Amy’s ACPT wrap-up (no spoilers)

Hi! It’s me, Amy, back home from New York and the American Crossword Puzzle Tournament. I went into the weekend feeling like this would be my last year competing at ACPT—my heart just wasn’t in it, I have no interest in “training” anymore, and I wasn’t even particularly looking forward to the tournament puzzles. I really didn’t care how well I did, and had no goal—whereas in past years my various goals were try to make the top 10, try to make the finals, and try to stay in the top 15 or 20.

Before the start of the tournament, I chatted with Stella Zawistowski in the Marriott lobby. Like me, she’d had a rough time in a recent tournament, with mistakes taking us both out of the running and (gasp!) out of the top 20. She said her goal this year was simply to be “clean” (finish all seven competition puzzles with no errors). So I adopted that goal myself. I’d had a few good years of clean solves, and I wanted to reclaim that mantle. I wanted to have nothing to feel lousy about afterwards, but had no idea how high I’d place if I approached the puzzles with non-speed goal (nor did I care).

I solved each puzzle in exactly as much time as it took me to solve it. No speed-demon shortcuts (such as looking at the next clue while still writing the current answer) attempted, no looking at the clock while solving, but no deliberate slow-down. When I finished a puzzle, I’d look at the clock and decide if I wanted to raise my hand right away or take a little time to check for mistakes.

I solved Lynn Lempel’s 15×15 puzzle #1 in 3:29 and so had up to 31 seconds to check my work. No blanks, no implausible-looking answers? Good to go. I headed to the hallway outside the ACPT ballroom, where many of the 11 2-minutes-and-change solvers and some of the 3-minutes crowd had already gathered. Unlike past years when I had specific aspirations, I felt zero pressure from knowing that I was already 25 points behind those 11 people.

Next up was puzzle #2 by Mike Shenk, a 17×17. It took me 5:40 to solve this one, and at least seven solvers beat me by 1 or 2 minutes, putting me 75 points behind the leaders. I enjoyed the theme and, like all of the tournament puzzles, Mike’s crossword had remarkably smooth fill and terrific clues.

Puzzle #3 was a 19×19 challenger from Brendan Emmett Quigley. Once again, I had a good time working the puzzle, and I finished in 10:02. If I’d been striving for a top-10 finish, it would have irked me to miss the 9-minutes-and-change echelon by a mere 2 seconds. Didn’t bother me, I swear. (Besides, if anything would be irksome, it would be that Dan Feyer finished a whopping 5 minutes ahead of me. Five!) I took a little time to glance over the grid, but not much because I wanted to get a prompt start on lunch. (Joon Pahk had noted that if one doesn’t wait until puzzle #3’s 30-minute session is over but instead speed-solves and hightails it out the door, it’s a lot easier to enjoy lunch and not have to race back to make the 2:30 start time for the afternoon session.)

A mathematician, an entomologist, and a physicist walked into a Yemeni restaurant … and ate lunch with me. Terrible punch line but the joke’s got a great set-up, am I right? My lunchmates included the brilliant math guy Kiran Kedlaya, entomology grad student David Plotkin, and Joon the college physics teacher. Good food. And speaking of Yemen, at brunch on Sunday (with my PuzzleSocial/Daily Celebrity Crossword colleagues), sitting next to possibly the world’s only other medical editor/crossword pro, Adam Cohen—I discovered that his name anagrams to ADEN MOCHA, which longtime crossworders know are both port cities in Yemen. It was a Yemeni kind of weekend.

After Saturday’s lunch, we launched back into puzzledom with Ian Livengood’s 15×15 #4. This one was a little harder than #1, as usual, and it took me 4:08 to wrestle it to the ground. Add another two minutes of lag behind the leaders.

Puzzle #5, the one with the fearsome reputation, was perhaps less soul-crushing than the typical #5. I enjoyed the hell out of solving this Patrick Blindauer creation for 10:17 of the 30:00 allotted for a more difficult 17×17. The top finishers put this puzzle to bed 4 minutes faster than I did, in 6 minutes and change. It’s really quite impressive, that fierce speed on puzzles of every difficulty level—the competitors who trounced me on multiple puzzles are Dan Feyer, Tyler Hinman, Anne Erdmann, Francis Heaney, Stella Zawistowski (who ended up with an impressive 5th-place finish thanks to her clean solves), Kiran Kedlaya, Joon Pahk, Al Sanders, Jon Delfin, Ellen Ripstein, David Plotkin, Howard Barkin, and Erik Agard. These stellar solvers include people in their teens, 20s, 30s, 40s, 50s, and early 60s—isn’t that age span neat?

We closed out the Saturday solving day with Liz Gorski’s puzzle #6, a spacious 19×19 that was markedly easier than BEQ’s 19×19. It flew by me in 6:45, which was unimpressive compared to the 3:something posted by Dan Feyer. Often there are several people tied for first place on a given puzzle, but Dan was alone at the top with his 2020 points on this puppy. It bears noting that Dan was also the sole 5:something finisher on puzzle #3. The bigger the grid, the wider Dan’s lead.

During dinner that evening, the scores had been posted for, if memory serves, the first five puzzles. And puzzle #6’s scores were up later in the evening. Massive thanks to Matt Ginsberg, who engineered the simple system of marking incorrect squares with a highlighter, scanning the puzzles in, and letting computers do the work of totaling up the scores and entering them into the score database. Before Matt brought Will his automated system, the scoring process was cumbersome and the judges stayed up late into the night. Contestants didn’t know the Saturday standings until a ream of paper was mounted on a wall Sunday morning. Matt’s system lets anyone with internet access check scores and view scans of their puzzles on Saturday, allowing ample time to catch—and fix—scoring errors. It’s brilliant, I tell you. Vastly improves the experience for ACPT competitors. Anyway, for most of the evening, I was listed as being in 13th place, which was exciting since I’d placed 25th last year. Late at night, a puzzle #3 error bumped speedy David Plotkin down a few notches and put me at 12th.

Saturday night’s main event was Matt Ginsberg’s presentation on his other brilliant crosswordy technical creation, Dr. Fill. Dr. Fill is an AI (artificial intelligence) doodad that solves most crosswords in a fraction of the time it takes Dan Feyer. Dr. Fill used to get tripped up by things like themes with backwards entries (as seen in an ACPT puzzle last year) or nonsensical groupings of words. Matt adjusted the algorithms so that Dr. Fill would prefer seemingly unrelated real words over unrecognizable words, be able to handle words spelled backwards, and whatever else seemed to trip up Dr. Fill in. Dr. Fill also reads Wikipedia to amass words that relate to one another. But while Dr. Fill did better this year than last year, Dr. Fill did not beat Dan (or me, or a bunch of people). I forget what Dr. Fill’s finish was—does anyone recall the number? I think Dr. Fill is a B Division solver, so it’s mighty good but not perfect. I imagine that Dr. Fill would plow through the typical Monday-through-Wednesday NYT puzzle in under a minute without errors, but tricky themes are easier for humans to grasp than for automated AI tools. (Apologies to Matt for any misrepresentations of his nifty technology.)

Sunday morning was brutal. No, Patrick Berry’s 21×21 puzzle #7 wasn’t all that difficult—it was daylight saving time lopping off an hour of sleep, which was especially unwelcome for solvers who’d traveled from time zones to the west and had already lost an hour. The puzzle took me 9:52. Interestingly, this time Dan did not hold or share the lead on a puzzle. Anne Erdmann made up 1 minute of her behindness by finishing this one 3 minutes ahead of me and 1 minute ahead of Dan.

It took a couple hours and an intervening talent show before the post-puzzle-#7 standings were announced for the top three A, B, and C divisional finalists and posted for everyone. The World’s Nicest Guy, Howard Barkin, had an error in puzzle #7, and you hate to see that happen because he is really just the most incredibly genial and generous spirit. So that bumped Howard down several notches and allowed me to park myself in 11th place. Whoo-hoo! (Note: This is merely an 11th-placer’s surprised and grateful “whoo-hoo” and not the slightest bit of gloating over Howard’s misfortune. It really is impossible to wish Howard anything but the best. Although I do need to chide you, Howard, for never getting around to showing off pictures of your kid. We really did all want to see those!)

Kevin Der’s finals puzzle was a terrific 15×15 themeless. Dan whipped through it, though not so fast because it was a tough puzzle. Anne was close behind, and she was tickled pink to finish second rather than third (the three two previous ACPTs, the A finals finish was Dan in first, Tyler second, and Anne third; in 2010, it was Dan, Howard, and Anne). Tyler struggled with some tough clues and the various traps Kevin had laid. Much of the audience began solving the finals puzzle in their chairs during the C finals. It’s believed that Joon had the fastest A-clues solve, at 6:32. I finished in about 7:50. Up on stage in the A finals, it took Dan over 10 minutes—and though it takes longer to solve a puzzle on the whiteboard in front of 500 people than to solve on paper with no one looking, I think Joon and I both would have had a good shot at the championship if only we had managed to be fast enough on puzzles #1-#7 to actually qualify for the finals. (Ay, there’s the rub. Me, I’m never going to be fast enough on the themed puzzles. Now, Howard, Al, Kiran, and Francis have all been in the finals, pre-Feyer. Jon and Ellen have both won. I think Stella, Joon, David, and Erik may all have the sheer speed to make it to the finals, too. And rookie Glen Ryan finished an impressive 16th—so he’s one to watch in the coming years too.) But apparently it would take Dan, Tyler, and Anne making atypical mistakes, being atypically slow, or not competing at all in order to make space in the A finals for any of these other contenders. These three have been ridiculously dominant since 2010.)

So by the time the tournament was over, I was (1) delighted to have been “clean,” as were the other 60 people with perfect papers listed here; (2) pleased as punch to have picked up a trophy for 2nd place in the Midwest region, behind fellow Illinoisan Anne; (3) happy to have finished 11th without giving up my lackadaisical, go-with-the-flow feeling; and (4) utterly exhausted from my standard 4.5-hours-of-sleep-a-night ACPT regimen. I was also recharged by the delightful company of hundreds of smart, interesting, and hilarious people throughout the weekend and the hours spent with dear friends I don’t see nearly often enough.

You know what? I hardly spent any time eyeballing my iPhone (aside from when the standings were posted). I credit that to the uncanny ability of ACPTers to keep others entertained and engaged. If you’ve never attended the ACPT but you like crosswords and you like people, I do encourage you to give it a whirl.

And if you’ve never solved any of the ACPT crosswords, treat yourself to a set of tournament puzzles for $20. Yes, I know $20 would buy you three crossword books and these are only 8 puzzles. But they’re terrific, and they just might whet your interest in attending the ACPT next year (March 7-9, 2014!). Oh! And if you’re curious to know how you’d place at the tournament, that same $20 buys you the opportunity to mail in your solutions and solving times and receive your points total and hypothetical ranking. I did the “play by mail” option in 2004 and resolved to attend in 2005—and haven’t missed a year since.

A warm “Thank you!” to ACPT organizer Will Shortz; his tireless compadre, Helene Hovanec, a logistical whiz; techno wizard Matt Ginsberg with his scoring system; this year’s gifted eight-pack of constructors; all of the hard-working judges and officials who volunteer their time (they’re listed at the bottom of this page); and all the old and new friends who made the non-puzzle times so much fun. And congratulations to everyone who won a trophy, who met a particular goal (many people try to beat a target person who has no idea, or strive for improving their percentile in the rankings, or shoot for a round number like top 100 or top 300), or who just plain showed up and gave it a go. (This is no knock on the people who couldn’t compete because of illness, injury, travel snafus, or lack of money. Hopefully you’ll make it to ACPT in 2014!) I hope the rest of you had as splendid a time as I did. I’ll be back, I’ll enter the competition again, and I hope to retain a goal of finishing clean rather than hitting a particular spot in the rankings.