I’ll save my final rank for later in the post. I am starting to see a pattern where apparently Ulrich and I make these very fun to watch (obviously not intentionally), though this time we had help.
I can’t do my usual detailed per-puzzle rundowns because most of what I had to say for each one was about what was happening around me and where my standing was. This year, we sat at a walled-off booth, solved puzzles on a projector, and had little to no information about how things were around us. Even if it does make this harder to write, I much prefer the lack of distracting information. I had tried to refrain from looking around me too much even when it was possible in previous years.
The one set I did get to watch was 7-10, featuring Bram in the lead with Zoltan Horvath, Kota Morinshi, and Will Blatt (teammate!) trying to make up a time disadvantage. Only the first to finish would advance. The puzzles were Slitherlink, Chaos (a chaotic number fill-in that lives up to its name and that no one really likes), and Akari, in that order. The Slitherlink at the start pretty much decided it. All the competitors had trouble with it, and the eraser came out a lot. Bram was unfortunate to be marked wrong multiple times, and the same befell Will once. Kota in 9th dealt with it pretty well though and advanced first. There wasn’t much shakeup after that. Kota (and others later) clawed through the Chaos, and then we were pretty sure Kota had sealed his win since Akari ought to be a strong type for him. I don’t know what actually happened, but he either bifurcated or broke it and had to do some erasing. However he had enough advantage on everyone to finish it first, and so 9th replaced 7th in the 4-7 heat.
The 4-7 heat (or rather 4-6 and 9) was next, and I had to defend my 4th place spot here. Hashi, Slitherlink, Flowerbeds, Suguru, Yajilin. I had a very good run here. I got most of the way through the Hashi, then finished it off with a good guess and apparently gained a bit of time. The Slitherlink here seemed to be easier, so I put that one away without too much difficulty. The Flowerbeds took some time, but I never got terribly stuck on it at any point and think I got it out rather quickly. The Suguru seemed to be the hard one here. From my conversations with some audience members afterwards, it sounds like I was the only one to peg an unusual propagation of a number’s position through a bunch of X pentomino vertices. This is a step I’ve never used in one of these, and it didn’t get me anything immediately, so during the solve it felt like chasing it had been a waste of time. But that information helped crack it later. The Yajilin I tore through so fast that I couldn’t really give an answer to Tom Collyer when he asked if I enjoyed it, but I do remember using a number of nontrivial steps to do it, so I’ll have to take a look later. Evidently when I turned it in, no one had gotten much of the way through the Suguru (didn’t surprise me; I needed awhile too), so I had a clean victory in this heat.
That put me in 1-4 against Ulrich (with over 2 minutes of advantage on everyone), Ken Endo (LMI veterans may recognize him better as EKBM), and Florian Kirch, and I had over 4 minutes of ground to make up. The puzzles were Battleships, Area 51 (a loop hybrid that David Millar makes; a recent one), Kakuro, Nurikabe, Unequal Length Maze, Neighbors, and Masyu. I got off to a great start by pegging a constraint on the 3-length ships in the Battleships (1) and solving it in short order, and it sounds like I had already made up a lot of my disadvantage on this one. The Area 51 (2) was a good solve except that I missed a constraint in a corner and had to suffer a minute penalty. Then I made an arithmetic error on the Kakuro (3) and got stuck in a no-solution rut before I found the tweak to fix it. That cost at least 30 seconds. Not good when I started so far behind, so I felt pretty bad about my chances at this point. Nurikabe (4) had some easy logic in the top and bottom, then a difficult packing problem in the middle. I was told that Ulrich and I both made some close initial guesses and were essentially staring at the same almost-solved grid, but apparently I found the tweak to fix it a bit faster. (Tweaking seems to be my specialty.)
If you clicked the link above, you know that Unequal Length Maze (5) is a fiddly Erich Friedman type without a lot of logic in it. An odd choice for a playoffs. Apparently this one went really poorly for Ken Endo, leading to his eventual 4th. In contrast, I saw a way to include a path pattern I often see in the solutions for these and finished it in 15 seconds. I was rather incredulous at this, my invigilator David McNeill was moreso (that’s my proctor, for American readers less familiar with British English), and Josh Zucker from the USA B team said it was his favorite solve of the championship. I wouldn’t say that if only because I don’t really think a type like this should have been in the playoffs. To be fair, there’s a lot more logic in this one than other Erich Friedman styles, but I sure as hell didn’t use any. And as noted above, someone got mucked up by it, and that doesn’t feel all that deserved. My congratulations to Ken Endo for his excellent 2nd place preliminary finish, by the way.
I’m under the impression I was still lagging behind Ulrich at this point. Next we were on Neighbors (6). This is a type that appeared in the Dutch Around the World round last year, for which I made this practice puzzle if you want to see the type. This one was very tricky, probably the hardest of the 7, and a bit harder than the one I linked to. It had some easy steps up top, but the bottom was almost entirely white and very, very difficult to get into. Everyone got into it differently, I’m told, and we all finished the slog that was that puzzle in around the same amount of time. That left the final puzzle, the Masyu (7). The round ends as soon as three people finish, so I wasn’t necessarily expecting to ever see this one, especially given my earlier errors. Being a nikoli.com veteran, I had a very good run on this Masyu and apparently gained even more time on it. Leading to the final result of…
2nd, behind Ulrich by 22 seconds (equivalent to about 100 points of time advantage from preliminary rounds). Very close! I made up almost exactly 4 minutes in that set. Florian was a bit less than a minute behind me, I think (don’t remember the exact number). Ken Endo was apparently still cracking the Neighbors at the time (I already explained about Unequal Length above). Also, you can’t blame my Area 51 messup for the close defeat because Ulrich, I am told, made the exact same error and also lost a minute!
(My apologies for not being able to say much about Florian’s performance in all of that. I didn’t see any of it and I’ve not been told about it either. I have no doubt it had its moments.)
I am actually mostly satisfied with this result if only because I don’t really like stealing titles in the playoffs. The 2011 win didn’t feel that wonderful (I wrote a whole post about that), and winning here from a 4th seed would have left a bad taste too. Likewise, my main point of discontent is stealing the 2nd position that I’m not entirely sure I deserve (in the end the finals only swapped me and Ken Endo). In general it seems the Japanese have had a really rough time with the WPC playoffs the last few years. I think of them as a much better set of solvers than the official final rankings say each time.
The main positive is that the first thing I was told after joining back up with the US delegation was that the set was extremely exciting to watch, mostly because of that small margin of victory. The projectors and lack of easels probably helped a lot with that too. I know I’ve said this before (I don’t know if it was in person or on this blog), but if the playoffs were a fun exhibition thing that got its own separate award but didn’t affect the preliminary standings much (maybe it’d be worth some good points just like that team round today), I think I’d have no problems with them. But it doesn’t seem likely that this will happen any time soon.
That is the end of my horrendously mistake-filled WPC 2014, where apparently my best solving was during the playoffs on the last day. Thanks to the organizers for a great set of puzzles and a well run competition with no broken puzzles and no playoff snafus (none that weren’t trivial to fix at least). There were a few logistical hiccups, like the hotel apparently being overbooked and staff who knew what was going on being a bit harder to reach for help/questions than usual, but nowhere near anything that can’t be forgiven. Certainly nothing that could overshadow my self-inflicted disappointments.
I’ve learned from my history on LMI that I tend to fall back into this mistake habit when I’m coming off from a break of doing puzzles, and that’s basically what happened here. There are a lot of things I could do and an endless amount of time I could put in if I wanted to put on a better show next year, but beyond being sure to get back in shape about a month before next year’s competition, I doubt I’ll do any more. There’s no denying that my excellent finish last year (accomplishing the goal of 1st in preliminaries) has sapped some of my drive to finish well at WPCs.
Not that I plan to stop going or reaching the podium, just that whatever weak types I have now will probably stay weak, and I’ll continue to need my stronger ones to compensate. That’s probably not good enough to have favorable odds against Ulrich in the preliminaries in any given year, but it’s good enough for me when there are a lot of other things in life that one can do in their spare time. I’ll still give him the best fight I can each year I can make it to the competition, which is every year for the foreseeable future.
Thanks again to the organizers, and let’s end the post here. This will probably be the last one on this championship.