Day Three (August 29th, 1997 ... sorry, scratch that ... April 16th, 2008) - Judgment Day

I got maybe five hours of sleep which was tremendous considering most nights 0. But at 6 am I heard birds chirping and was not getting back to bed. After thinking up some meta puzzles and general puzzles that might work for the MIT mystery hunt, I decided I needed to put on my pants and see some scores. Round 1-3 were up and I saw Jakub's super Round 2 but I had had three good rounds and was in the lead. I knew I was safe over rounds 4 and feeling ok over round 5 and by an hour later, I'd see I was still up 25 over Jakub O. after 4. Then, by 9 am or so, as I went to breakfast, I got the unfortunate news I'd actually made that transposal in round 5 and my 475 became a 450 and Jakub and I were exactly tied after 5 rounds, with over 250-300 points down to 3rd place. It was quite the story for others at breakfast as someone else was in the Snyder orbit now that was not an expected name like H. Jo or M. Ley or Y. Kusui or anyone else I'm missing that you bet money on. I joked about it with the others there, particularly Riccardo of the Italians, and suggested - since everyone wanted to see the two of us face off - that we should ditch the 4 competing for 1 title protocol, and just let the two of us go mano a mano, particularly if what was all 4 competing on 1 puzzle on stage which is a bad bad format.

We were still in striking distance of a team title though and I wanted more than anything to deliver that. I figured the playoffs were a danger zone (we didn't even know what format/# of puzzles/types of puzzles would appear) so the one sure thing was to bring 1st back for Team USA. I put on my competition red, for the key rounds of the championship.


Round 8 - Weakest Link
Individual Part (400 of 400 - David McNeill and I essentially tied for time and both sat at our team desk at the same moment)
Team Part (600 of 600 + 4 minutes (120 points) of time bonus)
Placement (2 - 90 points behind Czech Republic)
Puzzles Broken while solving: 0
Outstanding Puzzles: Extra Regions part of the team part - I actually saw every puzzle here so I can fairly say this was the best although the whole round was constructed ~great for what it was.


So, you start at the desk and get a samurai. Its got a lot of different stuff going on, but you solve it and then have a tedious and dangerous transcription step into a classic on another page that is really scattershot in letter going to very different position from A to Z. The classic is pretty standard. You get to the team desk. This part went well for me. I said finished and felt I was first, but on the other side of the room David McNeill of the UK had also just finished. I'll take the tie.




Now, in most weakest link rounds, you can't do much until guy #2 or #3 shows up. This was different. While the feeder puzzle at the table - the classic sudoku where you'd get more givens - was not much handle-able, there was still some info on what its letters would mean in other puzzles, and also the other puzzles had a lot of data. I went to work on all of them. That's right, all of them. I had all 5 visible on the table. I'd spend a minute on the irregular, place 4-5 digits, see if I could use the letters yet, then do a minute with the diagonal, then do a minute with the odd, .... I found the odd completely solvable, although I used uniqueness at least 3 times. Ok, not completely solvable, I had 2 of the letters solved and the 3rd was an either/or giving a deadly pattern that would require knowing what the letter in the classic was. Otherwise, finished. Still no one else at my table and I saw other teams doing a whole lot better number-wise. I gave my first look at the extra regions - absolutely beautiful arrangement to the extra regions! - two groupings of 4 and one lone 1, spread all over the grid in different rows/boxes/columns so that the communication between regions was neglible yet phenomenal. Made progress there. Went back to the irregular. And the diagonal. I basically had 3 puzzles stored mostly in my head when Wei-Hwa finally arrived. I eventually heard his weakest link disaster story (much later, after the round). He had seen the odd squares in the odd blue color as being in a diagonal. This is a good observation, as they were all along a diagonal. However, it did not mean this was a diagonal constraint, which was his mistake. I did not know where Jason was or what his problem was, but now having a second set of givens and tons of info started on the letters, I got 2 pinned letters for the irregular and handed that to Wei-Hwa to work on and finish. My notes have a synergistic state to Wei-Hwa's and he'd comment later that he could pick up my paper and use it no problems. Jason finally arrived, the classic fell 3 minutes later, and then we were all on puzzles. I did the diagonal completely from where I had it, worked on the extra region with Jason (who couldn't stand my numbers, and I can't stand or even read his), finished that with Jason, as Wei-Hwa was doing the irregular. We checked everything, and turned in second. Given we were 6th getting the team to that table, we caught up real quickly.




I wonder if the whole last 4 team puzzles could be solved without the team at the desk (at least by a crazy man like me) or maybe just the whole round with only 2 at the desk (except maybe not the classic). Its odd to not want a teammate when you think you are doing everything well yourself, but I spent a lot of mental energy to lay the groundwork for success in all 5 team puzzles and delivered a great performance here. We were now certainly in 2nd as a team.

Team Round 9 - Tiles
Team points (0 of 1200)
Place - (last, tied with Australia)

Jesus this round kicked my butt. As a team we knew we'd have to be first to finish the round, and finish the round significantly in front (or have a Czech stumble), to have a shot at the title. Our strategy was therefore the one we felt would achieve this. We identified the most difficult puzzle (correctly) as the diagonal and gave that to me with Jason taking his "I love cross sums (read: kakuro)" specialty of the Sum Sudoku, and Wei-Hwa taking the irregular. As two sets of eyes looking through a placement of tiles in any of these puzzles could duplicate efforts and potentially waste time, we solved this as 3 individual puzzles. It was not a wise choice.

I started on the diagonal not in the smart "there must be something easy way" but in the "I am a super computer, I can do the Spider Webs [WPC 2005] or Surely you're Hexing [MIT Hunt - Spies 2006] on paper/excel without computer assistance" approach. I created a matrix of tiles into two groupings of three sets, labeling pairs that must be in the same/different rows and any pairs that must be in the same/different columns. There was a lot of there there, but not a great starting point to be honest. It was easily 8 minutes wasted. I then made the obvious deduction of the center square and my super computer information became instantly important but not relevant as the center squares of the tiles told you most of those facts quickly too. Finally, I got the placement I wanted and solved the puzzle. With 4 squares to go, I had a 62 pair in box 4/6 that gave a non-unique solution. This bothered me but they said the diagonal had more than one solution (although I'd expect point-reflection to be the cause of this more than anything else).

Then I checked my diagonals! Fudge. Only I didn't say fudge. I said THE word. The big one. The queen mother of dirty words. The F-dash-dash-dash word.

I think I said it maybe 4 times. I don't say the word often. This time earned it. I went to stick soap in my mouth after the round.

I saw one way to fix the repeated 6 on the diagonal that needed a 69 flop in box 7 to also fix the deadly pattern to give one answer. Only that made a tile that gave 186 now give 189 and you can't change givens. Chocolate Fudge! ....

I went from asking my teammates who needed my eyes most next - hopefully to get us to 2 and maybe 3 answers in the time that remained - to having bubkis.

Wei-Hwa had his own stupid story in the round. He wouldn't recognize it until later, but on his grid, in that 3x3 piece in the irregular, a 6 was written on the paper as an 8 when he copied from tile to paper. He was solving a completely different puzzle. One that wouldn't be giving us any points.

Jason doesn't reveal his cards much - not that he's good at poker - but he couldn't do the sum and lets leave it at that.

0.

Zero.

ZERO!!!

We tried the round as individuals, and individually we each failed. It left a horrible taste in my mouth. One you might clean with a bar of soap, except it would make you blind. I was completely demoralized entering the playoffs but fortunately had lunch imbetween.

Did I mention we only now learned what the playoffs were?

The semifinals would have an even (okay, but basically a classic), a touchy sudoku ("What's a touchy sudoku?"), and a missing digit sudoku ("What's a missing digit sudoku?") while the finals would have an equation sudoku ("what's an equation sudoku?") and a quad sudoku ("what's a quad sudoku?") and a toroidal sudoku (HAHAHAHAHAHAHAHA .... Confidence restored) with the last classics final having 3 classic sudoku, listed separately on each line 1. Classic Sudoku, 2. Classic sudoku, 3. Classic sudoku, to make a point.

TOROIDAL SUDOKU!

We still didn't know format, and I still didn't know what the heck these puzzles were, but Jonathan pointed out that examples were posted near the playoff results. I read them quickly, demanded a paper copy, took that and my stuff and went to my room to sulk. Fudging team round.

TOROIDAL SUDOKU!

Yeah, my mind was wrestling the team round/toroidal sudoku yin-yang of emotions for a long while.

I did show up to lunch, had the comfort meal I discovered at the championship - hot buttered nan plus cold ice cream equals yummy - even if the meal is a bit sacrilicious to Indian cuisine. The flavor at this meal was butterscotch ice cream so it was buttered butterscotch nan. I had blazed through the examples of the playoff puzzles back in my room (although I expected they would be much easier than the ones in the playoffs) but was already game-planning what strategies would invariably show up on those puzzles in the playoffs. I certainly wasn't sharing that hand of cards with anyone else who might be on stage.


Lunch over, we entered the competition to start the Semifinals:

Semifinals - 3 puzz...




Wait - we need to tell the competitors the rules. Well, it turns out that now all the rounds have 4 puzzles. Oh, and the semifinals will be run like a horserace (ala WPC Eger) with 4 puzzles at 4 desks, and you do them in order until your invigilator checks them. The invigilator will take 1 minute - invariant - before he lets you pass, although if you are wrong, you are given the paper back with the equivalent of that 1 minute penalty. It is an excellent playoff system, and the 1 minute invigilation fixes the flaw in the Eger set-up of the 5 minute penalty/puzzle skip. However, I would have gladly known this before 2 minutes before we start. Any last surprises? Oh, you get a time bonus of three minutes based on score between 1st and 8th so I actually have a 1:15 to 2:45 head-start on everyone but Jakub. This meant day one mattered, not that I was specifically saving energy during it, but a surprise none the less. Again, nice, but should I learn this fact 30 seconds before I wait 15 seconds before I start? Probably not.



Aside: The Indian judges were invigilators. This word was used a whole lot. It made me giggle. I wondered if Dementors were like invigilators, since it seemed an invigilator might be the kind of judge I'd have when getting accused of a serious crime, not of writing a 8 instead of a 9 in a sudoku grid. I'll accept judges or proctors any day. Invigilators gives me a chill, until I remember my patronus.






Semifinals -
Placement (2? of 8 - only 3 finished the round, Yuhei Kusui in 1st)
Outstanding Puzzles - the regularly scheduled ones (Touchy and Missing Digit).
Broken puzzles: 3 (Touchy - 3 to 4 minute with erase; Missing Digit - 80 digits, then full erase; Diagonal - simple correctable error, but still 1 minute penalty + 1 minute to find it)


Touchy sudoku was the first thing at the desk. It is a new variant. You can't read of it online in their rules, but it is kickass. Basically, any number in the grid must touch some number that is consecutive to it. Its like a consecutive sudoku on a mind-altering drug, only the walls aren't marked. The 1's and 9's, which must touch 2's and 8's respectively, are good places to start, but this was not a normal touchy sudoku, it was a very hard one for a type none of us had seen before. I got started, made a too quick leap to placing a 9, but fortunately, it broke only a minute or so later, so I restarted a little more patiently and powered through it. Yuhei and I were both being invigilated at about the same time, then sprinted to the second desk tied for the lead. My invigilator, by the way, same as at the WPC, was Jean-Cristophe Novelli of France. I joked with him a lot; it kept me loose. I still fudged up a whole bunch here though. I don't do well when he's invigilating me. I wonder if invigilateur means anything en francais.

Missing Digit was next, and it was a type I'd learned the lesson during the Japanese Number Place Championship of 2007. 0-9 go in the grid, except on the outside some of the rows and some of the columns have a number given which does not appear in that specific row/column. I learned on the JNPC that you write the missing digit in the grid. Make it a 0/9 sudoku immediately. Even knowing this, even telling my invigilator about my "crazy" notation, I stupidly do not resolve the 0/9 part immediately when box 4/6 looks a bit wonky in my assignment and I've stuck the wrong missing digit in the wrong place. 80 digits and 5 minutes in, I'm staring at a cell that needs a 4 in a box/column that both already have a 4 and realize I'm in trouble. Well, only one other guy is at the second table (Jakub H) so I'm not that bad off, but Yuhei turns in at the same time I should be done and I realized I'm just playing to get to the stage now, not to win the semifinal. A full erasure (with no real relevant information from the horribly wrong path) and a complete resolve later, I now move ahead close in time to Jakub H to puzzle 3 as Jakub O, David M., and maybe Michael L. are at the second desk.




Third is a classic. I solve it like a classic. It works ok. I gain no time on Yuhei who is at the diagonal. What is a classic doing in this semifinal which is not for the classic trophy?

Last is a diagonal. I solve it like a classic but with a diagonal constraint. It works ok. I turn it in, relieved. Then, one minute later, I'm invigilated. And confused.

Ok, I check the rows, check the columns, check the diagonals. All look ok. Is the answer not unique? I look for problems like that, then finally see my 8 with a particular closed bottom loop masking as a 9 in my mind during the quick search, fix it by turning 8 to 9, and finish. 3 competitors are done and 2 are at the final desk (Jakub O. and David M.) as time counts down. Oh, the round runs for 40 minutes. We actually knew this before it started - although with time bonuses and invigilation costs, it actually was 35-37 minutes. Anyway, with 4 spots on stage, and 2 people a minute or two away from finishing the diagonal, I personally would have extended time shortly to let things finish naturally. The 40 minute limit was too punitive.

Well, at least they could grade whoever was closest to being done. Unlike in Eger, its easy to tell on a sudoku who is farthest. Sum up the correct digits. Maybe subtract incorrects, although allow them to be notes so really just use big certain looking digits. I'd actually fought for this scoring system before the horserace started, but the organizers decided to use "seed of player" at respective desk as the tiebreaker. I don't know who was farther, but it should have mattered. At least, "let them play" and not give David the bad news to bear that he was a minute from making the finals but the round timing screwed him. If we all started at the same time (just as likely when we set down at the desks for the semifinals - David would surely have advanced as he sat for over 2 minutes before he got a go at the touchy).

Overall, I liked the semifinal format but some of the organizational choices were a mess, learning 66% of the salient rules immediately beforehand, in oral English (when many competitors don't speak English as anything close to a primary language), and being given extra puzzles we weren't expected after lunch were all unacceptable. Having a classic was certainly unacceptable - you self-select all of those out if running two tournaments. While the regularly scheduled puzzles were good, getting the two boring other ones didn't add much to my life. Oh well.


On to the variant finals - my home territory - a toroidal at the end ...

Um, wait, we're doing the classics finals first. Huh? The sheet says semifinals/finals/classics finals.

Oh, that was just listing the puzzle types.

Well, its not fair to shift competitions and give some people rest between rounds while others compete for all titles.

Have a seat.

It doesn't have to be this way.

Have a seat.

At least tell me when it will start.

2:30.

At 3:00, Jakub O. (1), Thomas S. (2), Michael L. (4), and David M. (7) took the stage for the classics finals. I'd predicted beforehand that the classics finalists would all be in the semifinal playoffs, with maybe 3 out of 4 being in the top 3 out of 4 anyway. That prediction came out real good. Why run two events? Come back here later and I'll explore that more.

Anyway, we were again given a new set of rules. No one would have head-starts here. We'd get 4 puzzles, one at a time to solve. Each puzzle would be run until all 4 people finished (or 15 minutes had passed), then the next, then the next. Cumulative time would determine the champion, with any error costing you everything unless others had a similar number of errors. While it didn't matter, the organizers eventually decided, after I asked a pretty expected question and then waited 2 minutes, that correct cells only, not correct cells minus incorrect cells, would be used to determine the best score on an unfinished puzzle.

Now, the classics had been all over the board, and Jakub had done exceedingly well in those two rounds, but I expected he was using an even more reckless strategy than I am willing to given how well I'm spotting triples and naked singles these days (watch my solution to this puzzle again on Youtube if you aren't frightened of me by now). Well, I quickly learned that he places a pad of pre-sudoku grid-defined graph paper near his puzzle and quickly writes down his bifurcations onto the pad for better or worse. He's learned to use it to not waste time losing notes when you need to erase, but has basically practiced seeing info on one grid in the other (or, sometimes, copies over). I don't practice with a bifurcation pad, as I always hope that the puzzles are fairly tested and ready for competition. I don't want to train to be the fastest guesser in the world. Some puzzles need some guessing. Some puzzle types (spokes, snakes) almost require using it to converge to an answer. Sudoku is grounded in logic - not math, logic - and feels like it should be won fairly. The dirty secret of classic, "vanilla" sudoku is that steps below maybe that second or third line of tough strategies on scanraid.com (and rarely should a competition puzzle stray very far below that line) are never used by the fastest solvers in competition. Swordfish? Maybe when I'm having a laugh at home for a half-hour with a gem from Tetsuya. In a competition? Not likely.

Bifurcation gives you back-doors into puzzles. So, you test people above that line with hard examples above the line if you know what you are doing. I'll rant on this, maybe as long as in this set of posts over again, but it cost me one individual title now and one individual qualification title, and its not just that I'm an overly proud, immature man who can't admit when he has lost but that I want to lose because I'm bested intellectually, not dealt a crushing runner-runner miracle by the dealer to my opponent or some such.

Anyway, Jakub was clearly into perfecting the "nuclear option" as I'll call it of sudoku, willfully jumping to a bifurcation pad some significant number of times when he gets stuck. This is a high variance approach to the world of sudoku. I do my "one-minute rule" which works ok, but I lose a minute of mean time (at a great reduction in variance) by solving that way each time is too hard. So, off we go. Let's see who is best.



Classics Finals:
Puzzle 1 - GOA = Outstanding Puzzle
Placements:
David M. (2:32) (lead)
Thomas S. (2:36) (+:04)
Michael L. (3:08) (+:36)
Jakub O. (3:33) (+:57)



Note: my times are from the spreadsheet I kept onstage and my cumulative times have whatever math errors I made on stage during the event. Cumulatives are the right column.

So the first puzzle wasn't bad at all. It flowed smoothly, and earned its placement in the Times of London with its givens in the shape of G O A and its winner in the form of the UK's David McNeill. Only three puzzles to go and if results stood, the whole UK would be holding at the very least two parades, and we'd possibly have our first knighted puzzle solver. Interestingly, Jakub did the worst. Maybe it was live by the sword, die by the sword - this puzzle didn't need a guess and he could have gone down a wrong path - but I was glad to be near the front. I also could have sworn I was 1 second back. I finished and started to check because any error was death. When I heard David say finished, I raised my hand instantly. My reactions aren't that bad. My guess is he raised his hand, then when he didn't see organizer movement, said finished, but anyway the times are what they are.



Puzzle 2 -
Placements:
Thomas S. (3:01) (lead)
Jakub O. (3:15) (+1:11)
David M. (4:00) (+:55)
Michael L. (6:16) (+3:51)

This was harder than 1, and I think I found two naked singles that were important rather early. In the photos, David seems tied with me at the middle stages, maybe even an extra insignificant digit or three in the bottom 6 rows, except for one of those naked singles in Box 1 that I have and he hasn't. It would end up being about a minute difference. Not all the cameras are uniform or pencil writings dark enough so I can't tell where Michael or Jakub were at the equivalent time. Still, I had a large lead on one competitor now, and about a minute on the other two. Doing great. No errors (that I knew of). Clean. A feeling I hadn't had in a round in a while this championship.





Puzzle 3 - neat ring around the collar and/or inside puzzle
Michael L. (1:55) (+3:13)
Jakub O. (2:04) (+:44)
David M. (2:06) (+:28)
Thomas S. (2:33) (lead)


This puzzle had a tough middle. It almost certainly depends only on the givens. It also had an easy outside. Boxes 1/3/7/9 were therefore by design isolated from the others puzzlewise. A cute design. Everyone likely did the outside first, then had to deal with the completely bare middle in box 5 and all the bare triples in boxes 2/4/6/8 that would only fall when the middle did. By the end of the puzzle, I had the pressure of knowing everyone else was done with it, but took an extra 5 seconds to double-check my last placements. No reason to finish first overall, but with a single digit transposal costing the title.


Ok, I sacrificed time there, but still golden if I have another at that difficulty.



Puzzle 4 - This one went to 11 (or more accurately, this one went to an XY-chain, after an X-wing, simple colouring, some more x-wings, a y-wing, an x-cycle, and a uniqueness rectangle. Thank goodness it didn't need the kitchen sink elimination. Note: These are the scanraid steps. Some might be useless, but after any hard step it cycles back to the start and so I'm just cataloging that stuff in the order I see it pass below not just the tough line, but the diabolical line).

Outstanding puzzle? - definitely not

Michael L. (7:56) (+3:01)
Thomas S. (8:08) (lead)
David M. (8:50) (+1:10)
Jakub O. (13:31) (+6:07)



Ok, so when I first saw the puzzle, I subconsciously remembered many things. First, this was a pattern I wrote with the title "Four Square" for the kids in the Silicon Valley Puzzle Day Championship. It is a pattern where the cells pinned in the rows/columns with only 3 unknowns are critical to make progress. You'd love one to be a naked single. Alternatively, you can attack the 4 corners, but only if both adjoining boxes in rows and columns contain the same digit (a total of 4 occurrences, easy to search for). Now, I rued having given this to the kids as it turns out they never look by geometry, only by number, and anything "naked" is unknown to them. It may be a puberty-plus rule? I've also solved this pattern before in Nikoli books. I know already how and why this pattern with three bare rows/columns apiece is hard. I knew very early it would be painful. I did not need to use a "one-minute rule". I got the 5-6 easy placements, looked briefly at the bivalued cells for something screaming "X-wing", saw none of that, and then went to guessing town. Now, my first guess at the top (a 2 in R1C5) was a poor choice. It looked tantalizing - it was clearly a spot that would give an answer in about a minute and a half. If I was going to be really lucky, this would have given me an answer in 3:30. Unfortunately, it has something a good guess spot never has - no real value to the or of the either/or. So, when I proved it wrong, and erased a mostly filled grid to the laughter of the audience, it was almost back to square 1 except with a 2 in R1C6.



Ok, C6 is now the magic column, and a 1 in R5C5 not only forces a lot of digits, it also forces something nice below by uniqueness, or at least seemingly so. Again, went down this path for awhile. Learned (as you sometimes do in a guess) that R2C9 and R8C9 is 48 early on. This is real, not guess information I stumbled on, so I jumped out of guess mode to use that. It proved useless as the 4 in R5C78 doesn't resolve R5C4 or R5C6 (where its most necessary/potentially useful) and the 8 you wish it forced in R5C78 that would be critical isn't actually there as there is an 8 in box 6 already. Oh well. I found the right breadcrumb for that 11 impossible step solve by Scanraid.

Back to the guess - power through it for another couple minutes - fudge, its wrong too. Put in the right 1's, then because my notes are totally messed up, I convince myself box 5 is impossible. I erase everything, put in just the 2 in R1C6 and 1 in R5C6 that my guesses have now made definites, and resolve the middle and now it works and I trickle it out to solve the whole puzzle. Not my best effort, and definitely not my luckiest, but it was second by only 12 seconds.


This puzzle was not outstanding. Outstandingly difficult (beyond even a human with logic in a half hour) is not outstanding even if "outstandingly difficult" contains the word "outstanding". I'll let sudoku.com decide at some point but this was silly.

After the puzzle, where I'd expected like Barry Greenstein to be signing a copy of my book to the person who defeated me, I instead got to keep it myself. I figure since I have something to celebrate that I'll put it on camera shamelessly. It gives me a laugh and keeps me loose. And the Chinese get really interested in publishing it.




Ok, so I think I've won the first "classics" title. I ask everyone if they saw an error in my grids. It seems to some of us that the Indian organizers are learning they need to invigilate the final puzzles too, not just assume people who say finished are finished and correct. It takes 15-30 minutes to get the next final set-up, but they thankfully announce the winner of the classics before the regular finals take place. Still, Yuhei and Jakub H. have had all that time to relax while Jakub O. and myself suffered through the classics (and winning did wonders for my psyche, after the semifinal disaster of 3 broken puzzles out of 4, but could not have left Jakub O. unaffected). This rest by Yuhei, who most recently beat me by 5 minutes in the semifinals, should give him confidence. I feared Yuhei the most - already finished 2nd in a WSC playoff, and he was a much stronger finisher overall than last year, this year at 3rd and not 9th before the playoffs.



Finals:
Puzzle 1 - Even Sudoku - Outstanding
Jakub O. (2:02) (lead)
Thomas S. (2:56) (+:44)
Yuhui K. (3:42) (+1:40)
Jakub H. (4:26) (+2:24)

Now, I know evens are basically classics so I'm not worried about Jakub O.'s time here, mainly getting a little time on Yuhei. The puzzle (originally intended for the semi-finals) is actually of a pattern I've written 3 times. Everyone does a smiley face sudoku. Here, the 2 even orange cells were the eyes to a smiley face. Battleship sudoku has its smiley face on 141 (although I use the down ship as a correctly shaped nose and 2 subs as round eyes - so its much better. Its also the only battleship sudoku with no radar clues for ships, although I could have written many more with this logical variant). I wrote a smiley face for the kids in Morgan Hill at SVPD too. Back to back puzzles that I wrote for the kids? Is Tic-Tac-Toe next? Competitors in this year's US qualifying for the WSC would have seen a greater than/less than variant I wrote (that will be in my next sudoku book and no one else will see til then) that worked out to look like a smiley face too. Anyway, in a 9x9 pixel image, a smile is easy to do. This puzzle was easy to do. I lost time to Jakub O. but got a minute on Yuhei. This was warmup anyway. A toroidal awaited (81-1-1-1). I took the extra time here to watch the others solve a bit and see where they were relatively. Then we were efficiently onto the next puzzle (the Indians took less than a minute to reset, hand out puzzles, say start - that was real nice).

Puzzle 2 - Equation Sudoku
Thomas S. (5:38) (lead)
Yuhei K. (7:34) (+2:44)
Jakub H. (8:51) (+4:43)
Jakub O. (14:28)* (+8:06)
*(after the championship, it was revealed Jakub O. only fully finished 2 puzzles so I'm pretty sure this was his first DNF)

Ok, this puzzle was outstanding for me, but not outstanding to me. It involved a series of letters (nine total, multiple occurrences of each) being in the grid in different places. As Jonathan would comment later, you could immediately solve the sudoku part of it by writing letters elsewhere. However, there was also this very involved linear algebra part. You had a number that gave you a sum of the letters given in that row/column. R--GOA--L = 25 for example meant that the values (unique) to R + G + O + A + L = 25. I'm making up that row, but some row had GOA and other letters on the edge, some had WSC and other letters on the edge, some row had something else. I don't have the actual puzzle with me right now. Anyway, not all the rows had 5 things. Some had only 3, and one had a set of 2 that was exactly the members of a set of 3, giving an instant assignment. I figured going in that a puzzle with tons of linear equations was the kind of crazy math I could beat anyone in the world (outside of real math people - this was a sudoku event after all) at. I wanted to hit a complete homerun and break the field apart. I like leading from in front (Tiger Woods and no major comebacks anyone?)

I tear into it and have all 9 numbers -> letters within probably a minute after I get a key set of 3 equations reduced to 2 variables that form a cycle GO = ?, SO = ?, GS = ?. Something like that. Needless to say, no one else was close to breaking it that soon. The problem was I had to write these numbers over the letters on the grid and this led to a lot of incomprehensible numbers when I stared at the puzzle and certainly slowed my sudoku-part solving time. Letter-based overwriting substitution puzzles are the only puzzles that need that bifurcation pad. Solve somewhere else and return later. Ok, that and "circle sudoku" which is also not a variant. Anyway, on that puzzle I really wish I had a bifurcation pad just for the grid on it. Still, I finished REAL early, checked carefully given the mess of my grid, and then started to wait for others. Then I realized, I'd be waiting awhile.





I was done advertising my book, so I pulled out a very different recreation that was my counter-sudoku puzzle programming for the weekend, a xeroxed page from Super Tough Word Search Puzzles by Dave Tuller. The actual puzzle was p. 14, Getting Lost, which had some good words I circled instead of lined out including "Competition", "Contest", "Tournament", and "Title" - all other words, as any WPC solver knows, were lined out for efficiency. However, this was a "super-tough" word search and each entry was missing a letter in the grid. So, while the audience started to laugh that I was doing, of all things, a word search to stay loose, they may also have been wondering how a potential World Puzzle Champion was taking forever to complete it. Again, each entry was missing a letter in the grid. I'm glad I'd already finished p. 13 which involved anagramming all the entries to new words which themselves were in the grid. That would have been really embarassingly slow after I got that Britney Spears one taken care of. I think I peeved the organizers (at least the long audience reaction to my word search was lots of noise), who eventually turned off the cameras for the round. Still, my equation sudoku solve had busted the field. Unlike all previous puzzles where someone was within a half minute to a minute of the first finisher, this had a huge spread. I was in a comfort zone (which was key in Prague), and planned to stay in it. Up next was a quad sudoku.

Puzzle 3 - Quad Sudoku = Outstanding
Thomas S. (5:40) (lead)
Yuhei K. (7:26) (+4:30)
Jakub O. (8:28) (+11:54* - I did not know but #2 was probably a DNF)
Jakub H. (DNF) (*, DNF)

Ok, so the organizers may have hoped not giving the playoff puzzles to people beforehand would keep them from devising fiendish strategies. Now, on a toroidal that won't matter as I've been there in competition and done that. Well, the quad puzzle and example seemed innocuous enough. There would be 9 2x2 shaded regions in the grid. Within them, unique to each 2x2 region, would be one of the quads (wrapping from 9 to 1), namely [1234], [2345], [3456], ..., [9123]. Its a fun new concept. The example puzzle gave three givens in each of the 9 darkened regions and some outside so I solved it in about 80 seconds. I doubted this one would be like that. I bet it would have no digits in the 2x2 regions. I also bet the 2x2 regions would be staggered, but this isn't the case, which makes things easier. This meant that the outside givens in a particular box (say a 2 and a 5) might pin 2 digits (34) which cannot be in the quad, leaving something like [6789] or [7891] in the quad itself. I had subconciously figured this out before the puzzle started. My conscious mind caught up 15 seconds into the puzzle. Then I was again, off to the races.

First, I needed book-keeping on the side. I guess someone on their scrap paper could already have 123456789 written out - I didn't - so I wasted time here and on the equation sudoku writing down my bookkeeping. I personally think pre-written notes are a form of cheating. So, I write 123456789 and I go to write 45... below it, only I start 4678912 and then realize I made my first error of the playoffs. 5 is a number. It is between 4 and 6. After that was done, it was just using my now consciously deduced quad-pinning rule, and bookkeeping which boxes the quads appeared in, and then it was mostly a classic solve. I again was a minute and a half in front of Yuhei so I was back to the WODSEARCH or WORDEARCH or whatever the puzzle in the Tuller book should be called according to the drop a letter rule. Jakub O was third, but Jakub H never really managed to grasp how to do the puzzle. The title, for all intensive purposes, was down to 2 players, with only a big screw-up by me likely to cost it.


4 - Toroidal Sudoku - Super Outstanding
Yuhei K. (9:13) (+1:21)
Thomas S. (12:22) (lead)
Jakubs (DNF) (2xDNF)

So, my first thought was "I hope, like the example, its a regular toroidal, not an irregular toroidal". Its not that I don't like irregular toroidals - heck, that first irregular toroidal I solved inspired my first legendary puzzle, the United States Jigsaw Sudoku and its cool to see that in my original US jigsaw entry I was already talking about regular versus irregular toroidals as if the name should matter. Its that my magic left/right or up/down rule that I developed (while on stage in Lucca, and then discussed over at sudoku.com) does not work for them. Instead, you basically use irregular rules and just have the headache of edge wrapping to deal with. Yuhei is testably now twice better than me at irregulars in playoffs, and I bet he'd crush an irregular trophy although I'm confident I'd be second. I had 4.5 minutes though, and that is a lifetime in a single puzzle. Just need to be careful - and unlike every other puzzle - reconfirm each individual entry at least once as, if you make an error anywhere in the toroidal, you won't catch your error until 15+ minutes are up and you are toast.

The pattern was really nice, a square in the center, 4 biased shapes on the sides, none of those 5 shapes toroidally constrained. The 4 symmetric corner pieces though each contacted 3 of the four corners of the grid (so one piece was in the LL, UL, UR; another was in the UL, UR, LR; ...). I got to work. A nice trickle of one particular digit happened. Then a break-in with another particular kind of digit. It was a well-designed puzzle in a championship with so many memorable ones. I had a bit of a struggle then, tried working the 6s, and even almost wrote my first 6 twice in a column before jumping to the next 4 placements that would give. Caught that one soon enough though. In a tough irregular (or any tough sudoku really), I'll try painting early on to force something that feels "off" to be wrong giving me a placement. Painting is not guessing. Well, sort of. Its unwritten guessing. "Beautiful Mind" you see it on the page guessing. My fingers danced across the page on this puzzle. It often gives me a LoL type constraint I wasn't seeing as LoL, which is something that was true in the irregular toroidal sudoku I linked to above until the forum cleaned up easy versus hard LoL regions for the world. Anyway, my hands and eyes were racing all around the grid. It was dizzying. I hear Yuhei declare finished, but I know where I stand time-wise and I know that I'm still on a good path. I work and work and work, and finish, and still know where I stand time-wise, so I do a full check of rows and columns (but not the toroidal regions, that's too hard and long to do) and say finished. Have I won? By time, certainly.


I'm pulled off stage, congratulated. The Aussies (who deserve and will get their own entry) are there to congratulate me in front of the documentary crew (ok, I'll subsume them in the Aussie entry). Michael Harvey already wants an interview when I just want a drink and a shower and maybe a nap. I hadn't even been told I'd won so this was a conditional interview I suppose, and while he did not butcher DNA sub straights [sic] like last year, I definitely said many things more compelling than just: “I take advantage of my talents.” Not even my book-plug worked with him - no mention in the article. Drat!

So, with time before the awards, I snuck back to a teammates' room and sadly used a sugar packet, my one option, as sustenance. The awards then started - Hendrik seemed really tired but got through them - and the pictures were as always a bit too many but still manageable. No Charles Bridge filming to speak of (although I would spend an hour with the documentary crew the next morning).

Something that doesn't happen when there isn't a formal press conference is that the players cannot officially thank the organizers for their efforts. I would not take my 1st place overall trophy until I'd specifically thanked Hendrik and his whole team for putting together a fundamentally brilliant championship with so many stellar puzzles.




Dinner eventually happened - again returning outside after many lunches and dinners inside - with Indian dancers and music. There was karaoke outside the night before, but the competitors seemed to have left before too many caught onto it (Wei-Hwa and Jason sang some, but really just to the rest of the US team and some non-sudokuist guests). Unfortunately, no karaoke that night. ALL W(P/S)Cs should have something musical like karaoke. Dinner had a lot of discussions on what worked in the championship, and what really didn't. I continued to use the term I created on stage - "nuclear option" - to describe the effect bifurcation is starting to have on the championship. I don't mean to pick on Jakub O. - he has incredible solving speed and his Youtube video is impressive - but the approach he takes to crack impossible puzzles by bifurcation is not always how sudoku is meant to be solved (IMHO) but may be how you solve one if all you care about is speed. It sets a very dangerous standard for the future, and I hope he advances his general solving skills to rely on it less.

The music chased Wei-Hwa to the room early. I stayed and then hit a goal on my list with Will Shortz (the goal didn't have to be with Will Shortz, but it was something he meant to do too) and walked down to the water to stroll along the beach and contact my third major ocean a couple days after finally hitting my fifth major continent. We had a good conversation for the half hour we walked north and south along the beach, and there were fireworks both where we were at Mobor Beach and elsewhere along the coast. I had hoped to chat with the English (I'd chatted with Ferhat of the Turks and Riccardo of the Italians already who were the other people I wanted to say my goodbyes to) but somehow the UK team was off to bed before I was back to the beach. No stimulating conversation with Simon to close the championship. No new Brain Age weight (or whatever the game on the DS is).

The next day, after some but not a lot of sleep, I went out early with my camera now back in my possession to take a lot of early morning shots and between 6:30 and 7:30 AM the resort was beautiful and the weather quite pleasant. I learned the meaning of Wei-Hwa's statement earlier on stepping outside - "now for two minutes of blindness" - when my camera lens was completely fogged over for my first set of shots until I wiped it clean. Some Indians were digging a big hole in the soil before the beach. Not sure what it was for. The hotel was certainly paying them to do it, but it seemed like just another thing the incredibly populous staff of the hotel had to do.




Some birds were attacking the leftovers of the buffet dinner. I got back to my room - Wei-Hwa still asleep - and decided to just hit breakfast. Caught a quick break to blog one entry here to announce the results since wsc2008.com had none at the time, then ran to my room to make my documentary appointment. That went ok, then off to lunch, the airport, and a long trip home.





I hate travel. The terminal I was at in Goa had its AC knocked out, and my flight delayed 1 hour, so it was a pain. I also had to go through the ticketing line 3 times as first my checked bag hadn't been scanned and tagged beforehand, then I didn't have a new tag on my carryon bag to get stamped after it was scanned. Fortunately, some other puzzlers were around to kill time as I dehydrated. Wei-Hwa was on a different short connection (at an even different terminal in Goa) but we remet in Mumbai. While there were many nice interior places to wait near our London flight gate for 7-8 hours, the length of our layover, we could not get to that gate until an hour before the flight when my checked bag finally arrived at the desk to have me verify it was mine so it could continue on its journey. During that time, my attempt at using a power converter to start my laptop convinces me I've completely broken my laptop as it cannot start even though the battery is at 3 of 5. Another bad thing for a long trip home - having to think your laptop is dead. I'd backed up all the data before I left, as I am a smart person who understands the fickle winds of fate and theft, but jeez. (fortunately, on the US power grid, it charges and starts ok.)

Maybe we could have gotten back to the gate and the comfy lying down chairs 2 hours before the flight if not for that checked bag which I had to confirm, but I was really about to blow my top at someone and meant to not do it towards Wei-Hwa when he, say, kicks my behind in Race for the Galaxy (although actually I was over .500 against him, if just barely, for the trip as I recall). I learned again I was not an aisle person, or even a "could you stop that" person during the Mumbai-London leg. I was inside of two grandparents who were sitting behind two parents and a young child (all flying together). My early observation was the grandparents, who I had to ask to move to let me in, did not speak much English, not that this mattered if I wanted to complain. Anyway, the granddad next to me had some incredibly bony elbows that particularly liked hitting buttons on my entertainment controls on the shared armrest between us. Yeah - not fun. Then, during the lights out night-time segment of the flight, the kid who before the flight had demonstrated he could stretch and get a lot of his arm through the seat in front to where I sit, proved in his sleep he could repeat this maneuver. I was trying to save a glass of water to maintain hydration at this time, and it all quickly rushed onto the table and a lot onto my pants after his arm knocked it down. Fortunately, just water, but I was at a window seat, and the grandparents next to me were asleep, so I dealt with the drying issues as best I could for two hours or so.

Heathrow was only a 4-hour layover, most in the new Terminal 5. I already posted in a comment on an earlier entry my thoughts of that terminal but I'll repeat them here: The automated systems for handbag checking in security was phenomenal. The problem is most people are nowhere near as sophisticated as the system and so while Wei-Hwa and I enjoyed the automatic return of trays, the automatic detection of when we'd collected our things, the automatic shunting of a "bad bag" into a second line for additional checking, ..., the people before us and after us in line were hardly fool-proof at using such a new system, showing that the world will always find a better fool. The architecture was alright and at times stunning, although I don't go to airports to get in touch with my inner Frank Lloyd Wright. The funniest moment (unfortunately I do not take photos inside airports as this tends to be considered illegal) was one of the HDTVs at terminal 5B displaying a high-def "blue screen of death" as the pc it was linked to had crashed. Seemed to make sense for telling the story of the terminal - still some bugs to work out, but everything next-gen.

The flight back to SFO was also long, and I was still in a Window seat, but my kidneys cooperated on this one (and it was also always a lights on flight unlike a red-eye so I'm better on those I guess. No middle guy, so also the first un-full flight. Sweet. I was on my second full book of trip, "Freakonomics" which I should have read ages ago, but it was very enjoyable. Still, my brain was shut off for much of the flight as so much lack of sleep was giving me a pounding headache. I think I successfully slept twice for about 1.5 hours cumulative. I also managed to drool on myself (while asleep - I hope). The most noteworthy moment of the fright (yes, I chose the r purposefully) came 2 minutes from the gate. When those blasted seat-belt lights are still on, you do not open the luggage bins above. Why? Because. Someone opened a bin, their bag popped down, in trying to catch it they directed it my way. It crashed into my left thigh and left hand. Fortunately, nothing but a bit of a contact bruise, but come-on. That's dangerous and stupid and you can wait like everyone else until you can even begin to think about getting to the flight door. It's not the Amazing Race (or if it is, I didn't see your camera crew).

So, WSC 3 is now in the books. The press attention was DOA this year. The one great failure of the Indian organizers was getting any Indian coverage in my mind - it might have been there, but I've discovered no Indian stories. The Times of London and the Australian documentarians were the only people I spoke with, a far cry from the diverse panel in Prague including the German "reliving the cold war through action shots on the Charles Bridge", the sex columnist, ...; all were missing. It could be location, it could be the new state of normalcy as we are decidedly out of the sudoku fad phase in most places in the world. Maybe the lack of press stories is why my account is so much more vivid this year. Maybe its because the organizers delivered some of the most outstanding puzzles I've seen at a championship when they had the freedom to do more than just classics and do things really special. Wei-Hwa's USSC puzzles were great classics, but these puzzles in Goa were above and beyond anything I ever expected. This is the book of puzzles you will want to buy. You will hate having to read this entry and not see the "Outstanding" puzzles I specifically labeled.

Thanks to the Indian organizers, and the other competitors. The world is still strong in sudoku, but throughout the competition (even with all my missteps on day one), and certainly in the playoffs, I asserted again that I am without equal. Whether you use qualification math, or playoff math, I now have 2 world titles. The challenge is identifying which person has the third world title.

I was worried that there wouldn't be a single champion, with the addition of the classics trophy, but fortunately the world got a single champion. I hope organizers of these events learn to prepare puzzles in anticipation of the "nuclear option", and choose to use classics more than before, but not go to 11 so much. You could really just renumber that 11 to a 10, and then stick at 7 or lower and be just fine. I know you all respect what I can do, but don't totally Tigerproof a championship until you know the steps are good for the game.

Cheers. Thomas Snyder, World Sudoku Champion (x2)