Whenever you open crossword construction software to a new project, chances are that it’ll automatically give you a 15×15 grid. The reason is simple: Statistically, you probably are making a 15×15 grid, since it’s been the standard crossword size for decades.
But just because it’s the standard doesn’t mean it’s the rule. There are many exceptions, including Jonathan Raskin’s Wednesday, June 17, 2026 puzzle in the New York Times, which adds a column of width to create a 16×15 grid.
This aspect by itself isn’t super noteworthy: Last year in 2025, the NYT published 38 “irregular-sized” grids (i.e. grids that don’t adhere to the standard 15×15 weekday / 21×21 Sunday sizes). They’re on pace to surpass that number this year – at the time of writing, they’re at 23. (According to XWordInfo.)
But I was struck by this puzzle, and not only because of the irregular size. What interested me was that on the surface, this puzzle really feels like it could have been a standard 15×15. Which makes the choice for an irregular grid feel intentional.
Ultimately, I’d argue the constructor’s (in my mind) deliberate choice of 16×15 grid provides an interesting perspective on the question of standard vs. irregular-sized grids. And it’s made me reconsider my constructing practices and values….
Spoilers (click for solving link):
New York Times: Wednesday, June 17, 2026 by Jonathan Raskin
The Puzzle in Question
The puzzle’s trick is explained by the revealer in the center row: [Casino maneuvers carried out three times in this puzzle?] for RIFFLE SHUFFLES. (The horizontal entries with some shaded squares remain valid words in both their “unshuffled” and “shuffled” forms.)
You could argue that the middle RIFFLE SHUFFLES is what necessitates a 16×15 grid. Since the revealer is 14 letters long (an even number of letters), you need a grid with an even-numbered width in order to maintain rotational symmetry.
But I’d say that this puzzle has an equally valid, alternate 15×15 version with the middle revealer RIFFLE SHUFFLE, singular. For this variation, the entry is now an odd-numbered 13-letter word, which would work fine with a standard 15 columns.
In my opinion, the published and alternative concepts create essentially no difference from a solver perspective. You’d just have to tweak the cluing slightly from plural to singular: [Casino maneuvers…] to [Casino maneuver…]. Plus, the other theme entries are all 8 letters, which theoretically should fit fine in rows of both 15 and 16 squares.
Rethinking Standard Practices
So why opt for a grid that’s 16×15 instead of the “standard” size?
My guess is that this decision came down to the grid-filling process. The quantity of theme entries means that there’s a lot of unchangeable word-stuff that must be included. These entries naturally restrict the grid in terms of 1) where you can put black squares, and 2) limiting word options when you’re forced to cross two or more theme entries.
The wider grid eases that pressure a little bit – reducing possible instances where vertical entries are forced to cross multiple themers. Intuitively it makes sense too: More space means a lower proportion of theme- to non-theme entries, meaning a decrease in restraints.
Take a look at these two grid proposals. The first grid is a possible 15×15 version of this puzzle, compared to a stripped-down version of the published 16×15 grid:
Even ignoring the specific letters involved, the hypothetical 15×15 grid (Exhibit A) has a very constrained middle column. It crosses through all 6 theme entries and the revealer – as a result, options for placing black squares or choosing words for that column are limited. (For example, you’re basically forced to include the entry SST – for Super-Sonic Transport – in the center.)
Meanwhile, these problems are mitigated in the published 16×15 grid (Exhibit B). The added column offers just enough extra space to avoid a middle-column situation like in the 15×15. As a result, I’d anticipate it’s significantly easier to fill compared to the “standard” grid.
In theory, this puzzle could have technically worked as a standard 15×15. But practically, I think the added column creates a more adaptable and thus a likely stronger puzzle.
Final Thoughts & “Justification”
So what’s there to learn from this discussion? At the very least, I think this size-changing business is another tool to improve the grids we make, especially when they’re jam-packed with theme content and feeling unwieldy.
This has got me thinking a lot about the concept of justification. A lot of irregular-shaped grids have a pretty obvious “reason” for not being a 15×15 – such as when the middle themer has an even number of letters (like in this case), or when a key entry is 16 letters long. Sometimes there’s even softer justification, such as making it so a final 12-letter revealer can fit more comfortably on the 3rd-to-bottom row.
Still, for many draft 15×15 puzzles without a similar concrete “justification,” the irregular-sized grid might fly under the radar as an option. In my personal experience, I rarely consider adding an extra column unless the theme absolutely necessitates it. Instead, I’ll just keep grinding away at my awkward 15×15 grid, slowly lowering my standards while also hoping for a miracle.
I think that’s an oversight, and unnecessary. Because what has a bigger impact on the solving experience: good fill, or a grid that’s perfectly square? After all, the only reason I notice irregular-sized grids to begin with is because I’m in too deep with this crossword stuff; even then I can’t say it affects my experience or enjoyment of a puzzle.
In many ways, the 15×15 still reigns supreme. Many venues only accept standard-size puzzles (though this trend is changing). Still, different editors may have different opinions on the matter, and maybe some feel like there’s something uncouth about seemingly “unjustified” 16x15s.
Ultimately, I think it comes down to trade-offs. In my post about 2-letter words, I wrote about how certain benefits (e.g. enabling more ambitious concepts) might be weighed against resulting detractors (e.g. the solver finding certain grid elements uncool/un-fun). I don’t even think that what I’m advocating for here is nearly as radical as including 2-letter words. The irregular-sized grid is already being published with some frequency, and solvers and editors seem fine with it. Any perceived inelegance feels mostly negligible.
I doubt that adding an extra column is always necessary – people are still making banger 15x15s, after all! – but it’s perhaps worth a try in certain scenarios. The NYT puzzle is just one example of an irregular grid being useful despite not being strictly required, and I imagine that similar situations will crop up during construction if we’re open to them. (I can even imagine that there are situations where removing a column makes a grid easier to fill, though those feel like they’d be rarer.)
So do irregular grids need justification? I’d say yes – but as far as I’m concerned, good fill and a better final puzzle is justification enough.
Do you have additional ideas? Things I missed? Feel free to comment your thoughts! Or email me at thegridalchemist@gmail.com.
And be sure to subscribe below to read the next post as soon as it comes out!