there are puzzles being rejected at the NYT which have solid fill (themeless or with solid themes) that are being displaced by puzzles that don’t? Or do you think there could be many better puzzles if only editors were to insist more effectively? (How much extra effort would have to be expended by submitters collectively so that you wouldn’t be complaining about low standards?) Or are there just too many puzzles for maintaining high standards?

    If you solve a thousand puzzles a year, you’d expect 25 of them to be two full standard deviations worse than _average_ (assuming approximately a normal distribution ….)

    When I see fill I don’t like in a puzzle, I don’t even bother to shrug my shoulders. The constructor is trying to provide amusement for very low compensation, and a bit of bad fill never bothers me for very long.

    • I don’t have much of an idea of the puzzle supply, so I’d be guessing there. On occasion, I’ve been frustrated that a puzzle of mine was rejected while certain other ones made it in, but, of course, I’m rather biased in those cases. Considering the length of time between acceptance and publication, though, I imagine it would be pretty hard to be too choosy.

      The compensation is an interesting factor I hadn’t considered. Would the quality go up if the pay rose? It might, but so too might quantity. The point is moot to a certain extent, as most editors don’t have much control over the pay.

  14. Thanks for this post, Tyler.

    I’ve been ranting on this subject for 25 years, long before Crossword Compiler, and I hope you and the post writers know where I stand: CRAP = REJECTION. PERIOD.

    It really is a big surprise to me that so much crap still gets into so many published puzzles, when CCW + a robust answer database has rendered it virtually unnecessary. While I am all for record-breakers and push-the-limit ideas that allow a few clinkers now and then (though I won’t publish these) , there really is no excuse for the continued prevalence of this stuff seemingly everywhere.

  15. I think autofill is a big problem, but I think a bigger problem is word lists (and also clue lists). By definition, if your list is composed of previously-used entries, then your puzzle will not have any fresh fill. And, if you choose the top choice every time in the manual fill, uh, you’re pretty much getting the autofill.

    One big reason I’m using CrossFire these days is that I can type in my own entries at any time while filling a puzzle, even if it’s not in my list yet or if CrossFire doesn’t think it’ll be fillable. The idea that the fill should come from a predetermined list, even a huge one, limits fill quality.

    • I think Crossword Compiler’s lists do provide a healthy number of phrases that haven’t appeared in a grid, though it requires diligence to throw in all the fresh stuff one encounters in the world. Efforts like Alex Boisvert’s Collaborative Word List Project and Todd McClary’s Autofill Project are important and beneficial endeavors in this regard.

      And I never start a themeless without some sort of marquee entry I haven’t seen before. So, if nothing else, I’ll have that.

  16. Regarding all of this talk about CC and autofill… you may be interested to know that when it comes to quad-stacks puzzles, much of the construction work is not done with computer.


    • I certainly buy that. I’m trying to focus more on the result rather than the process. Hell, if someone refines their word list to the point where they can autofill and generate a clean puzzle, that’s cool with me.

      • Cool with me too. My point being that for me quad-stack puzzle construction took me right back to basics… just like the old triple-stack construction in the pre-CC days.


        • Sure. And to be clear, I’m not completely opposed to these stacks on principle. Certainly, you’ve proven that triple-stacks can be done very well indeed. I just think quad-stacks, by nature, require too many concessions. One of them might yet prove me wrong, and I would be the first to praise it.

  17. Really thought-provoking post, Tyler, and what interesting comments. I have nothing but a tremendous amount of respect and deep admiration for all the top solvers who solve dozens of puzzles daily with only solving time in mind—but it troubles me that so many say they can’t remember a puzzle as soon as it’s over. I know it’s an important skill to go through a puzzle quickly and improve your time, but I think the speed-solving culture has made the online crossworld a far tougher audience. Yes, I regularly solve NYT Mon-Wed for speed, but I always go back and look at the clues I missed and admire the theme (even if it’s an add-a-letter or something else we seem to brush off as trite now), because the constructor put an effort into that too.

    I think the fill is really important, of course. I write my puzzles in Excel, resizing all the rows and columns and using XWordInfo and OneLook and Cruciverb, so I really select every word quite carefully. But sometimes a theme really is worth the concessions made in the fill (and sometimes a constructor overvalues how much fill you can compromise, and solvers disagree), so to not even bother to look at the theme, no matter how simplistic, seems wrong.

    • I actually don’t speed-solve terribly often, and in any case, the theme can certainly help one get through the puzzle quickly. I started solving fewer puzzles this year in part so I could appreciate each one more.

  18. love the discussion, lean towards dan’s comments on most of this…
    Smooth fill is extremely important to me, but theme still trumps all and I can’t stand the continual put downs of the pangram…to me sometimes that IS the good fill as far as I’m concerned!!! Crunchy crunchy Scrabble Scrabble!
    I know crunchy and yet smooth sound like they are opposites, but they ain’t!

  19. Don’t get me wrong. There are some puzzles I never speed-solve (Matt Gaffney’s weekly contest puzzles are an example), because I want to admire the details. I didn’t speed-solve the NYT Titanic Sunday puzzle either, nor do I speed-solve Liz Gorski’s Sunday masterpieces. And I never speed-solve cryptics, because the themes and wordplay are the whole point.

    But “easy-puzzle” themes are usually uninteresting to me whether I’m speed-solving or not, and yes, it’s probably because I’ve seen most variations on Mon-Wed themes gazillions of times. So if I’m going to remember a theme, it HAS to stand out, and sort of by their natures, Mon-Wed themes usually don’t. I’m sorry – I know constructors put a lot of work into puzzles whether it’s a Monday or a Saturday, and I have tons of admiration for that (not incidentally, because I can’t do it myself), but I’m afraid I’m simply not their audience. Which is OK – millions of people are. But all I have is my own perspective.

    That said, I’m not irritated by less than desirable fill in puzzles with triple stacks or quad stacks or some other construction feat. I cut puzzles like that a lot of slack!

  20. I enjoy outside-the-box, risky theme ideas the most – but only if the constructor has taken the time to ensure that the whole puzzle itself is not neglected. If the theme is brilliant, but to uncover it you need to bust through nasty obscurities, roll-your-own acronyms and abbreviations, etc., then it’s just frustrating. As far as themeless puzzles, I want to expect the unexpected. Include lively words and phrases (whether completely new or not), colloquial and classical references, a thorough mix of knowledge and quirky clues from all over the place. The best ones surprise you around every corner.

    I do think that autofill may be one factor in some puzzles that don’t meet this goal. It seems like any tool – you can use it carefully and in very specific cases like a master artist to great effect (perhaps to check a corner for viability), and ultimately produce a finer result. Or you can simply wield it gleefully like an overcaffeinated chimpanzee, and turn that potential masterpiece into a crazy quilt of OLIOs, ERNSTS, RETAP, etc.

    Of course, those of us posting here are likely the smallest fractional percentage of solvers. I wonder what the larger view is, if any.

  21. Okay, a few things.

    However, I’ve noticed a trend in some of today’s puzzles that I believe is detrimental to the fun of crosswords and the art of their construction.

    Without examples, this sure looks a lot like a straw man argument. Not saying it is; just that it looks like it. As some have pointed out, the fill quality overall was a LOT worse ten years ago than it is today. So if there is something here, it might be just a dead cat bounce. Maybe it’s just a blip in the long-term trend of fill inexorably getting better.

    If it *is* a real trend, then we’re looking at a few culprits: (a) new constructors (b) a race for fancy grids or (c) autofill. I can’t imagine it’s purely (c). First of all, you should NEVER simply autofill a puzzle and send it off (that would tie in with (a) here). Letting the computer guide you is fine, and occasionally (in my opinion) better than filling it by hand because the computer might suggest something you never would have considered.

    (b) might be an issue. It seems like these days you have to do something crazy to get Will’s attention and he seems to tolerate bad fill more than a lot of people in this discussion. So we get quad-stacks and 64-word grids and seven theme entries and of course the fill suffers.

    But you know what I think some of it is? Partly (a), but there have always been new constructors. I think some of it has to do with the fact that these new constructors don’t have Peter Gordon to help them anymore. The reason I’m borderline fanatical about fill is because he hated partials and such and worked with me to make my puzzles better. All of the puzzles I had published there were essentially co-op ventures with him and I became a much better constructor because of it.

    In general, I think new constructors don’t need a mentor as much with the amazing tools that are out there, and so they’re not taking one. So I guess what I’m saying is, bring back the apprenticeship system?

    Sorry for rambling.

    • No need to apologize; good points, all. As for the first part, I just meant to say that I notice it much more than I used to. That might simply come with my own increased experience.

      Excellent point about apprenticeships. The tools you mention can’t grasp the aesthetic quality of puzzles that’s a big part of this issue.

    • Speaking to your point about apprenticeships, I know that I learned so much of what has made me a “decent” constructor from Doug Peterson. I am so appreciative of the time he took helping me out when I was first getting started.

      It was incredibly helpful to me and I am sure others would gain too from a similar experience. Sometimes you just have to ask for help.

  22. I now think that things aren’t getting worse, fill-wise, so much as they are getting less and less understandable / excusable. There’s too much info out there about what’s good / bad, too many resources (electronic and otherwise), and (one would think) too many editors who should know better than to let shit slide.

    I think there is a combination of auto-fill / laziness / inexperience / insufficient oversight (i.e. the Nobody-is-Peter-Gordon factor) that is causing the *apparent* uptick in inexpertly filled grids. When I (a relatively inexperienced constructor) can make a corner of someone else’s puzzle manifestly, objectively better inside of a couple minutes, I get angry and resentful. Other much more experienced constructors I know feel the same way, to the point of routinely (justifiedly, amusingly) fulminating about it (in semi-private).

    I think the idea that “theme trumps all” is terrible. When your great themed puzzle has merely adequate fill, you are fine. But let a couple real clunkers get in there—ESP. if they add difficulty to the grid—and you’ve got problems. People get unhappy and your theme is left sadly underappreciated, if not forgotten. Kevin’s “Titanic” puzzle is a good example of this (Kevin is an excellent constructor, so nothing at all against him). Bad short fill in too many places, combined with difficulty, left most people grumbling rather than oohing and aahing. You can say “that’s their loss” I suppose, but satisfying a solving public is generally the name of the game.

    Everyone knows that crosswordese is necessary from time to time, and you often pay for great fill (at least a little) in the shorter stuff. It’s a matter of balance. It just feels like a lot of stuff that *could* have been great (with some work, a better constructed grid, fewer/more theme answers, etc.) is being served half-baked. Underdone. Great ingredients. Lame meal.

    I also agree that innovation is great. But the quad-stack is not “innovative.” The low word-count puzzle is not “innovative.” The pangram is not “innovative.” The first and third of those might be good, but not usually. The very low word-count puzzles should be attempted by only about four people on the planet, tops.

    Thanks for the thoughtful post, Tyler, and to commenters for the thoughtful responses.


  23. What great points, everyone! I wish I’d happened on the thread sooner but then I might have missed all the later gems.

    I agree wholeheartedly on the need for greater diligence in general in creating fill, theme and clues, and for the perpetuation of mentoring, in all its forms. I also think we all put too much on the shoulders of the editor. Yes, the editor is the arbiter of what’s acceptable and is the final checkpoint before publication – but what lands on that person’s doorstep needs to be better. The only way that product will be improved is by the constructor learning what works and what doesn’t and not settling for less – which can in most cases only happen once the constructor has had plenty of experience and mentoring, sufficient to have a sense of what will or won’t fly without being told. Every editor I’ve been lucky enough to have worked with has helped me improve as a constructor. Some have been more hands-on than others with their shaping and criticism, but when there’s been little or no guidance that has been fine too, as it ultmatley is saying something as well: “Hey, you’ve been at this a while – I think you know why this doesn’t cut it: you don’t want me to re-fill it for you, do you? That theme had 4 great entries, but did you really think you could sneak that fifth one past me and the rest of the world? I’d say ‘close, but no cigar,’ but then again, I’ve got a stack of puzzles I prefer and another mile of submissions to get to so, do what you have to and I’ll see you on the next one that floats my boat a little better, okay?” Well, that’s how I like lick my rejection wounds, at any rate!


    Really, I think the one thing that should not be tolerated is the laziness factor. Sometimes you just have to bag a theme or take many stabs at grids – and some of those times might include losing entire fills, some of which have prematurely inserted clues, some of which were just the dandiest anyone’s ever seen … Tough break! Rip it out! A bad corner that needs refilling? Are you serious?!!

    Thanks, Tyler, for an excellent forum to channel my insomnia!

  24. Another reason to avoid bad fill is that newcomers to crosswords (and there must be some new blood coming in, from time to time) have to learn all those four-letter words in order to go from beginner to intermediate level solving. That stuff is easy for me, since I started doing puzzles in the Maleskan Epoch, but someone just trying out crosswords could fairly ask “What’s an ELLER?”

    • Im pretty new to more advanced crosswords. I’ve only been attempting the NY Times puzzle for 2-3 months. I still find the Sunday puzzles pretty tough, but I’m getting better at the midweek puzzles.

  25. Great discussion. I am just an average solver at best and have little to no interest in constructing. The criticism on the Rex Parker blog has greatly increased my appreciation of crosswords as an art form. Before becoming a regular reader of the site I never really noticed dreck (nor constructor’s names). It was an ignorance is bliss situation. Now that I am more cognizant of what can be done I am disappointed when puzzles are weak. Weak fill is the most common cause of a weak puzzle in my experience. I cannot recall a theme ruining a puzzle, but the instances of bad fill ruining an otherwise good puzzle are easy to list. As further evidence of this, I believe all the pejorative neologisms on the Rex Parker site refer to fill.

    Having said that, I also sense that the issue is being overstated here. In baseball there is often lamenting about the absence of .400 hitters today. This was once seen as evidence of the lesser quality of players today. The statisticians have pretty much convinced most people that what really has happened is a reduction in the range of quality – great hitters don’t get at-bats against bad pitchers anymore. The best are as good as ever, but the worst players today are far better than the worst players of 50 or 100 years ago. I suspect a similar phenomenon here. The best constructors are still great, but the worst are pretty good. Statistically some dreck is going to get published. So we sometimes get weeks like this one in the NYT where on Sunday a constructor undertook a challenge and failed (IMHO) because of the fill, followed by a run of six days of great puzzles. So I doubt that there is more bad fill. Possibly, since the general quality of puzzles is better, bad fill is just more obvious.

    So, while I don’t think there is an overall trend of bad fill running rampant, if a constructor doesn’t want to be in the bottom 25 of this year’s top 1000 puzzles s/he better take care with the fill.

  26. One thing that hasn’t been mentioned much (if at all) in this discussion is cluing. For me great clues make a puzzle even more than great fill or theme does. Crappy fill can be made interesting, and thus excusable, by a good clue just as a trite theme can be.

    What’s frustrating to me as a newbie constructor is I get the feeling from editors that they’re rejecting puzzles solely based on the interest the theme generates in them and that they aren’t even looking at the puzzles themselves. Puzzles should be judged as a whole: theme, fill, cluing.

    Bob Klahn’s puzzles don’t always have the most original themes but when I see his byline I’m excited since I know that I’m going to LOVE the clues and the challenges he presents with them.

    • Alan, I think you’re never going to win that one! The editors are looking for theme first, theme second, and theme third. If the fill isn’t good enough, they can ask for a rewrite; if the clues stink, they can do their own rewrite. I love Klahn’s CrosSynergy puzzles too, but they generally wouldn’t be accepted at the NYT or LAT — and they shouldn’t, because the themes aren’t entertaining enough.

      • That’s a good point; the theme is the part of the puzzle where the most novelty and ingenuity is needed, so its priority in a submission is, to an extent, warranted. But the fill needs to be there too.

      • Ugh, Dan, then I say UGH! Theme, theme, theme…whatever. I still say you can build an extremely interesting “add a letter” theme puzzle with great fill and great cluing that would far exceed some of the “interesting theme” puzzles we’ve been seeing. It’s a shame…feels to me like we should be emphasizing quality of the whole over “theme-interest.” Disheartening.

        • I can understand this. With the number of puzzles out there, it’s always more difficult to come up with great themes, but every puzzle can be filled cleanly.

    • Here here. Clever cluing is often far more interesting (and at times even more humorous) than clever themes. I am more likely to forgive a crappy corner if some of the cluing of that corner makes me smile.

  27. Ha ha. When I see a Bob Klahn puzzle, I know I’m going to be pissed off and frustrated. Doesn’t men I don’t enjoy them, but it’s never, ever a smooth ride.

  28. More on topic than the last post (please read “mean” for “men”), I share the displeasure with odd abbreviations and partials — but not the occasional criticism of old-time crossword-ese. I started doing crosswords when I was about 8 (literally at my Nana’s knee), and snee and adit and the like are my old friends. Things from pop culture that are gimme’s for Rex (for example) are annoyances to me, but I live with them. If I can bother to learn and remember references from rap/TV/movies, y’all can learn and remember some of the classics. Just sayin’

  29. I am a solver, not a constructor. I did my first NY Times Sunday puzzle during WWII.

    There is a growing trend to using circled letters which I abhor. With rare exceptions, these ‘cute’ artifices add little interest to this solver, and I would like to see fewer of them. I note that they add some challenge to the constructor, but rarely add challenge to me.

    • I wouldn’t go quite that far, though sometimes I think the circles take away from the aha moment of grasping the theme.

    • I love circles! I pick up the Times in the morning, but don’t solve it until the evening. But I always glance at the puzzle first and check the theme clues. If there are circles, or shading, or something unique about the grid, or if I have no idea where the theme is going, it heightens the anticipation.

      Of course, circles or no, the solve doesn’t always measure up to the anticipation, but I wouldn’t trade it. Tyler’s right that sometimes the aha moment can be spoiled by circles, but I like the… er… pre-ha moment of looking forward to finding out why they are there.

      To say something on-topic, though, I think Tyler has it right that if fill quality is occasionally low it is due to the feeling like you need to reinvent the wheel with a theme. In fact I’m more optimistic than some of the commenters – I think we’re just getting more sensitive to fill the more we discuss it on blogs and mailing lists, not that the quality is going down – but sure, there are puzzles that drag because of iffy fill. It would be interesting to know if there’s a relationship to theme letter count.

  30. Great discussion. As a solver, I tend to appreciate each day’s new offerings, and don’t get involved in the rating systems that are available. I guess I’m more interested in rating myself against the challenge of the puzzle (or perhaps against my crossword “nemeses”). Some of my favorite constructors are Byron Walden (more please, Byron!), Bob Klahn, the Patricks, and Tony O.

    As for fresh fill, I am on puzzle 128 out of 144 of Natan Last’s new book “Word”. The fill is definitely fresh. I even blushed a couple of times. But I also chuckled a few times to see the fill including old movie stars, the “E” of BPOE, etc. So, creating that many puzzles apparently couldn’t be done in the vacuum of things that happened in the last quarter-century. But then I realized that this was a good thing; for someone in their teens or twenties wanting to start doing puzzles, they might pick up some of the crosswordese they would need for more widely available puzzles, and I bought copies for my daughters (who had already filled in a couple of grids in mine).

    • That’s the biggest benefit of the modern style of crosswords, I think. If I’d solved or constructed during the Maleska era, it’s quite likely I wouldn’t be doing them today.

  31. My foremost criterion for evealuating American-style puzzles is in the juggling of possibilities. Say your have a 4×5 area Clued with numerous possibilites, so you have to weave some sensible fabric out of the increasingly numerous permutations. That’s my kinky sort of delight. And back to Tyler’s essential point: a proponderance of crap fill is evedence of mediocrity desipte any thematic brilliance.

  32. (Oops, two typos — seems like I ouoght to be more careful in these parts.)

  33. Ir couldn’t agree more with that puzzle guy. Last week’s ny times (April 15, 2012) had an ambitious them based on the titanic but the fill so so god-awful it made the puzzle not with doing. Great them, bad fill = lousy puzzle. I also agree that some fun cluing can at least make up for some bad fill.

  34. Older NYT 15x puzzles tended to have 3, or maybe four theme answers. As a constructor, I find it’s much easier to fill such a grid cleanly, than when you try to go for 5 or more theme answers. Heck, once I had a NYT puzzle published that had 15 theme answers, with several of them intersecting. My original fill was pretty wobbly, but Will had Frank Longo fix it, and his result was very smooth. So it can be done, even in many extreme high-count themes, with extra effort and lots of talent!

    Not that Will insists on 5 or more theme answers. 3′s still sell, too. They just need to tickle him enough. Wish I could think of lots of “ticklish threes”, but doesn’t always happen that way.

    As a solver, I just want a fun experience; learning something along the way, while still feeling I’ve conquered the puzzle, is even better. What’s crud fill? What’s a good theme? What’s the trade-off line? I’ll go with what Will chooses for publication; he’s the top expert in the field, I think.


  35. True story. Once I submitted a puzzle to NYT with “VAYS” as a fill entry. Clue was: “Ve have ___ to make you talk!” Will accepted the puzzle, but changed the entry. I was crushed. So, there’s an example of My judgement, vis-a-vis crud fill. Don’t even get me started, on some of the weird themes I’ve made poor Will endure — but they had great fill!


    • Heh. I can certainly understand why VAYS was tossed from an NYT puzzle, but I still kind of like it for some reason. It might fly in a wackier alt-weekly or something. It’d be funny in a Something Different puzzle, too.

  36. Great conversation.
    Like Marc Z, I was blissfully solving puzzles for decades, ignorant to the artistry. Rex’s blog, aka Crossword Appreciation class, taught me lots. Both from Michael and from the constructors who post.
    I’m glad for the sharp criticism – I think it helps both constructors and solvers.
    Having said that, learning to appreciate fill/clues/construction hasn’t made me dislike more puzzles – it has made me like all puzzles more.

    I now enjoy looking over the finished grid, doing a replay of my solve, and really appreciating some of the subtleties. Speed solving? Why??

    As for themes, I love them! Good ones are another level of puzzle. There’s always Fri/Sat for sheer crossword purity – but give me some fun themes the rest of the week.

    Fill? Clue it cleverly, and we’ll forgive you! All I ask is that you make thoughtful decisions to insert dreck… ;)

  37. Totally agree!!! I also hate puzzles that use expressions like” I don’rt give a hoot”….I learn nothing from things like that..they are just filling with vernacular instead of information or true witty word usage…and they are always so long..
    And most of all, musical groups or bands or trios ..ugh.
    Thanks for your essay on subject. Joyce

    • This is proof that people want different things in their puzzles. For me, it’s a given that modern pop culture references are superior to older references, even somewhat independently of whether or not I’ve heard of them. As I said, obscurity is a tricky beast.

  38. What a watershed discussion– and so polite, compared to the industry discussions I’m used to! It may not have been quite so nice if Tyler were to call out more specific examples, as Dan requests, and the authors of those particular puzzles got word. Though Dan is right that limiting the discussion to unspecified “crap fill” and a general injunction against quad stacks, pangrams and other stunts does leave matters vague.

    So that’s a trade-off decision.

    Which is appropriate, because crossword-making seems to me to be one trade-off decision after another. The initial argument, as I understand it, is that the wrong values are being assigned in the trade. The grids are now too open, the uncommon letters too forced. The stunts leave the puzzle stunted.

    It’s an important conversation. Personally, I often feel overwhelmed by evaluating the goodness and badness of each word in a puzzle. Sometimes I nod enthusiastically at the orthodoxy, and sometimes I feel like either I’m an illiterate, or everyone else is. What’s that fine line between a “fresh” entry that no one’s ever seen before and an “obscure” entry that no one’s ever seen before, and why does everyone know it but me?

    Crosswords seem to acknowledge central authority a lot more than other forms of entertainment do. If Will Shortz or Peter Gordon says something– and if they’ve EVER said something– it’s often simply accepted. I have plenty of respect for Shortz and Gordon and the others who toil away at this field, and I know that the civility I mention above is related to that respect for authority. But it’s still hard for me to go along with that.

    ETUI is one of the classic examples of crosswordese. A couple of editors allow it, but nobody really seems to LIKE it, except in an ironic, in-jokey way. Yet I notice that when I Google ETUI lately, I find a couple of references to it in conversation or Wikipedia articles that have nothing to do with it as “an example of crosswordese.” So is it coming back, and if it is, should crossword critics be sharpening their knives for it or accepting another language shift?

    My biggest qualm, really, is the criticism that seems to be piled onto long entries that have common letters. I agree that, in a very generalized way, common letters translate into dull grids. But I find entries like TENNESSEE TITANS and SESAME STREET perfectly fresh and interesting, and I’d certainly rather see either of them than another AJA or OXO, even if the latter two do cross JURY BOXES or JURY FIXER.

  39. I like T Campbell’s point that our priorities are slightly out of whack. I don’t actually solve conventional crosswords very often, but as a cryptic constructor, MY bete noire has been the tendency of brainy folks to make their poor cryptic grids hold up three to five interlacing themes, to the detriment of basically every damn clue in the puzzle. Since we seem to be sharing mantras, mine is, “a puzzle needs to breathe.” A lot of crappy fill (or, in a cryptic, uninteresting wordplay in the service of dull fill) is the first sign of dyspnea.

    Also, since we’re all repeating THIS, too, let me add, “…not that I’ve been consistently perfect at this myself.”

  40. I wasn’t going to bring up ETUI, but since T did: for many people who do needlework, etui is not an obscure word. I own an etui (and I know how to use it!). I do not own an adze, but I do not object to it as fill. I think it is something I should know about, and I am open to learning other new terms from other crafts, disciplines, etc.

    • Thanks, Jan!

      I don’t own an etui, but I know it’s a legit entry so thanks for pointing it out! It’s way superior to say ERNE… No one owns an ERNE, and if they did, they’d more than likely call it a White-tailed Eagle! Of course it’s also a lough, but I digress…

  41. Thanks to Tyler for the smart, thoughtful essay on crossword fill. As an editor whose work gets commented on a lot, I thought I’d chime in with some thoughts.

    Sometimes, I admit, I may be guilty of permitting a few too many subpar entries in the Times crossword, because a) I think an ambitious puzzle theme or constraint justifies some stinky entries, or b) I think the poorer entries are outweighed by enough good ones. This is all a matter of opinion and taste, of course. Also of balance.

    I do disagree that the quality of fill is declining, at least in the Times. Check out the puzzles of 10-15 years ago. They are not as clean or as interesting as they are now. And if you go back before 1993, when I started, you’ll find the fill much inferior. That’s the way things were. Standards have risen.

    I doubt that subpar entries in Times puzzles are the result of lazy use of autofill, or any sort of laziness, as some have suggested. In my experience, most contributors are trying their best.

    I will revise corners of grids myself to remove unappealing entries. For bigger changes I’ll return puzzles to the contributors for revision. Occasionally I’ll ask Frank Longo for a major fix. I don’t routinely rework entire grids, though, because a) I don’t have time, and b) I want a puzzle to be mainly the contributor’s own work.

    Perhaps nowadays there has gotten to be too much emphasis on quantity of theme material or novelty/razzle-dazzle of themes over the quality of the fill. Dialing back a bit on this might improve the overall solving experience. This is something I’ll think about.

    There is always pressure to boost themes, though. A couple of weeks ago, for example, Rex Parker criticized a theme as being “thin,” which I thought was a fair criticism (even if I disagreed with it). But he also criticized, in stronger terms, the quality of the puzzle’s fill. My reaction was … you can’t have it both ways. Given the theme, I think the constructor did an admirable job. Having better fill probably would have required an even “thinner” theme.

    Here are some other things I see on the blogs, repeated a lot, that I disagree with:

    1) Many times bloggers and commenters call certain words “obscure,” or say they know them only from crosswords, and I think “Really?” This is stuff I know from reading and everyday experience. At the very least I wish bloggers and commenters would have a little more humility about what they don’t know. Sometimes, through no fault of the solver, he or she just doesn’t know something that I think a Times reader should.

    2) There seems to be a growing feeling among bloggers and commenters that anything obscure is bad. I strongly disagree. I don’t understand why *all* the difficulty of a puzzle has to come from the clues. Some of it can just as fairly come from the grid. Presumably people who do crosswords love words and knowledge. It should be pleasant to learn something occasionally.

    3) For some people, any difficult word on a Monday puzzle is considered a flaw. No. A Monday puzzle is supposed to be easy, but it can still contain a difficult word or two if the crossings are fair and the overall effect of the puzzle is easy.

    Other thoughts:

    Like everyone, I love crosswords with juicy fill and no dreck. But I don’t think every crossword has to be like this. Sometimes it’s nice to see how far the English language, or some particular theme idea, can be stretched. As long as the creaks aren’t too numerous or noticeable, I think the result is worth publishing. Some people here disagree, but I favor the widest possible variety.

    Always bear in mind how little crossword constructors receive for their work. For the Times that’s $200 for a daily, $1,000 for a Sunday — not much, considering the amount of effort and skill that goes into making a puzzle and the fact that I accept only about 10% of submissions. Criticism on the blogs is warranted when a puzzle could have been done better. Over the years this has been a major factor in the improvement in crossword quality as well as making solvers more discerning. What I object to are angry rants. These are disrespectful and hurtful to the constructors, who are not doing this for the money. I think it’s also hurtful to the crossword community as a whole by inclining readers to be negative. There is a better way to express criticism than by slams and rants.

    The calm, civilized discussion here, of course, has been perfect.

    –Will Shortz

    • Pure Consumer here

      Re your points 1 and 2: I read Rex daily, Amy frequently, and while each frequently say that they only know something from crosswords, they’re humbly stating that that’s what they know, and how they know it. “I know this only from crosswords” doesn’t equate to this shouldn’t be in crosswords, That’s not what they’re saying.

    • I am completely with you on objecting to angry rants. Yes, they are “disrespectful and hurtful” to the constructors. I’m very new to Rex, but almost every day, I wince at some of the mean things people say and how the constructor, who no doubt worked very hard at the puzzle, must feel. I wonder, too, if the negativity hurts the community in another way; after seeing how mean people can be, I would be very hesitant to try my hand at constructing. I wonder if the hostile comments are scaring off would-be fledglings with fresh ideas.

      • This is starting to remind me of people’s objections to Eugene Maleska. He was notoriously ruthless with puzzles he didn’t like. Will and his peers are far more supportive, and I think that’s part of the reason for the vast improvement in the NYT puzzle since 1993. Of course, a blogger is not an editor, but it’s something to keep in mind.

  42. I, also, doubt very highly that constructors are slacking off by using autofill. Just speaking for myself (as a relatively newbie/wannabe constructor), the only time I’ll ever hit that button is when I have, say, a 4×5 empty corner and am curious as to what autofill will come up with. It’s usually dreck. But, even if it came up with something solid, I’ll say “hmmph”, delete it, and fill it in myself. (If my name is on it, blame me for it.)

    I do wonder if constructors are relying too heavily on, say, Matt Ginsberg’s clue database as an arbitor of what is acceptable fill.

    For instance, two or three years ago I was still very much struggling to fill grids at all, nevermind with acceptable fill. With what was a tough area (for me, at the time) to fill, I used for possibilities and came up with T TEST. I’m pretty sure that it did not appear in Matt’s database, until it made it in there as a clue from said puzzle of mine. Today, the entry TTEST has this history:

    TTEST (5)
    1 Th LAT 11 Statistical hypothesis trial
    1 Th CHE 09 Way to determine if a sample is indicative of the norm
    1 We Tau 11 Way of checking a null hypothesis
    2 Tu- >1 02 Method for checking a statistical hypothesis
    1 00 Statistical hypothesis procedure
    1 00 What a statistician might perform

    T TEST just showed up this week in yet another puzzle. I doubt that anyone wouldn’t agree that T TEST is low quality fill, and I suspect that it’s appearance in the database may be contributing to its further use.

    Consulting the database for fill possibilities is something I do constantly, along with pattern searches at and I suspect that nearly every other constructor does the same, but I am always surprised that some reasonably common words/names do not appear in the database, at all.

    For example, I just recently posted a puzzle that included VINLAND. I checked Matt’s database to see what clues had been used previously (so as not to repeat one), and discovered that it is not there. VINLAND is fresh fill?! (Same with TATTLETALES in the same grid, though TATTLETALE has appeared numerous times.)

    Just another observation.

    Thank you, Tyler, for getting the ball rolling on this amazing thread.

  43. All this should be directed at editors and not constructors. If editors want better/perfect fill, they should demand that the constructor rework the puzzle. Shortz lets borderline stuff go and never tinkles with grids. Gordon the opposite.

    • In my experience, Will has never hesitated to ask me to rework part of a grid if he thinks it can be improved.


    • Shortz lets borderline stuff go and never tinkles with grids.

      Hmm. Will Shortz informed us of just the opposite in his above post. He’ll refill a corner but, for anything major, he’ll send it back to the constructor for reworking as he would “want a puzzle to be mainly the contributor’s own work.”

      • Second that. My first NYT puzzle had two sections (W and S middle) slightly tweaked to make it a little more friendly for its particular day.

    • Wow, I wrote “tinkle.” Anyway, what i mean by “tinker” is taking a perfectly acceptable grid and aiming for something a little better.

  44. Well-said so far – I just wanted to give a (very) novice constructor perspective. I’ve done my best to fill without auto-filling, and to pay maximal attention to fill that includes “fun” answers; not necessarily the Scrabbliest letters, or the most attention-grabbing phrases, but overall a fair, enjoyable solution that I am (currently) capable of. I try to review and revise all areas of the grid for improvement after completion, so that’s a start. Overall though, I can see the differences between what I and an experienced constructor can create. Newer constructors may still have to learn the nuances of grid design, cluing, optimal fill, etc.

    I understand that especially for newer constructors, their early attempts may be criticized more harshly in the blogosphere; that’s OK to a point, if it’s a fair and civil discussion, but not too negative or at all nasty. More than anything, I don’t want to see new talent discouraged from making the attempt because of a fear of public backlash. So for those who discuss such things online, be honest in your opinions, but also be fair and reasonable in your discourse. Thanks.

    • And I would add to Howard that the same applies to those trying new things. I suspect the first triple stacks had suboptimal fill. Similarly, quads are in their infancy. The first of anything show it can be done. Then come those who improve on the concept. Let’s encourage those willing to go out on a limb and take us in new directions.

  45. I had planned on posting a comment here a few days ago…then life intervened. Anyway, seems that the conversation has continued. Not a lot to add at this point, but a few miscellaneous thoughts:

    1. Everyone is talking about bad fill, but no one is saying exactly what it is. There have been only a few examples here (ETUI, for one), and even then there’s a difference of opinion. I suppose we could empanel a group of crossword pros to pass ultimate judgment on what’s good and what’s bad, and we’d probably get a few convictions, but many times I think we’d end up with a hung jury.

    2. Crossword fill is not getting worse. Quad stacks notwithstanding, on average it’s getting better. What has changed even more than the quality of fill is the demand for better fill from a vocal group of solvers on the blogs. Look back at the early days of Amy’s site, and you won’t see the litany of offenses that are a daily feature today. The puzzles aren’t worse now, but today everyone seems to be more aware of the flaws. People want sparkling fill in every puzzle and grumble when they don’t get it. I’ll admit that I don’t find the griping any fun to read (and like Dan et al., I am not a fan of ratings either), but if all the noise moves puzzles toward having better fill, that’s not a bad thing. Still, fill is just fill. To my mind, it’s secondary to other elements in a puzzle. And I find the most enjoyable puzzles to be the most adventurous. Let’s not lose that.

    3. Quad stacks are getting better. The only things that have bothered me were the recycled 15s and a couple of repeated words, but I’d guess we’ll see little of that going forward. If Joe and Martin keep making them, and making them better, I’m willing to do them. You might not find it art you’d put on your wall, but I think they’re a new type of crossword challenge worth solving on occasion.

    Lastly, thanks to Tyler for kicking off the conversation, and to all for the many thoughtful comments. It’s been fun reading.

  46. Tyler said: ” I’ll be blunt here: If I never see a quadruple-stack of fifteens again, it’ll be too soon. Sure, it’s eye-popping, but it’s no fun at all to muddle through the inevitable handful of weak answers crossing the stack. ”

    Okay, then, why don’t we — at least hypothetically — talk about what was particularly loathesome about the Dec. 31, 2011 and Mar. 31, 2012 Quad-stack puzzles in the NYT. I haven’t read any blog opinions — other than the NYT Wordplay — in about 2 years now, so I’ll listen to any fair criticism that isn’t simply pre-fabricated upon reading the byline and sizing up the grid as extreme… as is usually the case in these blogs.

    No, these puzzles aren’t going to have the same qualities as the 72-worders with sparkling seed entries. Instead, these puzzles present a different type of challenge for solvers… the specter than there are no 3-letter entries to gain a foothold in the center. And yet, both these puzzles are paper tigers because they are cleanly enough filled to melt away completely. What is underappreciated here is that there is careful attention to minimize the use of propers and to use common words as much as possible.

    I get the impression that most of the opinions circulating in the crossword community are formulated by fast solvers who sometimes even forget to go back and appreciate themes when they are present; Anything that disrupts a fast solve is perceived as an unpleasant muddle. So, those are our current opinion makers… and yet most common folk solvers are not like that.

    So, let me tell you how these Quad-stacks provide an enjoyable solve to those solvers not concerned about fast completion. Again, it is what I refer to as the paper tiger effect: Initially the solver looks at the puzzle and thinks that all that white space is going to make it impossible. Yet slowly but surely, with each pass of the clues the solver slowly chips away at the puzzle, and actually manages to complete the entire thing. This is the type of solver who walks away with a sense of accomplishment… the feeling of having conquered something which initially looked impossible. In my opinion, if a solver has that type of experience, then I’ve hit a homerun.

    So, I will gladly cater to that solver who enjoys the challenge of clue misdirection … the type of solver who doesn’t mind taking the time to ponder words from a couple different angles before figuring out the word play. The important thing here is that this type of solve is among the variety of styles that Will selects for NYT. It would indeed be very dull if all puzzles were the same variety. What’s important is that each variety has it’s own purpose which satisfies a niche in the solving demographics, and I’m perfectly happy providing the paper tiger style.

    • See, a large part of my point is that it’s possible to deliver the satisfying solving experience you describe without a grid design that forces fill like REMET and OHI. I still relish tough nuts from Wilber, Walden, “S.N.”, and others, and there’s nothing particularly special about their grid patterns. I still sometimes find a corner that I’m on the verge of giving up on, but then I finally break through. So I have to disagree that quad-stacks and other stunts are unique in their ability to provide this sort of gratifying challenge.

      As I implied in replying to MAS, I will concede that I may have overstated my opinion on quad stacks, and I will further concede that yours are about as well-executed as I imagine they could be. At the same time, I still don’t think they’re worth it, by and large, and I wouldn’t miss them if they were gone.

      • That’s cool, Tyler. I just think that these extreme grids get too much of a bad rap … some critics don’t seem to want to concede when the fill is near flawless … or not any worse than other puzzles. Clearly, the Quad-stack region is liable to have less sparkle than those other puzzles, but the sparkle resumes in the non-Quad regions. If one of these days I do manage to get an extreme puzzle to really shine all the way through, I still doubt the critics would acknowledge it.

        • I think we’ve got the heart of it. I would counter that one could make a puzzle that sparkled all over and delivered the same satisfaction in solving. I suspect you would reply by saying stuff like quad stacks delivers a unique and worthwhile experience, yes? It comes down to a philosophical difference, which is fine.

        • I (largely) didn’t enjoy solving your first quad stacks, and said so, quite bluntly. The fact that the “sparkle” resumes in the non-stacked areas irked me, because I knew the rest of the puzzle could’ve been like that! That said I do think the last quad stack puzzle DID shine all the way through, and I admitted as much! (Plus I remember enjoying all the traps I fell into, there were many!) I do wish that (and maybe the one before it) were the first quads of yours I’d seen….

          I should add that – as a constructor who has tried, once, to fill in a double stack, and thereafter given it up as a mug’s game – I admired the extreme difficulty involved, it just didn’t translate into fun-to-solve puzzles for me, because of the annoying supporting answers those earlier puzzles had.

  47. On the sparkling fill/adventurous theme continuum (or dichotomy, if it is a dichotomy), count me solidly in the adventurous theme camp. If I run into a drab piece of fill, I fill it in (it takes two seconds to write in ETUI) and move on, with my eye on the prize of what magic lies ahead in the theme. If I finally get to the point where I’ve sussed out the theme only to discover that it’s another all-of-the-last-words-in-these-entries-can-precede-the-word-HAT kind of themes, I feel as though I’ve been had. It’s a buzz-kill that no amount of sparkling fill can fix.

    I also agree that since two of the major blogs are written by speed-solvers, their tone (and the tone of many of the comments on their blogs) tends to skew in the direction of “any thematic quirk that slows me down is bad”.

    As for me (and I suspect most constructors), I write the kind of puzzles I enjoy solving, so that’s what I’ll continue to do.

    • You and Joe both seem to believe in this speed-solving bias on the major blogs. I’ll let them defend themselves, but for my part, I’d like to hear what evidence you have for this.

  48. Here’s the thing about speed-solving (at least in the “probably won’t ever make it into the ACPT finals unless there’s a mass casualty on Friday night” category): I’m not battling the clock when I solve, I’m SOLVING THE PUZZLE. What the clock reads when I’m done gives me a way to estimate how difficult the puzzle was and where it lands in the NYT continuum.

    I have found it incredibly insulting for several years that some people who are not speed-solvers feel compelled to criticize the act of speed-solving. “It’s like swallowing a Michelin four-star meal in 5 minutes.”

    Here’s the thing: Not everyone’s brain is wired to work that way (the near-instant recall, the near-photographic memory of words). This is no insult at all–Martin Herbach is brilliant and knowledgeable and finishes every puzzle, but he’s not a speed-solver, it’s not how he works crosswords.

    You know why speed-solvers solve quickly? Because that is how long it takes us to answer the clues (honestly, it’s as simple as that). We solve fast because that’s how long it takes us to solve the puzzle. (Plus, it’s fun to push our brains to do that quick recall.)

    We don’t get mad at any complicated theme that takes us longer to unravel. Many of fastest solvers devour Matt Gaffney’s Weekly Crossword Contest and the Fireball crossword, and we relish the mental work needed to fully conquer those intricate themes and metas. We also love tough themeless puzzles because they make us think hard, rather than just zip through a grid with little effort.

    When I work a puzzle with a dense theme and the fill gets squeezed by the theme, the entire puzzle is less enjoyable. Five long theme answers and 69 unthemed? Yes, I want those 69 answers to provide me with entertainment too. When my Scowl-o-Meter starts flashing again and again (AGN!), I start wishing that the theme had been reined in a little (or laid out differently) so that the rest of the puzzle could have been better.

    Also, it’s insulting to hear that Joe and Peter think “the bloggers” (presumably meaning me and Rex) don’t slow down to appreciate the theme. What the hell are we doing when we write about the theme (and the fill, and the clues) and launch a conversation about the puzzle? Crikey, anyone who solves the puzzle and then writes a review of it has spent MORE time and thought on the puzzle than the typical solver. You know how many solvers aren’t even aware that crosswords HAVE themes? Or have trouble identifying the themes? They visit a blog and pick up that info, and they gain an increased appreciation of what makes a good puzzle good.

  49. Let me put it this way: It was great when Ryan and Brian were blogging and developing their puzzle solving skills right before our eyes. Rising from D division champs to C division champs (or whatever). It sent a message loud and clear about how mid-range solvers could improve themselves… by demonstrating how it was done. Yes, they had opinions about what they considered good fill and crosswordese, but they always did it with self-effacing humor. And yet, these guys were genius enough to run their own crossword competition. It would sure be nice to see a new blog fill that void. In fact, it would be fun to see a tongue-in-cheek blog showing half-finished puzzles containing a small fraction of wrong answers … and having a blogger defend his wrong answers. Now that would be something worth reading on a daily basis. How better to highlight crosswordese than to say: I couldn’t finish it.

    Any takers?

    • Joe – love the idea…perhaps I can modify my blog to have more of this focus.
      But… I would offer that this is already a big part of the top xword blogs. Look at any Rex Parker post/comments…they are rife with writeovers, or what we call Word Evolutions.
      Dermatologist’s concern [waRt-->Boil-->BURN]
      Late, as a library book [overDUE-->PASTDUE]
      And we often justify, in a very self-deprecating way, either our bonehead mistake, or defend (as in the case of PASTDUE), why we really truly feel it’s wrong.

      And I already have a Epic Wrong Answer Hall of Fame on my blog referenced above ( – look for it on the left sidebar.

      I am thinking that your idea would be a fun thing, and I have the technology that is just spot-on for what you describe – both highlighting the replay, complete with wrong answers, and tagging the answers as-you-solve.

      But is it different enough from Amy, Rex, Deb, and what we poster already do there every day?
      I’m looking forward to hearing more.

      • That’s all there is to the idea. Just display half-finished puzzles with wrong answers included, and adamantly defend them! It would be hilarious. For instance, I recently put in TOAD rather than TOTO for [Biter of Almira Gulch] … I think one could reasonably assert that a toad is found in a gulch, even if others insist that it’s more of a hopper than a biter.

        All apologies that I am not familiar with your blog.

        • LOL – I certainly don’t count mine as one of the above mentioned “top blogs”…that was shorthand for Amy’s, Rex’s, Deb’s, etc…though upon re-reading my post, it kinda sounds like that’s what I meant. Anyhow – I will post some such partials on my very obscure blog soon.

  50. @Joe Krozel…re: your statment:
    “I haven’t read any blog opinions — other than the NYT Wordplay — in about 2 years now, so I’ll listen to any fair criticism that isn’t simply pre-fabricated upon reading the byline and sizing up the grid as extreme… as is usually the case in these blogs.”

    You have deflated my xword-ego. I naively admit that I thought many constructors would visit at least the top 3 or 4 blogs to hear not only the blogger, but also the solvers on days when their puzzle is published.

    It is also surprising to me that you can make that statement, and then, continue to describe both the bloggers and the commenters…if you haven’t been there in 2 years, how can you discuss them?

    As an dedicated daily solver of the NYT, and an avid reader/poster at 2 blogs, allow me to share my thoughts:
    First, re: speed solving…
    Sure, a large cross-section of folks are speed solvers. Many more(myself included), are not. You would not have the misconception that you do if you read more blogs. But even those who speed through will sit back afterwards and give it some thought.
    To quote you again: “Anything that disrupts a fast solve is perceived as an unpleasant muddle. So, those are our current opinion makers…” Again – read some current blogs. Most certainly not the case.

    But you really ought to take the time to ready what people think about your puzzles just a bit more often, especially when you went out of your way here to say “I will gladly cater to that solver who enjoys …” I wonder how you know what the solver enjoys, if you no longer check the blogs?
    (No, they are not a complete cross-section of all solvers, but they are an important demographic.)

    Many blog comments are
    a) well thought out— not “pre-fabricated”,
    b) both praising and critiquing
    c) insightful, engaging, and educational. Sure, there are jerks there too, but they are pretty easy to ferret out.
    Wordplay, with all due respect, adds little value to the puzzle for me. Why? I don’t learn enough from it. You can’t learn when everything is presented on an even keel with only praise.
    Whether or not I agree with the sharp criticisms, or the high praise, it gives me great insight into the art & the mechanics of construction. One of the highlights is always when the constructor drops by.
    And I perhaps naively think that the constructors learn something from us, too.

    Once again, thanks for this thought-provoking topic.

    BTW – Will – I think your observations on obscurity are a whole other discussion.

    • I haven’t said much here, although I am in agreement with Tyler’s initial point (remember that?) about quality fill. At the same time, I also agree with Will’s point about balance and stretching the boundaries about whether a word might be considered obscure or just new and (subjectively) shiny.

      Am I playing both sides of the coin? No, I don’t think so, and that is my point. This is a creative endeavor and there shouldn’t be any one “right” opinion about what makes something good or not. In any art form you should be able to find both perfection and flaws. And there is no innovation without experimentation.

      More importantly, I think we need to be reminded sometimes that this is ENTERTAINMENT to most people. We are a hardcore group of solvers, constructors, and editors, so of course we all feel closer to the product than most people. If you ever have the luxury of sitting in on user research sessions, however, you will learn very quickly that most people are not even aware that puzzles have bylines. Or that the puzzles have themes. And they don’t much care. Talk about ego-deflating.

      Which leads me to the comment above about Wordplay (you knew that was coming, right?) If you believe that everything in Wordplay is presented on an even keel with only praise, then you are not reading my blog very carefully. When there is something that bothers me about a puzzle I will say so, but much of the time it’s in what I neglect to mention rather than what I do say. There’s a reason for that.

      My job is very different than those of Amy and Rex (who I will also remind everyone do this same job daily and without compensation, so I vote we be respectful and kind to them). Not only am I not allowed to give away more than a certain percentage of answers in the puzzle, but I am also charged with keeping discussion within New York Times standards. That certainly doesn’t prevent people from liking or disliking parts of a puzzle, and it doesn’t prevent me from talking about certain aspects of construction or interviewing constructors so readers can get the behind-the-scenes stuff, but yes, WP is a civil (mostly) place where people can feel safe discussing the puzzle. At the end of the day, it’s a BUSINESS, and we would lose readers if it was operated any other way.

      On the other hand, if civil discourse prevents you from learning about puzzle construction, I will point you to the same place I point anyone who is interested in learning: to the Sage Advice section of, or to Patrick Berry’s book.

      • Deb- excellent points. In my attempt to keep from running on at the keyboard (as is obviously my wont), I tossed that sentence out there in a bumbling, incomplete way. I apologize.
        While I recognize this is going off-thread, I feel I should restate and make a feeble attempt at redemption:

        I go to Wordplay for your quick synopsis, and for the interplay from your family of commenters, who spark interesting tangents. It’s always fresh and fun. And yes, you point out dreck where it is deserved. Point taken about the significance of what you DON’T say too!
        You do, in fact, have a difficult balancing act to pull off, and we realize that. Kudos for doing it well.

        I go to Rex Parker and Fiend because there I learn lots about construction and style, and that has hands-down turned me into a far better solver. Which makes solving more ENTERTAINING.

        All of which leads me to thank you relentless bloggers for helping me recognize a great puzzle when I see one (hmmm… was I better off being oblivious to poorly clued bad fil?)
        Tyler – I hope your post will at least be another reminder to editors and constructors to do the right thing by bad fill.

  51. No, I’m afraid I stopped reading the other blogs when the level of abuse got unreasonable. And now almost two years later I still get reports from other constructors that the abuse level remains high. No thanks. One does not stay in an abusive relationship.

    I continue to read the NYT blog because it remains civil — and solvers are still allowed to bring their gripes forward politely. But lately I’ve noticed that even that is changing. (Not sure if I’ll need to abandon that as well some day).

    You know, if I could make a recommendation to new constructors, I’d say go read the blog archives from 3-4 years ago when things were more civil. You can learn about pitfalls by reading the feedback on other constructors puzzles. And after a while you start noticing the same useful advice coming forward: Don’t cross two propers to create a Natick square; Don’t fill your sub-60 puzzles with lots of -ER and RE- words; etc.

    Yeah, I even used to leave a comment now and then on those other blogs — still do on the NYT blog — and that’s the thing that the bloggers and many solvers seem to resent: constructors expressing the constructor’s viewpoint. I got some severe tongue lashings over that. So that told me that the blogs are meant more for solvers to interact with other solvers… and that’s fair enough by me. No need to stick around.

    The NYT blog could probably be considered safe haven for constructors to leave their comments. Or, if a solver really wants the constructor to know a viewpoint they should feel assured that the constructor will read it there. As always, politeness helps. But if commenters there start resenting constructor interaction, they will likely lose it. And at that point the blog system will truly have become disfunctional.

  52. Oh Tyler, there’s a business opportunity here. You might start your own brand of puzzles, free of the “bad fill” you refer to (tho I’m not sure what it is). Let the market, the paying solvers, decide. I’m serious.

    That said, I’d like to support and applaud those constructors who aren’t afraid to take risks. If your puzzle has been maligned, spat on and destroyed by the blogs — you’re probably doing something right. Creativity and innovation find resistance in the mainstream.

    I agree with other commenters — Tyler, isn’t your beef with the editor or publisher, the person who has the power? The constructor is doing a job for a specific puzzle market, and doesn’t decide what gets published.

    Telling individual constructors to improve the NYT puzzle is like telling the bank teller to — stop being lazy and and lower the Prime Rate.

    In defense of puzzlemakers — the fine people I know have unique abilities, they’re sometimes eccentric and wacky (like me). But they’re not lazy. They’re interested in improving the craft and take it seriously.

    Thankfully, the editorial powers (Will, Mike, Rich) are tolerant and interested in advancing the art of puzzlemaking, as it relates to their specific markets.

    Finally — I’m not clear on what you mean by “bad fill.” Are you bothered by specific words?

    The old “crosswordese” debate is just that — it’s old. And it’s vague. One person’s obscurity is another person’s reality. I use an ETUI every day, and yet there are people who insist it’s a fake word. They don’t use one, they’ve never bothered to learn about folks who make and use ETUIs — therefore it doesn’t exist. Scary logic.

    That’s intolerance. Not earth-shattering if we’re only talking about puzzles. But scary, if folks like that get into positions of power. Intolerance is bad business. It’s just bad.

    I’ll take a crack at the definition of “bad fill” or “crosswordese”:

    Crosswordese (n.): The undefined vocabulary of puzzlemaking that, according to some constructors, always appears in puzzles made by other people. When the same vocabulary appears in their own puzzles, it is referred to as “fill” and is therefore justified.


    • Regarding the business opportunity, you find me the time, desire, and steady stream of good themes and I’m there.

      Certainly, the editors are the last line of defense for putting quality puzzles in the paper. But if the constructing community as a whole worked to eradicate junky fill, the overall product would become better. We don’t get paid a lot, so I hope pride in one’s work would be sufficient to make this happen.

      Bad fill is obviously going to mean different things to different people. For me, it’s partials, abbreviations never used in real life, and boring obscurity. I think your invocation of “intolerance” is a little extreme. Furthermore, I find your “definition” of crosswordese needlessly pejorative and at least as accusatory as the charge of laziness, which, by the way, is something I brought up only to dismiss.

      It seems we’ve reached the point in the conversation where my original point starts to get distorted. I never said, nor would say, that fill has to be completely perfect. What I did say is that concessions are inevitable. My beef, quite simply, is where I feel the constructor could have done better for whatever reason.

    • ‘Liz, “bad fill” is any fill that is neither a word nor known or an inferable abbr.

      For instance, Let’s say that you’re reading the the clue: {He played the guitar solo on the Beatles’ “Taxman”}, what would the possibilities be?

      You’d probably think (were you a Beatles fan) that the obvious answer is “George Harrison”.


      Aah, but what’d be your second guess?

      Perhaps recalling that Eric Claption was such a best friend to George that he (Eric) rescued his wife from him (George), and want to guess that the lead guitar/solo on “Taxman” was played by Eric Clapton.


      The guitar solo on “Taxman” was played by Paul McCartney.

      I’ve always loved that little nbit of trivia, ‘Lis’beth. And, right now, I don’t recall what brought to mind. Just shows to go ya.

  53. Mr. Krozel…
    My new blog post is “Baring my Soul”.

    Bear with me, as I’m much better at puzzles than I am at using Blogger (ugh).
    Anyhow, check my video answer to your request (My defense of the new sport of Taser Tag) at that post (scroll down), and the page that I had already been compiling,

    (BTW, I even included Dr. Fill’s epic “DOTHEBOYS” screenshot from ACPT…)

    Due to Blogger vagaries, you need to scroll down a bit in both cases to get to the goods.

  54. As a solver, I like puzzles from all days of the week, but themes are what I enjoy most. There’s a concept behind the puzzle, an idea, some wordplay, a payoff beyond filling in a nice collection of answers and clues.

    I think that’s what drives a lot of solvers, although I totally get the it’s-all-about-the-construction point of view. In fact, I can’t remember my 91-year-old mother once appreciating a theme; it’s 100 percent about solving the puzzle for her. My dad was the opposite.

    But that doesn’t mean the fill isn’t important in a themed puzzle. It’s like a supporting actor in a movie or the band backing up a singer. If the fill is good, it makes the whole experience better. If it’s O.K., at least it doesn’t detract. But if it’s bad, it undermines everything; it erodes all the good will built up by the theme.

    Beyond that, I love a great innovative theme, but not at any price. A theme that’s all special effects isn’t fun. If the fill is stretchy to boot, it’s even less fun.

    Moving to unthemed puzzles, if a grid has a low black-square count, that’s on the constructor. If he or she can pull it off, great, but if not, it’s the solver who pays the price. And the whole reason for constructing a puzzle is to entertain solvers.

    With any puzzle, you’re never going to please everyone, but clever, smart, and clean will always win the day.

  55. I feel like the guy who begged the magician to reveal the secret to his tricks, and when the magician complied the guy lost his delight for the craft. This discussion, and other facts recently uncovered about what goes on behind the constructor’s curtain, have left me more informed—and more disenchanted.

    Autofill? Guess I’m naive, but I had no idea. Sounds like the constructor’s equivalent of a solver using google, or simply hitting the ‘reveal solution’ key. I’m not a constructor, but the use of a computer to fill in the blanks seems perilously close to a crossworder’s equivalent of plagiarism. Not only that—but how can a constructor take any real pride in letting a machine do any of the creative work? If a puzzle-maker gets painted in a corner, I’d like to believe she has the integrity to retrace her steps and devise a new route to satisfaction rather than ask a Garmin to do it for her….

    Until someone—maybe ACM?—mentioned how a submitted puzzle could occasionally be so heavily gutted that the original was almost unrecognizable, I didn’t know that happened, either. I figured Will—an editor, after all—enjoyed the firepower to make a very few small corrective moves, but I presumed that once a fairly low threshold of crap fill was reached that the puzzle would simply be rejected.

    If a constructor is going to put a puzzle in front of me and cost me $2.50 (pen on newsprint—that’s right, no computers, I have to have the paper in hand), then he/she is tacitly requesting my approval—or condemnation—of the puzzle. I’ll happily provide either as I see fit. It’s not an attack on the creator, but on the creation.

    Besides, why should the constructor take it personally? Between autofill and editing, who can know how much of the puzzle is actually that person’s contribution, and how much is Will’s, and how much is that faceless computer program? I used to note the byline on especially fine puzzles, but now I discover my affection may be misplaced. How do I praise autofill?

    Maybe it’s the low payout for this stuff—the faster a constructor can produce, the higher the income, so efficiency overrules accepting full—and slower—authority over the product. That’s another thing I misunderstood. I presumed the real reward was simply seeing one’s name as the puzzle’s author. If it really is a money thing, then my disillusionment grows even deeper.

    How about we get some new classification for the truly special crossword: “This puzzle certified to be 100% organically created by the named constructor.” No autofill, no Will involvement—maybe no money changing hands! Just a pure, unadulterated, creative jewel by a craftsman demonstrating unique personal skill.

  56. As a longtime solver from the Margaret Farrar days I can, like everyone else, marvel at where crossword puzzles have come today, with regard to theme development and fill. Most of my friends and relatives (paper / book solvers) are Monday-Thursday solvers, and have never read or even know of the existence of blogs. As much as I try to avoid overused fill words, give me an “oreo” anyday to a partial or obscure word with z’s or k’s. While not speed solvers, I think most solvers want to get to the theme elements, and sometimes words that are often repetitive in crosswords are familiar “friends”, the mortar that holds together the bricks, in a way. I think if the theme is an interesting original one, and the additional longer entries (7, 8, 9 or more letters) are fresh, have a nice sound to the ear, I don’t mind the oreo or even the etui being there, especially if the cluing is pleasant. It is so difficult to ferret out new interesting theme concepts, or grid designs, that sometimes I suppose a constructor needs a little “help from his friends”, is not being lazy or unambitious, and I really think that the vast majority of solvers out there appreciate the “gimme’s” and don’t hold the constructor’s feet to the fire for them. And come to think of it, Oreo is a pretty nice sounding word, with pleasant connotations.

  57. Tyler, would you be willing to contribute to a master list of ‘banned fill’? It could be used to evaluate any puzzle with a simple summary number, indicating the total fill which appears on this list, e.g. “It was a good puzzle but it reached a 13 on the banned fill list.”

    • No. Can you imagine the arguments over what should be included or left off?

      It’s important to keep in mind that people will have different opinions on puzzles. Assigning a concrete score seems to run counter to that.

    • Quantification is the essence of evaluation. I’d be willing to do all the work, and I’d even exclude myself from the selection process. Certainly a procedure could be implemented to allow a fair decision on the inclusion or exclusion of any given example of bad fill. Of course, if this would do nothing to improve the quality of the current milieu of crosswords, then there would be not point in attempting it.

      • I’m afraid I just don’t see the point. A large number of people will inevitably disagree with the omission or inclusion of a large number of entries. When you look at bad fill, and the issue of obscurity in particular, subjectivity dwarfs objectivity.

  58. With no ultimate reference materical on construction, it occurs to me that most constructors pick up their own habits, discipline, do’s and don’ts from their personal experience. And If a fledging constructor doesn’t engage a mentor (I know of several, bless their hearts), then the blogs are the next best thing.

    Now my personal standard is that the entire grid is up for trashing until the last letter is filled in. But wasn’t always that way – used to be that the more I had invested in a grid, the more likely I was to try and hammer something in that didn’t want to be there. So my persoanl experience is that I rarely go all the way with the original grid design (but once I hit the 4th do-over I ususally shelve the project for another day).

    So that’s my musing: that perhaps some constructors stick with grids that oughta be abandoned because they have so much invested by the time they start to see the dreck …

  59. Pete Collins questioned the existence of a dicotomy between good fill and adventurous theme above. I would argue that there is no question, it’s a necessary fact. Once you’ve chosen your 3/4/5 theme entries, the fill for the rest of the puzzle is restricted by those choices and their placement in the grid, you’ve got to find fill to fit around these entries, or find new entries. The new entries are likely less appealing that the original ones (otherwise they’d be the original ones), so you’ve got to accept either a thinner theme, or accept the fill you were trying to aviod. Theme density drives the restrictions on fill, and the higher the theme density the less flexibility the constructor has in choice of fill. Andrea’s commented frequently that, at least in the NYTimes, 3 is the new 4, 5 is the new 4 for theme entries, so the escalation of theme density has to effect the alleged quality of the fill we’re discussing.

    I, personally, care little about themes on a day to day basis, but when I think of memorable puzzles I’ve liked they’re ones where the theme knocked me out. When I think of memorable puzzles I’ve particularly disliked, they were ruined by the constraints of the theme. Given the choice, as a solver, I’d skip the theme in favor of cleaner grids.

    Oh, the whole ‘I defy you to define bad fill’ argument is a red herring. Very few individual entries are intrinsicly bad fill, excepting of course random Roman Numerals, ENE/ESE/NNW…, ALER, NLER, NBAER, .. It’s the sheer number of them, their self-segregating themselves into little ghettos. OREOs are fine, as long as they’re not encased in an ETUI and stored just inside an ADIT.

    • Very well said. This illustrates why I hold a themeless’s fill to a higher standard even though the word count is lower. There are no given letters restricting things, nor is there a theme to be viewed as a mitigating factor if the fill is poor.

      • It would have been said better if I had proofed, and said “4 is the new 3, 5 is the new 4″, but thanks anyway.

  60. Hi All,
    I’ve been doing the daily times puzzle and many of the weekly “good” puzzles for about 2 years. (I also struggled with the Sunday Times puzzle in the 1980′s, with a crossword dictionary at my side.)

    Here are some of my additions to Pete’s bad fill
    > random Roman Numerals, ENE/ESE/NNW…, ALER, NLER, NBAER

    many of the 3 and 4 letter instances “abbrev for”
    winning and losing lines of tic tac toe
    obscure initialisms

    I also personally make a distinction between crosswordese that I’ve learned to expect a few times a year (e.g. SNEE, SMEE, ESME)
    and fill that seems much less infrequent and hence more obscure – ALEGAR, CRETNA, AMMA

    Where such words fall depends on your experience. When I first became a solver I was outraged by words such as ADIT, now many of them are gimme’s.

    To the list of valued blogs already mentioned I would like to add
    “An Englishman Solves American Puzzles ” – I lament the end of crosswordman’s daily comments, and recommend his “Cruciverbal Cheat Sheets” to anyone wanting to learn crosswordese as a second language.

    Finally I want to say [off the fill topic] that star ratings are useful to me. I’m not a fast solver and I use star ratings to help prioritize which of the CS and LA Times puzzles I’m going to do. On the other hand I recognize that when a puzzle gets 4′s, 3′s, 2′s and a 1 that people have different tastes and standards.

  61. This remains an interesting read, though its branching discussions seem to be taxing the limitations of a blog comment thread. Is there a Facebook group, or some other multi-thread online community, dedicated to crossword construction issues? Should there be?

  62. Perhaps I can offer a slightly different perspective on the matter, if anyone is interested. I’m also a novice crossword constructor, but my crosswords are published in Croatia. Here, instead of single daily crosswords, we have dedicated weekly magazines for them, so tens of thousands of new puzzles are published every month. There are also occasional cunstructor contests so standards for croatian crosswords are pretty much set. In recent tournaments, puzzles with stacked entries have all done well, although many of them had severe compromises in the fill. But, it was generaly tolerated because the puzzles are judged by other constructors who appreciate the effort needed to pull off something like that (current records for different sizes of crosswords are 6×14 stack, 4×20 stack and so on…). Others highly placed crosswords had huge swaths of white space (up to 8×9), or extremely low number of black squares (6 in a 19×7 puzzle, 7 in a 15×8…). Crosswords with sparkly entries and interesting letter patterns were few and far between. With help from google, it’s easier to brake records and the technical aspect had become more important than the artistic one. I admit I’m guilty of creating at least one crossword with no fresh entries at all (in order to get a 7×7 swath of white space in it). However, the current plague of our crosswords, in my opinion, aren’t the obscure entries, but the stale long entries required to pull off feats like that. Especially annoying are the long entries with periodic consonant-vowel-consonant progression, which make the crosses of the long stacks more managable. In recent months I’ve seen dozens of crosswords sharing 3 or even 4 long entries (e.g. EKONOMSKI MATERIJALIZAM (economic materialism, the single “s” being the only thing disrupting the flow), ASTRONOMSKI DALEKOZOR (astronomic telescope, again with the “s”) and so on..Also, you can notice the string SKI, the suffix of croatian adjectives, repeating in these entries. This, together with TI (suffix of our verbs), enables creation of large swaths of white space in a same way your RE- and -ER words do.

    So the greatest danger to the overall enjoyment of the solver, IMHO, comes not from the creative new efforts, but from record-braking tries that turn the rest of the crossword into an autofill task.

    As per crosswordese, there are no requirments on symmetry in our crosswords, and no word length requirements (you can have isolated and unchecked letters, although it’s considered estethically displeasing when they’re not along the edges of the grid) so you are not as constricted when cornered (har-har).

    Please feel free to share your opinion (and of course, excuse my poor english).

  63. A clarification or two on my earlier comment. Regarding constructor “laziness,” I mostly meant that in response to years of queries at cruciverb that have invariably been prefaced by something like this: “I have this one trouble spot and ????? (fill-in-the-blank with whatever you deem to be a stinky piece of fill) will save me from ripping out half the grid.” My blanket response to that is “you may want to try ripping out that part of the grid first.” That’s the phenomenon I’m referring to – trying to force through certain things because, perhaps, you’d gotten to that point relatively easily and are now presented with a constructing challenge that will most easily be solved by backing into a section of fill you know is not great, or, worse, is just unfair or intolerable to many. I know those bits will more often than not come into play for all of the reasons that have been pointed out here – but I would feel better as a solver knowing that pains were taken by the constructor before having resorted to certain bits of fill.

    Some other thoughts on the terrific points being brought up here.

    Trade-offs/balancing – I think rather than make any attempt to have a list of “crosswordese” and/or mostly unacceptable abbreviations, names, etc., what might be enlightening at minimum would be to present two or three different fill options for a corner and see what comments people make, and why they would vote for one being better than the others.

    I’d say the context is always a main consideration. Choices among fill possibilities are impacted by the intended degree of difficulty of the puzzle; theme, long fill and previously filled areas of the puzzle (did you already have an UP, IN, ON, THE or whatever before the trouble spot? Do you have JIM in one corner and now can’t use JAMES?); the constructor’s own reference and perceived reference of the solver, and convention. Debating different possibilities for an entire section of a fill brings up all these points and more, and it would be much more informative than just deciding which is worse between ETUI and SNEE.

    Vocabulary words and individual blind spots – Will’s point, that has been echoed by Liz and others, that certain words can be tough for some and probably new to many, but whether you’ve “learned it from crosswords” or not does not preclude the entry from having appeared in the newspaper, books, dictionaries and notions stores, say – perhaps it’s just new to you and you’ve learned something. Though they mostly involve proper names, my personal ETUIs – “crosswordese” that rate as regular old words -are many. I can’t change the fact that I went with my family to a mining museum years ago and, there, at the entrance to the mine, was emblazoned “ADIT” – it was my “Slumdog Millionaire” moment, how I learned this word.

    Names and sports – I don’t think this has come up much but, wow, are there a lot of solvers who dislike seeing names and anything to do with sports! Some of us have fond memories of watching the ALOU brothers play baseball. ASHE means a lot more to me than a stadium name or a list of major tournaments the guy won. On the I-know-it-well but, yeah, it’s probably too obscure to perpetuate level, my father worked in an acting troupe with ZASU Pitts and my mother dated CLU Gulager – I’ve known those names well before having touched a crossword. So, though my soft spot for names needs to be squelched occasionally I do love to include an unusual one in favor of an unispired word. Still, trying to be objective – in other words “fair to the solver” – a dull-ish word might indeed be a better choice than a snappy name.

    I could go on but will instead go seek a support group of any kind – I mean, if it’s AA or whatever, and they’re truly supportive, they’ll let me talk about crosswords, right?

  64. The concept of “bad fill” brings to mind Supreme Court Justice Potter Stewart’s most famous quote. In the Jacobellis v. Ohio case in 1964, he wrote that hard-core pornography is hard to define, but “I know it when I see it.”

  65. So now I’m doing the last puzzle (#144) in “word.”, Natan Last’s hip crossword book. Guess who shows up at 32A? ZASU Pitts! Crossing one of the Zs of MS FRIZZLE, of the Magic School Bus. Thought you’d enjoy that, Tony O. As I was sort of saying before, we all need facts from each other’s generations for these crossword puzzles to happen. Kumbaya.

  66. I think my philosophy about valuing theme or fill comes down to one piece of advice: Amaze your solver.

    What you do should blow people away, or you shouldn’t do it. Now, for different people, that means different things. Some people can stack long entries and amaze solvers, and some people can create interesting themes and amaze solvers, and some people can weave interesting words and amaze solvers, and some people cannot do any of those things. If a creator is in any of the former categories, they should continue to do those things. If a creator is in the latter categories, they should strive for one of those things, but probably not all at once. That is where it usually goes horribly wrong.


  67. Im new-ish to puzzle construction and just wanted to thank all of you for the rich, thoughtful, non-terrifying discussion. This was perfect timing for me.

    Of course, as someone trying to learn how to write puzzles, I’m left with more questions than answers: like, how do you avoid bad fill (which maybe we can agree becomes an objectively recognizable standard at some point, personal interests aside)? And: what is it that makes a good puzzle? I never believed that every puzzle would find in me a worthy and suitable solver; I’m not promiscuous in my solving in that way ( though maybe I wish I were). I’ve tended to think that oftentimes, a given puzzle simply isn’t “for me” — either because I find the fill, skillful or not, beyond my range of interests, or beyond my ability, or whatever. (the recent STARGAZE would be an example of something I found completely invigorating and refreshing; in can think of a few others exemplifying the reverse.)

    I’ve been sitting at my computer trying to hammer a few of these out and thinking, who ought these to be for? And who are they /actually/ for? And does answering this help me find my way through the pitfalls of construction? And does how they’re received depend on this distinction? And is all subpar fill created equally?

    • You get better at filling puzzles by making more puzzles. You gain a greater appreciation for what’s good, what’s not so good but can be forgiven, and what should definitely be deep-sixed. As for what makes bad fill, that’s part of the issue, but I think there’s manifestly good fill, mainfestly bad fill, and some stuff in the middle that neither sparkles nor stinks and generates debate. It’s certainly not all equal.

      What makes a good puzzle is a more complex question still; you’ll never please everybody. The best you can hope for is to boost your odds of making a well-received puzzle by putting together a solid theme and surrounding it with fill you judge to be good.