1200: Traditional football game this afternoon, hurrah! However, it's raining hard, so they're going to organise a minibus from the front of the hotel to some indoor sports centre somewhere. We had spectacular thunderstorms last night, with sheet lightning so impressive that we weren't sure it wasn't from a firework display.
1202: Straw poll! Who's going to win? We're looking at about 70% Voigt, 15% Roest, 15% "don't know". Roger Barkan is regarded as the outsider.
1204: Off we go. Six puzzles: a cross sum including 0, fill in hexagons to put one A, B and C on each row, a very faint battleships variant (you can't tell whether pieces marked in are submarines or endpieces), a very strange-looking inductive logic puzzle that has us all scratching our heads, a 9x9 Latin Square with unique diagonals and a puzzle about filling squares in a grid with stars in a slightly Latin Square fashion.
1208: Nervy start. At least Roger Barkan may have abandoned one puzzle already. Perhaps "abandoned" is wrong - perhaps more "took a look at it and decided against it".
1209: Ulrich finishes the cross sums for an early probable-lead. David Savitt on my immediate right has got his laptop online too. Good wireless access here, but a bit pricy. But you're worth it, especially when someone else is paying.
1210: We think it's 1-1-1 here, as it should be - Niels has finished the star battle and Roger the Cross Sums.
1211: Ulrich races through the ABC hexagon grid. (Perhaps it's just an easy puzzle, at this level.) Grief, the Latin Square looks hard - not many numbers placed for you already at all.
1212: Niels is on the ABC hexagon grid as well but seems to be losing time in comparison.
1213: Voigt cracks the battleships variant in about two minutes. We think the score is 3-1-1. Ulrich wasn't the 70% favourite without very, very good reason. However, I think he's left the hardest (or, at least, most time-consuming) three for last. Roger Barkan is taking a long, long time on the Latin Square.
1215: The #1 seed has moved on from the pentomino logic puzzle, but it's not clear whether he has answered it or not. (Apparently not.) It's a bit of a strange one, more a riddle than a puzzle - might get it in 20 seconds, might not get it at all. It's one where you might guess, rather than calculate, the right answer.
1216: It's not a one-horse race: Niels Roest has finished the ABC hexagon grid, but is clearly at least a couple of minutes behind Ulrich. Ulrich is on the Latin Square at the moment - if that slows him down nearly as much as it has done Roger Barkan (about whom we have not forgotten) then the pack will close up. If part of the game is recognising the easy puzzles and doing them first then that's not an unreasonable thing to challenge.
1217: We think it's 3-2-1 at half time, but perhaps something more like 3.2 - 2.3 - 1.7 in practice.
1219: Roger Barkan throws his hands out apologetically at the Latin Square - an unusually expressive gesture when people don't have the time to pose for the audience - and flips the page. I fear that he might have just spent the last ten minutes to no avail. People spending time on these and not actually reaching a solution is unusual.
1220: Niels has solved the Cross Sums and moves onto the battleships. He's not out of the race yet; he might soon be ahead by the margin of a completed Star Battle to a half-completed Latin Square.
1221: Sometimes it's times like this that you wish your Friends list was in chronological order, not reverse chronological order. Or is that just me?
1223: Ulrich is making good progress at the Latin Square, he may have broken its back. He's probably still ahead in balance, but only if he can do the Star Battle nearly as quickly as Niels Roest did. (Which is probably a fairly reasonable assumption.)
1224: Apparently Ulrich has looked at the Star Battle twice and not fancied the look of it, so it's anyone's championship...
1225: Looks like a solve on the battleships variant for Niels Roest to me, which may put him 4-3-1 ahead. (Or 4-3-2 - I think Roger has been closing the gap in the interim.) Niels has definitely done the relatively easy four, though, with just the really-too-long Latin Square and the heaven-knows inductive Pentomino logic left to go. Five minutes left!
1226: Niels abandons the Latin Square, knowing it'll take him too long to solve. Good move. He's probably going to spend all the remaining time on the inductive logic pentomino problem. If Ulrich can complete the Latin Square then he'll bring the score up to 4-4 (barring mistakes - hah, mistakes!) and will win based on the performance over the first two days, but the Latin Square is looking very, very tough.
1228: Roger is back on the Latin Square and this time he's overwriting it. I think he knows he won't win, he just wants to beat that motherpuzzler for the sake of pride.
1229: Ulrich's got the diagonals of the Latin Square, but it's going to be very tight indeed... one minute to go!
1230: Roger throws his hands up again, this time at the Star Battle, and Ulrich races to finish the Latin Square! He does so and gets a premature round of applause. Roger squiggles all over the battleship in frustration, which gets a big laugh, but distracts the others slightly. Has Ulrich completing the Latin Square with about 30 seconds left created a win on a 4-4 draw?
1231: We think it's Niels! He got the pentomino! If he hasn't made any mistakes then he's got five, which must be enough...
1232: Niels's marker has just shaken his hand, one of the organisers has put his hand up to show five fingers, and Ulrich's completely random guess at the pentomino question isn't right. We think it's Niels by 5 to 4 - or, rather, by a pentomino and a Star Battle to a Latin Square.
1233: IT IS NIELS! UPSET! STANDING O! Ulrich second, Roger third. (Roest 5, Voigt 4, Barkan 3.) Chapeau to Gary Sherman, the lone person to go against the grain of Voigt predictions and select Niels Roest.
1234: Lots of applause, lots of photos, a fair amount of video and discussion of moving to lunch. The discussion topics will be the suitability of an extremely long Latin Square and an inductive-rather-than-deductive-logic problem in the final, when we haven't had an inductive logic problem like that anywhere in the championships to date. The policy of taking 13 for the play-offs rather than 8 also will come into question: remember, Niels was only #9 seed coming in.
1237: Everyone wants to shake the new World Champion's hand and fanboy Dickson is no exception.
1239: The conclusion is that Niels had brilliant strategy on all three rounds, which is what won it for him. Ulrich can probably justifiably feel slightly aggrieved that he's only - only! - on the second step of the podium this time, but it's not as big an upset as the way he won (or, rather, Wei-Hwa lost) his first championship in 2000.
1242: Aaaaand... relax. RTL (are they Belgium? France? The Netherlands?) are interviewing Niels. Niels (wearing extremely natty blue-denim-at-the-front, cream-cords-at-the-back trousers) is doing the interview in English, arms firmly folded, and doesn't look completely happy about it. Industrial hazard, mate - when you're the world's champion, you become the world's property.
1245: Or am I getting that confused with "to be World Champion, you've got to take on the world"?
1246: Probably a sign that this has gone on long enough. People are milling around and chatting. The British team, gamely enough, are solving the puzzles from the play-offs. As for me, I skipped breakfast and am minded to go and grab lunch, not least because I can smell chips. Mmm, chips.
1247: Chris Dickson signing off from the World Puzzle Championship in Opatija, Croatia. So long, everybody, and here's to Hungary in 2005!
1st Niels Roest NED 5/6 2nd Ulrich Voigt GER 4/6 3rd Roger Barkan USA 3/6