American Crossword Puzzle Tournament

Since I haven't been keeping up with the trees here, I thought I'd derail myself to hype the main event of my vacation, which begins next week. Starting Feb 20, I will be competing in the American Crossword Puzzle Tournament in Brooklyn, New York. This will be my second time in the contest, which was created and continues to be primarily the work of Will Shortz, editor of the New York Times crossword puzzle. If last year was any indication, he culls out the best of the best for this event, presenting seven puzzles by constructors well known to most of us who are geeky enough to participate in the event. I think the 7 contest puzzles are meant to approximate the range of difficulty that the NYT uses, starting with fairly easy puzzles on Monday, then getting progressively harder till Saturday, with a moderately challenging but oversized Sunday.

The puzzles last year didn't quite adhere to the same difficulty range. Of course, I am speaking from only one year of experience, so take this all with a grain of salt. But I would say the easiest puzzle was probably not quite as easy as the typical Monday NYT; the middle puzzles were probably all Tuesday-Thursday level. Instead of a pair of killers, as is typical on Fridays and Saturdays, the tournament has only the dreaded Puzzle 5 to sort out the top solvers from the rest of us schmoes--it was hard, but I would say not quite as bad as some of the NYT weekend offerings. As in the newspaper, some puzzles have a theme that may give extra hints to solvers, while #5 was themeless last year (I can't remember for sure, but I think the rest all had some kind of theme).

The hotel sets aside a huge room for the contest. Long, cafeteria-style tables are filled with contestants, who get one puzzle at a time. Everyone starts together, and a timer begins to tick off the minutes. Proctors around the room watch for hands; when you finish, you turn in your solution, the time is noted, and you may leave the room. The next puzzle begins about half an hour after the deadline for the previous one. In between, people talk about the puzzle; play other games; and shop in the lobby for puzzle books, games, and (at least last year) crossword-themed art. There's a lunch break, which means a lot of impromptu groups walking to one nearby restaurant or another. I suspect there are elite, highly competitive solvers who go to their hotel rooms to stay focused and eat special crossword-solving diets and such, but I didn't meet any of them. Saturday night (and Friday night) there are informal games and social activities, then there's puzzle #7 Sunday morning, and the live finals Sunday afternoon. The top three solvers from the A, B and C divisions all work a final puzzle on a stage with live commentary (wearing headphones to drown out any help from the audience or commentators). They all have the same grid and solution, but three different sets of clues makes each round significantly harder than the previous one(s). It was very exciting to see the A final especially--this video shows eventual champion Tyler Hinman as he agonizes over the last couple of squares, and finally figures it out to win with the only correct solution (again--accuracy trumps speed).

My skill level puts me solidly among the schmoes. I can do well on the early-week puzzles, and pretty quickly by most people's standards, though I'm not nearly as fast as the best. As of last year, I'm not sure I had ever successfully completed a Friday or Saturday puzzle. I can now say I definitely have (a few times), but I am far from the level where I expect to be able to do so on a given weekend. After four puzzles, I think I was near the top quarter of the field (someone who really cared about this could go to the site linked above, where all the results are listed and broken down in more ways than I care to list, to calculate my actual position, but I'm satisfied to guess). Puzzle five dropped me down to somewhere between the 35th and 40th percentiles. I had a good #6, and then missed one letter on Sunday to finish 244th out of 674 (MOAPO!!).

Those 674 people came from all over the country and beyond. The majority were from New York, of course, and most of the others were from nearby states, but plenty of us made the trip from across the US. I don't have my contestant list, which included brief bios, but the results show a handful of "foreign" competitors (which I'm sure includes at least some Canadians--I don't really know how exotic we got). Being in Texas, I was torn between entering the West division or the South; I competed in the West, which meant I was in the same region as the eventual winner (for his fifth title), Tyler Hinman. Not that in mattered--there was a similarly elite solver, Trip Payne, in the Southern division, as well (This year I saw a handy map -- don't know if it's new or if I missed it last time -- that puts TX in the South.). No, like most of the contestants, I knew in advance I was not going to win in any of the various subdivisions of the field (well, I was hoping to squeak out a top place among the rookies, but knew it wasn't likely).

The primary divisions are the skill levels: A through E, with most of us falling into the C division. To get into B or A, you have to have competed in the past and done exceptionally well (top 20% or better, or top 3 in the lower division). To get into D or E, you have to have done worse than at least 40% of the field in an earlier contest. As a rookie, I was automatically a C. If I'm reading the statistics correctly, there were 301 of us last year; of those, I finished 108. That's just about what I expect, since I solve the NYT puzzle everyday online and the applet there tracks times--I routinely finish fairly close to the line between the top two thirds (though in good weeks I can make the top 20-25% on Monday and Tuesday, and I often don't finish Fri or Sat at all--or cheat with google to do so).

A word about that--scoring is pretty complex when you first try to read the rules, but the gist is, you get points for each correct answer, a bonus for each completely accurate puzzle, and more bonuses for each minute you finish ahead of the allowed time limit. I think most people finished most of the puzzles in the time allowed, although puzzle 5 ended with most of the room still working. So speed is good, but only if you're accurate. Many people cursed their own failure to check a cross that would have revealed an obvious mistake and saved over 150 points. Again, all the details are at the ACPT site.

By finishing behind the top 20% but ahead of the bottom 60%, I kept myself firmly in the C division for this year. I would have needed another 1400 points or so to make the C finals last time--just about what I would have gotten if I had solved correctly on #5 and #7. So it isn't impossible to imagine. But I won't be holding my breath. So why am I spending all this money to do it again?

It's fun. Not only are these some of the best puzzles anywhere, but nowhere else can I finish a puzzle with a funny theme and talk about it with others who shared my enjoyment (except in the virtual world). If I mention a puzzle to most of my friends, they smile politely and generally try not to look bored. At ACPT, everyone is talking between puzzles about the good jokes or the clever clues. Some people bring puzzles they have made themselves and leave piles of them in the lobby for those of us who need more than six a day. Some dress in crossword-themed clothing. All are receptive to an invitation for lunch or a trip to the bar after the event is over. Most are like me, fully intending to finish well below the winners, and they simply do not care. The rankings are virtually irrelevant to all but a few of us. It's about the camaraderie and the shared passion for puzzling.

So if you made it through this whole write-up, the ACPT just might be for you. See you in Brooklyn!