Why is 17-Across special?

What is the most common clue in NYT crosswords? We were curious, so we wrote some code to figure that out.

Before running the code, I tried to guess the results. We’d probably be looking for clues pointing to common, useful answers, i.e., short answers with lots of vowels. OREO seems like a good target but there are a million ways to clue that cookie. Two categories of clues seem promising.

Answers with only one clue

Some useful answer words have only one reasonable clue:

• ALAI is almost always [Jai ___]
• IRAE is usually [“Dies ___”]
• ASTO is often [Concerning] or [Regarding]
• EERO is usually [Architect Saarinen]
• AER is almost always [___ Lingus]
• CRO has to be something about [___-Magnon]

Clues with multiple answers

Perhaps the most common clue can point to multiple common answers:

• [Pro ___] might be TEM or RATA or FORMA or BONO…
• [Mauna ___] could be LOA or KEA
• [Santa ___] could be anything

On a recent Facebook post, I invited people to guess the most common clue. Some wag joked that it’s probably something like, “See 1-Across.” Joe DiPietro immediately chimed in, “probably more like ‘See 17-Across'” which, by now, you’ve guessed, is correct. That clue has appeared 140 times in the Shortz Era, and over 100 times before.

Why 17-Across? That seems random.

Let’s do the math

Clues that refer to other clues imply there’s a theme, and that means an early-week puzzle. Themed crosswords generally have three, or four, or five long answers running horizontally that have some sort of connection. It might be a quip or a quote, or some other relationship described in the first theme answer. Later clues can point back to that first one.

You can usually identify theme answers because they’re the longest horizontal ones in the grid. Here’s a typical early-week 15×15 crossword grid with four long answers.

There’s not a lot of white space. Wide-open grids come later in the themeless Fridays and Saturdays. Answers outside of the theme will be shorter, so the top row typically has two black squares.

That means the second row generally has two black squares too because the first long theme answer is usually at the beginning of the third row. (Sometimes it’s the fourth row. Why not have your first long theme answer in the top or second row? You like puzzles, so I’ll let you think about it. I’ll explain at the end of this post.)

Every square in the top row has to be numbered because each has an associated Down answer. Two black squares means that we use grid numbers 1 to 13 up top. The three Across answers in the second row use up 14, 15, and 16. That means the first long answer, the one that other theme answers can refer to, is, yes, 17-Across.

That odd mathematical fact pushes [See 17-Across] to the top of our Common Clues list. As expected, [Jai ___], [Pro ___], and [Mauna ___] are next.

There are a few more surprises. If you’re curious, here’s the list of all the most common NYT clues in the Shortz Era.

The third puzzle row puzzle

Did you figure out why the first theme answer can’t be in the top or second row?

Every white square is a part of two answers, one Across and one Down. Every white square in the top row starts a Down answer, and in American crosswords every answer has to be at least three letters long. For that to work, every long answer in the top row must have an answer immediately below it that’s as long or longer. For the same reason, every white square in the second row must have a white square below it, meaning another long Across answer word. But, we want the longest Across words to all be theme answers. So, as mathematicians say, we have a contradiction. QED.

But crosswords aren’t math, and if you do enough of them you’ll find some belie this general rule. Now, when you find one, you can brag to your friends about why they’re rare.

Since you aced that puzzle, here’s a final question: Is it ever possible for a crossword to have a clue for 2-Across?