New York Times, Friday, July 4, 2014

I pity the poor fools who have puzzles near Patrick Berry. (Cue the sad violins.) Some thoughtful readers have told me that they don't like the fact that I pick a Puzzle of the Week, and I appreciate that feedback. But 1.) I like pointing out fantastic work and 2.) that's what some (many?) daily solvers tend to do anyway. For me, it's a good reminder that there are other people out there with much, much better construction skills than me, and if I want to be one of the greats, I need to keep working at it by studying, practicing, improving.

This PB was no different, giving me such unadulterated pleasure. So instead of qualitatively analyzing the puzzle as per my usual, I'm going to do something different: attempting to QUANTIFY why this work is so good.

People often ask me how they can get a themeless puzzle into the NYT, so I've given this a lot of thought. I've come up with a formula that I'll revise and evolve over time, hopefully keeping it simple enough for the non-mathy types. As a finance guy most recently, I liken the evaluation process to the decision whether or not to acquire a company. You buy something for its ASSETS, ignore the neutral stuff, and discount for its LIABILITIES. You can then put a price on ASSETS minus LIABILITIES, yeah? (Roughly.) For me, I think the odds of an acceptance become high when:

1. LIABILITIES < 5 and
2. ASSETS minus LIABILITIES > 10.

What do I mean by ASSETS? Stuff that sings. This is subjective, of course, but here's my assessment of the snappy answers Patrick provides us today, each of which I'll count as one point each:

• ON A LEASH
• IM BUSY
• AGRICOLA
• MARGIN OF ERROR
• BEFORE I FORGET
• BATTLE SCARRED
• I ROBOT
• IRISH PUB
• PET PEEVE
• BATTER UP
• VIES FOR
• HAVE A NICE TRIP
• TURNED RED
• ESCARGOT
• RYE BEER

And the liabilities? Things like partials, abbreviations, esoteric foreign words, pluralized names, etc. Each one will count as one point, except for "puzzle-killers," ug-ug-ugly answers which effectively take a puzzle out of consideration all by itself (RSI, for example, which killed one of my themeless submissions). Here's my assessment of Patrick's liabilities today:

• (insert sound of crickets)

The final count: ASSETS = 15, LIABILITIES = 0. So, Patrick meets the first criteria with flying colors. And the second criteria? ASSETS minus LIABILITIES = 15. As an analyst, I'd put a STRONG BUY recommendation on this one. (Never mind the fact that there's no price already set, you smart-aleck broker/analyst types.)

Will, if you're reading this, perhaps you could comment? Am I close in my assessment methodology or way off?

It's a thing of beauty, especially considering it's a wide-open 66-worder. (That's another point in the ASSETS column, actually.) And the cluing for IRISH PUB, ESCARGOT, BIPED, POT... For all those constructors looking to get published in the NYT, I'd suggest studying this one in detail. Try deconstructing and reconstructing it to see what you can learn through the process. Many of the great artists copied the masters for years before finally coming into their own, and that process was key to their emergence, right? Well done, Patrick, another beauty from the master.

Hi Jeff,

While I don't award "points" to crossword entries (as in your example), your general way of assessing crossword quality is the same as mine.

When I look at any crossword submission, which I prefer to do on paper, I mark up the answer grid — a check mark for an entry I like (on rare occasions two checks for one I really, really like), a minus for an entry I don't, and an "x" for one that's wrong or truly awful. The size of the mark indicates the depth of my feeling about it, so a bit of crosswordese might get a small minus, a stupid obscurity a big one.

When I'm done, I pull back and look at the puzzle overall to make a decision. An "x," of course, almost always means rejection. If there are no x's, then I judge whether there are enough entries that excite me to outweigh the ones that don't.

You're right that too many minuses will sink a puzzle even if there's lots of good stuff. I don't set an arbitrary limit on weak entries, but more than five, especially if they're bad, will put the puzzle in dangerous territory.

The bar for acceptance keeps getting raised. Many of the puzzles I published 10 years ago wouldn't be accepted today. Even marginal yeses of a year or two ago might be nos now.

That said, to me crossword quality is a matter of balance. Unlike some other editors and commenters on the crossword blogs, I will accept some crosswordese and obscurity in a puzzle (assuming it's fairly crossed and clued), if the upside is strong enough. Sometimes I think it's worth having a subpar entry in order to set up other material that's great. The greater the great stuff, a little more subpar there can be. Squeaky clean isn't always best.

My bottom line is that a puzzle should be judged as a whole, as ordinary solvers judge puzzles, not on just the least appealing entries.