Saturday, June 18, 2005

Post for WPC Qualifiying Test Comments

This post is being made primarily for discussing the puzzles on today's World Puzzle Championship qualifier. The WPC qualifying test is one of a select few venues where there are a significant number of new or seldom-seen puzzle types, ideas or twists are presented, and I plan to comment on at least a couple of them over the next day or two as I find the time.

As always, your comments are also welcome.

16 Comments:

I was chatting about #2 after the test was over with someone who had written and as a result of the chat (and the suggestion that the solving process might be mostly trial and error), I whipped up an annotated solution to it. You see, puzzle #2 was one of mine, and after all, We Service What We Sell. ;-) I've listed the solution here, for your edification.

#1: Column 2, with 13 squares total, must have four pieces.
#2. Because unominos are not allowed, row two cannot not have four separate pieces; it must have three, and they must be 4, 4, and 3.
#3. Row 1, with a total of seven, cannot have a distribution of 2, 2, 3, because one of the 2's would have to poke its nose into the second row, which has a 4, 4, 3 distribution.
#4. Between the two of them, row 1 and row 4 contain a 4, a 2, and 2 3's. (they cannot share their 3's because they are too far apart.)
Additionally, there must be a second 4, because of row 2. 4 + 4 + 3 + 3 + 2= 16, so we now have a complete distribution.
#5. Given our complete piece distribution, 7 must be 4 + 3 everywhere (because we have only one 2) and 11 must be 4 + 4 + 3 everywhere (because we only have 2 3's. This leaves only one location for the two in the bottom row.
#6. The rest of the pieces may be placed by straightforward deduction with the above information.


As is often the case with original WPC-type puzzles, the idea came first and the implementation came afterward. Perhaps surprisingly, my first few tries at implementing this puzzle idea failed, not because I wasn't able to come up with solvable puzzles, but rather because the first configurations I tried did not have unique solutions - at least, not without some part or parts already filled in. I wanted to avoid filling in some parts to start the solver off because I was going for a "black cocktail dress" puzzle - classic, simple and elegant - and I already had simple puzzle idea at an accessible level. All I needed was an elegant presentation, and for that I wanted the small square grid to be empty.

(More to come on some of the other puzzles - on someone else's puzzles - soon.)

By Blogger Craig, at 10:25 PM  

Cheers for the interesting backstory on Piecemaker, Craig. However, given the elegance of the solving technique you envisioned, I'm almost disappointed to say that it wasn't really the way I went about it.

#1: I was kind of expecting to see a 5 from the instructions, given that it was the easiest possible number to break up. So that gave me a 2 and 3 in the bottom row.
#2: From there, I looked at the 13. That had to be four different pieces, since even three fours would only give twelve; more importantly, however, it was 3 less than 16, which was the contents of the entire grid, and so the only piece that wouldn't be in that column was a 3. Thus, the 2 was easily placed in its proper spot, and the remaining 11 in three pieces was obviously 4, 4, and 3.
#3: This meant that the grid's composition was 2, 3, 3, 4 and 4, and thus with the two already placed, the remaining totals shook down to unique combinations of 3s and 4s.
#4: I then proceeded to fumble around for the better part of ten minutes trying to figure out the four remaining pieces' position. Because really, it's not a Byron solve without managing to cleverly suss out the hard part easily, and then wasting time botching something obvious. The puzzle used its elegance to lull me into a false sense of security.

Come to think of it, the same thing happened with Crisscross Crash. I meticulously listed the number of occurrences of each end-letter in hopes to connect the pieces together logically; unfortunately, one of the first moves that followed was a real lulu. I figured ON THE WAY OUT needed extension, so I stuck THIEF on the end...and didn't question doing so until about 10-15 minutes later, after first realizing that the Down entries had 11 letters (not 13), and then proceeding to KEEP wasting time trying to make things link up properly with FETAGLIMMER (THIEF, of course, being in use. :) )

Fortunately, I did manage to get it right eventually; this sort of thing was the main story of my USPC, actually. Puzzles that I struggled on at first, and eventually managed to unbotch in most cases. Well, up until I ran out of puzzles I was actually good at before the last twenty minutes, most of which I wasted on the stupidest decision possible: Trying to brute-force solve the Icon Maze. After that, I promptly foolishly tried to guess a solution to Observers...and sent it in at the last minute. Bad, bad idea.

I could really sympathize with the US Open golfers that afternoon; I counted my probable score as 265, and it felt like the equivalent of a hard-fought round of 71-72 on Pinehurst.

A real puzzle-by-puzzle analysis coming later, probably; as for now though, cheers to Craig and the others for the challenging round. You guys gave me a good little bout of uncertainty!

By Anonymous Byron, at 6:05 PM  

I'm looking forward to the puzzle-by-puzzle analysis, Byron! (I'll be adding additional commentary of my own tomorrow.) But while I'm here, I should mention that as a puzzle designer, I make sure that every puzzle I make is solvable by more than mere trial and error - and I gravitate to puzzles where little trial and error is required - but I do not generally try to artificially force one single path to the solution. Some puzzle types naturally gravitiate towards a single path to solution, but others (like the crisscross crash puzzle) gravitate towards multiple solution paths.

By Blogger Craig, at 10:40 PM  

I'm going to keep commenting on the test slowly bu surely for at least a little longer, in hopes that I can eventually cover most of what I wanted to remember to say. Tonight I'm going to cover Picnic Lunch and Famous Hungarians.

Picnic Lunch is a cool puzzle based on a cool idea: what if you did a criss-cross puzzle where none of the words crossed? (Good what-ifs very often turn into good puzzles.) It was set up to present a fairly wide rage of possibilities to begin with, but as the solving progressed, possibilities eliminated themselves due to the placement of words by elimination and the like, and very little trial and error was actually required. The end result was a puzzle that was neither obvious nor hard, and that is a very good combination for a WPC qualifier.

Perhaps surprisingly, the Famous Hungarians variety word search puzzle left me cold. I should mention here that it's not the underlying concept that didn't sit well with me; rah=ther, it was a couple of major aesthetic things about the execution of the idea. The one that bothered me the most while solving was that the names to be found were both given names and surnames; when you're constructing a puzzle like this that requires a robust execution to make a good impression, it helps to put your best foot forward, and just choosing first or last name "randomly" for the puzzle feels sloppy or disorganized to me. The other thing, which probably affected the solving a lot more, was that there were too many too short name-words for my liking. The puzzle was obviously solvable anyway (I mean, I solved it!), but I wonder how much harder it to solve it was as a result of this.

I'm fading fast at this time of night. Off to bed with me for now, and I hope to comment some more tomorrow.

By Blogger Craig, at 10:44 PM  

My Criss Cross Crash break-in method revolved around the shortage of W's. ONTHEWAYOUT had to be either horizontal (linked with THIEF, if I recall) or vertical. If it was horizontal, the W had to cross something... it couldn't be CHINWAG, so it had to be SOUWESTER, which led to a contradiction.

Thus it had to be vertical, which again forced an intersection with SOUWESTER, and from there it was pretty simple to see where things fit.

Somebody else I talked to (Maelstrom, maybe?) did something similar using B's rather than W's.

By Anonymous Spelvin, at 9:12 AM  

Hello again! Firstly, congratulations to the American winners who destroyed my below-peak result, especially Zack Butler on his astonishing 385 which ensures that I have at least one more person to really worry about at the WPC. Now, without further ado, the first of 4 parts to my belated puzzle-by-puzzle breakdown of the test:

1. Battleships - 5 points for being bloody easy, and another 5 because Battleships are part of the bread and butter set of puzzle types that any real WPC hopeful should know well. This was a nice starter course puzzle of its type; I would imagine it probably helped many of the competitors get settled in. For me, however, it started a run of somewhat silly decisions, as I got impatient and started scrawling out the solution on graph paper, half-done by the time the page had printed. Still an easy 10, in any case.

2. Piecemaker - 10 points. My detailed thoughts were already posted above, but to summarize: Brilliant deduction of the 2-piece and the row compositions, absolutely crap follow-through. I actually felt it was more difficult than a ten-pointer after the test, but was quickly proven wrong as I administered the problems to my father and a brother, both of whom got the Piecemaker and neither of whom are likely to be seen on a WPC team in the near future. Apparently, it's easy and I'm a screw-up.

3. 7-Letter Hungary - 10 points. Where the Battleships has lately been the annual ease-in puzzle, the Crisscross is the second puzzle that an inexperienced puzzler really should tackle, as it rarely can't be solved without some pretty simple process of elimination on a pair of words with the same starting letter and lots of intersecting words (or even better, the square of words provided in the righthand instance of this one). Again, sort of bread-and-butter, as Crisscrosses are the skill set from which much of the interlocking-words type of puzzle need to be reached.

4. Icon Maze - 25 points, if you're able to get through the muddle of icons and the routes between them. Unfortunately, I didn't really give this one a go until there were 20 minutes left in the test and I ran out of things I was good at. Thus, where I should've done as I did afterward and solved it with good, proper logic to claim my valuable 25-point prize, I was entranced by the file and folder icons, and thus driven to pursue a DIRECTORY-BASED SOLUTION, in which I started at A on my graph paper and drew lines to where it could go from A, and then more lines further down the paper to where they could go from those icons, and so forth until I was hopelessly entangled and time was running out. Come to think of it, had I just put A-Z in a big circle on the graph paper and drawn arrows representing possible paths, I might've had a better chance. Logic 1, Brute Force 0.

This brings me to my first World Championship-Caliber Protip~!

Protip~! #1: One of the most important things you can do in a WPC/USPC-level competition is to plan on which puzzles you know you can solve, and which ones are going to hurt you in the end. This requires some experience with the various puzzle types, and a solid understanding of your skills and limitations; if you don't think it through fully, you can lose focus and have a chaotic, even disastrous round like my second at WPC 2003 in Papendal in which I spent far too much time on a Paint By Numbers/Fill-A-Pix pair of puzzles at the back of the booklet and missed the many easier points I should've taken first.

Back to the puzzle rundown:

5. Hungarian Goulash - 10 points, and 5 bonus because observation is a critical skill in just about any type of puzzle, so while you might (as I used to) think "Why are they giving away points for a bunch of difference-finding puzzles?", what it's really doing is penalizing people whose powers of observation aren't keen. This ensures that sharper minds move forward faster in the championship, and that the Americans win another team title. :)

6. Cross Sums - 20 points, at which I shout with glee upon reading the instructions: Cross Sums are one of my two specialties (Number Place being the other, so last year's Cross Number Sum Place for 35 points sure got a good reception over here). This Cross Sums was a toughie, well worth the 20 points; I did a chunk, hit an impasse, came back later on, got the diagonal solved, and got outta Dodge. I still haven't bothered finishing the rest, but that's more forgetfulness than laziness.

Protip~! #2: If you're doing the USPC, and you've got enough to fill the Answer Key for a puzzle, and you're sure you're right, don't worry. You can always finish the puzzle later; the rules specifically state that a "correct answer" submitted will earn the points. Mind you, I usually forget to take advantage of this fact, but in the above case, I was finding the solving too slow to waste time; there'd be plenty of that on the other puzzles!

End Of Part One. Turn tape over.

By Anonymous Byron, at 11:06 AM  

Part Two: Picnic Panic, Or: An Instructional Refresher.

7. Count Me In - 10 points. Oh joy, a shape-counting puzzle. Count me out. That's what my reaction was; it was perhaps the only puzzle I didn't do anything at all on, and in that regard, I regret it; I probably could've used part of that last 20 minutes to at least get pretty close to the answer, since this was a nice, to-scale type of shape and it seemed like a pretty easy thing to systematically solve. Which brings me to:

Protip~! #3: Sometimes, even a puzzle type you don't like can come out easier than you expect. Take a look at any puzzle that you don't hate with the passion of a thousand fiery suns; you might just get something. I dislike Hiroinomo puzzles, but the last couple WPCs, they've been there, and somehow I've solved them. Last year, I got an Adaptation puzzle (room-division sort, like on Erich Friedman's website). I never get those, but I gave it a try and I got it, and I was glad for it.

8. Picnic Lunch - 15 points, +5 because A) word-fitting puzzles are an inevitability; B) this puzzle is an example of why it always pays to read the instructions in detail. I saw "Words don't necessarily use one of the given letters, and some words may use more than one", and entirely missed the most important part: "No square is used more than once." So I wasted the better part of 10 minutes fruitlessly trying to stuff the words together in an interlock, despite noticing that there were 64 letters in the whole word list. It just didn't click, and I missed a 20-point puzzle that I solved afterward in about 3-4 minutes. And so, no better time than now to make:

Protip~! #4: Read the instructions. Really, read them thoroughly. I mean it. If you don't get something, look at the example (if provided). If there's something ambiguous, assume the most logical interpretation. Don't be afraid to discuss it with people at the WPC; in fact, if you're at the WPC and you don't discuss the instructions with at least one other person, you're only hurting yourself. In 2 days of puzzles, there's usually bound to be something to ask about. That's why we have the puzzle instructions discussion sessions. There are very few stupid questions. Not even if it's WPC 2000 and everyone's getting a crash course in Mr. Potato Head anatomy. "It says that the words can jump over one square, but they can't jump over two. Can they jump over three?" Not a stupid question. A funny one to experienced solvers, but a legitimate one.

9. Division Subdivision - 15 points for substance, 5 points for impeccable style and invaluable math skills. Ahh, math. That's more like it! Pull the divisors and last line numbers out from the front and back of the line, and work from there; it's a nice, fun type of problem. Unless you hate math, in which case you probably don't want to touch it with a 39.5-foot pole.

10. Spinners - 10 points. The rotational-symmetry thing's been done before; in fact, I quite liked it the last time. This time, however, I couldn't really get going. I couldn't figure out where to start, and how to get forward from the places I did try starting; I could figure out in about four minutes, however, that if this stuff involved order-3 symmetry, it would be even harder. I quit early; it involved order-3 symmetry. I claim a minor victory.

11. Arrow Ring - 10 points. After doing a speed-run of Arrow Ring puzzles earlier in the week from a Nikoli book at about 3-5 minutes each, this one seemed to take me longer. It pretty much came down to understanding the puzzle type, and then to figuring out where black squares couldn't go if the loop was to be single and closed. This type has seen very little exposure Stateside to my knowledge, so for intermediate puzzlers, this one should have been thought of as an "If I have time" puzzle; for inexperienced beginners, it shouldn't have been thought of much at all. In either case, it would serve a puzzler with WPC aspirations well to give the puzzle a good try afterward, as developing an ability to tackle the unfamiliar is a significant advantage with the broad range of puzzles covered over a two-day period.

12. Factor Maze - 15 points. It's like a regular maze in solving, really. Just go down the different possible paths until you find a dead-end, then don't go down that path. The problem, however, is that with a running count of the current number and the previous one, you can grow very disoriented very quickly. Use arrows on the puzzle to keep track of the correct path and direction, so you don't lose your place. It ends up crossing over itself and through itself quite a bit, which doesn't get mentioned in the instructions; however, since you can only get out via the 3 and 4, it's only fair to assume that crossing the path is not a big deal. Fun for me, aggravation for the poor family members I inflict it on afterwards.

Part II is dead. Long live Part III!

By Anonymous Byron, at 12:01 PM  

The Madness Of USPC Part III

So far, you've seen me botch the early high-point puzzles, skip the tough-looking 10-pointers, and waste time in general. I don't mean to undercut my apparent reputation after last year's performance, but this probably hasn't helped. :)

So, in the interest of self-preservation, I present a special bonus World Championship-Caliber Protip~!

Protip~! #5: Consistency was a good thing. At least for me. If you look at my results in WPC 2004, I spent the entire championship hovering around fifth place; thus, I came into the playoff very near to fifth place. Making a solid round spectacular is always nice, and a great ego boost; but while a good round's momentum can carry forward into excellent results, so too can a bad round just get worse if you're not careful. I wasn't quite careful enough in this USPC, and it's going to show fairly soon in this report (if it isn't already self-evident).

13. Crisscross Crash - 20 points for difficulty, plus 5 for the fun that always comes with overlapping things for a solve. Unfortunately, when you don't pay quite enough attention, you run the risk of confusing 11-letter spaces for 13 and driving your efforts right off a cliff. Fortunately, there weren't all that many ways to screw up before I realized my mistakes, and it became easy again. HERBREZHNEV was an easy call to make, as was CHINWAGETHIGHON, but unfortunately, I didn't have the sense to try sticking them in the grid with other possibilities, opting instead to concentrate on things like COPYRIGHTEDRAFT and ONTHEWAYOUTHIEF which really weren't givens at all, but got mistakenly treated as such early on. Solved, but humbling, as my momentum just didn't get going for a lot of this test thanks to mistakes like that.

14. Zeros - 15 points. As 15-pointers go, Zeros was a nice, easy, fun one. Almost a freebie, really; just figure out what numbers have to add up to each row and column, cross-reference a little for the tougher ones, and fill in the grid. A good confidence-boost in an up-and-down test, and worth a try for pretty much everyone in the course of solving.

15. Observers - 15 points, or -5 if you do like I did. Funny enough, I just solved this one here by intuition as I type this. Unfortunately, my intuition was flawed when compressed into the space of 3 minutes at the end of the test, and so instead of double-checking in the final minute, I tapped out an answer key frantically and submitted. Thus:

Protip~! #6: If it's the last minute, and you're in a test where incorrect submissions will count against your score, even if you really want the points, you should double-check, then stick the answers in, THEN hit the server timer to make sure you've got time left to submit, and if you can, do. If you can't, don't. Not worth panicking and trapping yourself in a puzzle you won't solve if there's forty seconds left. A zero looks prettier than a negative number anyway.

16. Three Coins - 20 points. This is one of those cases where puzzler's intution is a good thing. I didn't think about formulas, or math much, or anything. I didn't realize that given A, B and C, only combinations AAB, ABB, AAC, ACC, BBC, BCC, or ABC could make the three numbers given. I just thought "29, let's turn that into a pair of one and one of another and see if we can make the other amounts with some of those and a third value." And working my way down from 14-1, it wasn't long until I hit upon 11-7, and from there, the third value 6. It wasn't anything sensible. It was intuition. And sometimes, a puzzler's intuition is his best friend.

17. Groundhog Day - 15 points. I much prefer the matching tiles type of puzzle to the matching big pictures sort that used to be present in USPCs; there's a greater concentration on making the sort of observations that come in handy in other puzzles in this one. Generally, the rule of thumb here is to look through each tile with its gophers in a specific alignment, and either find a match or cross out the lot of 'em. Another gimme 15, and after trying a lot of puzzles for the first time and not getting them earlier on, I'm not going to argue with a gimme 15 at that point.

Come back after the break for the thrilling conclusion to USPC Countdown! Or, rather, Countup. Silly ascending numbers.

By Anonymous Byron, at 12:39 PM  

USPC 2005: The Thrilling Conclusion (or, technically in my case, the chunk I mostly did between an hour in and an hour from the end)

18. Numerical Jigsaw - 25 points. I don't like Numerical Jigsaw. I didn't make an honest effort at Numerical Jigsaw, even though I had an idea what I was aiming for in one of the middle strips. Unfortunately, I should have, because yet again, intuition would have likely paid off. But my BIG mistake on this one was truly simple:

Protip~! #6: If the puzzle gives you...I don't know...a bunch of pieces to rearrange, and you've got the right to use special tools, like oh, say scissors, you might want to consider using the scissors. Makes it a lot less difficult to consider the possibilities, especially when it's worth 25 points.

19. Digital Number Place/Sudoku. If there's one thing I do better than Cross Sums, it's NP/Sudoku. I would go so far as to say that should a championship ever be held comprised of NP/Sudoku puzzles, I would give Britain's best at the very least a run for their money. And I'm not just saying that in hopes that Chris Dickson will read it and mention it on his LiveJournal, and word will spread, and Britain's finest will be outraged that the Canadian upstart DARED to declare himself no lesser a Sudoku master than any of them, and the British newsmedia will organize such a championship and fly me to London free of charge.

Ahem.

19. Digital Number Place/Sudoku - 25 points. I'd seen one previously on a PQRST test (which I generally always try, but rarely submit answers to, given that I seem to lose rather badly in those sort of competitions), and it was disappointingly easy. This one, however, was very, very satisfying. It's like a regular NP/Sudoku, except that the clues do not give you any numbers, merely eliminate some of them from certain squares' choices. From there, it becomes a question of placing numbers that can't be placed anywhere else in their row/column/rectangle, and 1s, being forbidden from a great deal of squares, are the logical start.

20. Famous Hungarians - 30 points. Ah, the annual overvalued word search; but wait, thirty points..for this? Oh my, they've thrown so many small names into the list that it's actually rather worth the 30! This one was more difficult than usual, with a lot of names crossing into the central abyss, and a well-distributed grid. Even though I solved it, I'm still crossing my fingers, hoping that I didn't make a transcription error on the submission. I don't THINK I did, but I have a memory like a sieve sometimes.

21. False Field Fences. Yeah, that seems the right number of F-words. This puzzle was a very irritating beast to tame, as my solving strategy consisted of "Try this and see if it leads to any contradictions." Educated trial and error, really, was the order of the day; not basic trial and error, which makes for poor puzzles, but the sort where you try to do something, and have to use proper solving techniques from there to see if it can possibly work. Then, if it doesn't, on to the next option you go. I suppose this is the sort of brute force I employed a couple times during the test; I don't like brute force as a rule, but in my case sometimes it does the trick.

22. Toroidal Number Place/Sudoku - 30 points. This, sadly, was my least proud moment of the test. I started solving logically, built up a good head of steam, and found myself at a contradiction a couple minutes in. However, instead of trying to figure out where the contradiction had wormed its way in, I noticed a pattern. As a puzzle purist, I don't like pattern solving. It defies the overall concept of solving the puzzle logically, especially in NPs if the same trios of numbers keep sharing rows inside boxes. But this was a timed event, and I figured it was worth a shot, so I erased the whole chunk of work I'd done, and started solving with one basic rule:

In each shape, the top three digits are 6,7,9 in some order; the middle three are 1,3,8 in some order; the bottom three are 2,4,5 in some order.

And to my continuing surprise, this worked. I mean, as I think about it, it had to certainly satisfy the shapes and rows; they never deviated from the pattern in the numbers already filled in. But the columns were not a given, and so this was not a reasonable strategy to pursue in the course of logical solving. But it worked, so hooray, 30 points!

Overall, this test was not the Pinehurst No.2 I quite imagined; I still took one over par, but all the while, I'd been doing it on Augusta, the course only made really difficult by my own errors. But it doesn't really count against me as much as it may seem from the rundown. Had I met my full potential, I could've used that 20 minutes of one incorrect solve and nothing else at the end to solve Picnic Lunch within 3-4 minutes, do the Icon Maze logically within 10, and maybe even Count Me In if I had the time. That means that where I had -5 from what'd been a 270 worth of solves, I could have potentially obtained another 20, plus 25, plus 10, for a Byron Potential of 325. That doesn't feel so bad. Besides, there's always Protip~! #7:

Protip~! #7: Luck is an inevitability. Every puzzler has ups and downs; and every puzzle naturally clicks better with some puzzlers than with others. You have to ride the waves that come with it, and look at the short-term picture as a piece of a long-term success. You had a bad round? Don't worry, the next one might be more your type. You had a good round? Congratulations! And in most cases, the good will generally outweigh the bad, whether you're battling it out for World Championships or just setting personal bests. So keep at it, and to all the puzzlers and potential puzzlers out there, good luck!

I'd just like to thank Dan Katz and Chris Dickson for their respective blog-namedropping; it's always flattering to get a shoutout from people who really enjoy solving puzzles, and a great rebound for what was originally a slightly-wounded ego after that test. Thanks as well to Craig for the post that encouraged this whole series of comments, as I wouldn't have thought to write a full rundown without a space for commenting on the championship being created.

I shall finish this post with the most useful World Championship-Caliber Protip~! for all manner of competition, from the recreational Sudoku-solving duel to the WPC itself:

Protip~!: Don't make errors!

;)

But really, have fun. It's the least you can do.

By Anonymous Byron, at 1:42 PM  

Thanks for posting your extended commentary on the test, Byron - it was both informative and extremely enjoyable. As someone who hopes to tag along with the Canadian team someday, I especially appreciated the protips and found that at least some of them squared with my experience with the test. (You're not the only one who didn't read the Picnic Lunch instructions closely enough, for example.)

Random thoughts here and there while I still remember them:

I noticed the toroidal sudoku pattern about, hmm, a little more than halfway through solving, and used it to finish the rest of the puzzle really quickly. I was kind of hoping someone would notice it sooner and use it to their advantage; I'm pleased to hear someone did.

I also really enjoyed the digital sudoku - what a great twist! And it resulted in a pleasurable progressive solving experience to boot.

For the Criss Cross Crash puzzle, my designated "in" was the Z and V combination in BREZHNEV, although I consciously made sure there were a few other low-frequency letters in the mix, while taking pains to make starting and ending letters repeat. I think the combination worked quite nicely to even out the difficulty to a suitable level.

My experience with the observers puzzle was almost unnerving, probably because I understood the structure invovled in the puzzle almost immediately. Because of the number of observers that could not see all the numbers, some numbers had to be completely hidden behind other numbers. Knowing that, it was not difficult to place the requisite three-in-a-row alignments and distribute the numbers. I think what actually was unnerving me was the feeling "there's got to be more to it than this..."

One last comment for tonight, on the spinners puzzle: I wasn't watching out for any 120 degree rotations, but somehow managed not to get bitten by it by placing the one 120-degree piece last. I think I managed to get away this mostly because it was such a small example of this puzzle (with the edges necessitating more conventional 180-degree rotation pieces.

By Blogger Craig, at 10:57 PM  

Well, the odd thing is that two newspapers (the Independent and the Daily Mirror) are hosting rival Su Doku championships, with postal qualifying rounds and face-to-face finals. You could do worse than getting in touch with either or both of them and seeing if you can wangle an invite. The Indy haven't been talking about this on their web site much but see http://www.mirror.co.uk/news/tm_objectid=15512533&method=full&siteid=94762&headline=the-great-su-doku-challenge-name_page.html for the Mirror's effort.

There is due to be a live Su Doku game show on July 1st filmed in London - see http://www.starnow.com/Audiences-wanted/listing_detail.asp?l_id=4264 passim. Unfortunately I don't know which production company is making it, but you may well be able to investigate further.

No guarantees, but if ever there were a good time for a grandstand challenge at global Su Doku mastery to be made, now would seem to be it.

I'll see if I can propagate your challenge further soon.

By Anonymous Chris M. Dickson, at 11:37 PM  

Thanks for the good words and advisory respectively, Craig and Chris; they're both much appreciated. I'll see about getting in touch with the papers, but my grandparents' wedding anniversary is going to conflict with the game show in any case. Memo to self: Grandstand more.

The sad thing about my Observers result is that I know the basic structure of the puzzle. Figure out what each one can see, figure out how to place things behind other things (usually beginning with what you're given), get a fully-satisfactory answer. But for some reason, no matter how much I understand the puzzle type, I simply cannot solve it under regular solving conditions. It's like the memory disappears whenever I'm confronted by one. With my luck, it'll appear in the next WPC playoff.

Two erratum for the commentary; I accidentally labelled two Protips each #6, and on a more serious note, my scoring summary just came back, and I'd apparently failed to follow my Ultimate Protip~!:

I forgot to count the Cross Sums on my score. 285! I suppose I've made par after all.:-)

By Anonymous Byron, at 12:05 PM  

People who are interested in this may well be interested in the new World Puzzle Championship community weblog. All WPC fans are welcome to comment or post. (If you would be so kind to list this as a puzzle blog in the sidebar then I'd be very grateful!)

By Anonymous Chris M. Dickson, at 12:54 PM  

WPC?? Wow never knew this was so serious. Good Luck everyone.

By Anonymous Kenneth, at 11:54 AM  

Who knows where to download XRumer 5.0 Palladium?
Help, please. All recommend this program to effectively advertise on the Internet, this is the best program!

By Anonymous Anonymous, at 2:51 AM  

those test are very hard !

By Anonymous rubix test, at 2:45 AM  

Post a Comment

<< Home