As mentioned in my last posting, I've been fascinated with vowelless crossword puzzles recently. I decided to try to make (and finally publish here) a word puzzle on one of the grid variants I use all the time for sudoku.

As in Frank Longo's book, the either/or vowel nature of Y is handled by never having an answer using this letter. All the consonants in the answers for the clues go into the grid; all vowels are discarded. Arrows are provided alongside the box numbers to help orient your entries into this cube; entries "wrap" around the faces of the cube.

Enumerations of the clues are provided as spoiler-text in the comments for those that need a little more help.

Enjoy, and have a safe 4th of July holiday!

My current puzzling fascination is Frnk Lng's Vwllss Crsswrds. I must say it is taking up a fair bit of my time nowadays as I'm on a bit of a (self-imposed) break from writing/solving logic puzzles. I'm tempted to try writing some myself - I know I will have a super tough time getting the wide-open fills Frank does, but it seems an interesting tapestry for cute themes and the power of onelook.com to come together. Maybe I'll do that for a future week.

This week, I'm recycling a puzzle I wrote last year for some friends that could fit into a Puzzle Hunt somewhere, but where fresher ideas have now pushed this one aside. It fits into that favored category of puzzles that went missing from the USPC this year, the word search. As with most Hunt puzzles, there are a set of steps to perform that will lead to a single-word answer. For those that just want to know a little more what to do, I'll eventually put the rules in the comments below.

nickbaxter has made a rare posting, but they always are USPC reports so this one is worth checking out.

Also, one of my calcudoku imitators has put up another solid variation. This one solves very nicely, like a hand-crafted puzzle should. Go check out cyrebjr's "Cryptic [Calcudoku]" here.

Should have slept more than 5 hours instead of taking this test. Answer entry on comet threw me a bit (it also wasn't clear that multiple [non-horizontal] segments would be in each row [column] from the example). I clearly wasted 15 minutes on the hardest diagramless kakuro and never got it; my ability to do math is compromised at these times of sleepy solving even in a puzzle type I've already thought about construction-wise. It seemed these were well constructed and push the difficulty envelope much more than my first diagramless from 2006. Its still a hard puzzle type compared to others in the set.

The optimizer was average but I cranked out a solution. Relatively untouched were Hamle 34, diagramless 23, news 124. Feels like another 3rd to 5th place finish with a high likelihood of data entry errors. Disappointing compared to what I'm expecting to see tomorrow when USPC results come out. Its not that the puzzles were bad - they actually were quite good - I just am burned out by these "wake up to do a test" mornings.

ETA: Apparently a lot of people struggled (and others didn't compete) so I got a 2nd. Congrats to Mehmet Murat Sevim (who has been close several times but finally pulled one out).

Just to prove there is no rest for the weary, Saturday is the 6th OAPC; this one features more puzzles in the style of my diagramless crossword/kakuro variant from ~three years ago (my puzzle serves as the example). I'm 50:50 on sleeping through the start and doing it later versus competing again, just as I feel like sleeping through the Hashi time trial tomorrow even though I have a great shot of winning it.

For this Friday's puzzle I share this Inside/Outside Corral, my version of the 40-pointer on the recent USPC that I wrote before the test to learn some things about the type and how it differs from standard corral puzzles. Mine is not as hard as Dave Tuller's, and I was certainly not trying to be minimal in the givens, but I hope you enjoy this challenge.

Rules (as on USPC): "Draw a single closed loop along the grid lines. If a number is inside the loop then it equals the count of interior cells that are directly in line (horizontally and vertically) with that number's square, including the square itself. Similarly, if a number is outside the loop then it equals the count of exterior squares that are directly in line with that number's square, including the square itself."

When I finally got back to them, it took another 15 minutes (I had totally misassumed a forced lower-right corner that I couldn't) to do Sheep/Fences and ~7 minutes to do Sweet Sixteen. So, like all USPCs, there was a 30+minute question for me. I just didn't do the faster puzzle before I got stymied on the high time challenge.

In other news, I finally received my preview copy of the Sudokumentary (Colours by Numbers) that covers the Australian team's exploits at WSC3. I'll be arranging some kind of showing here soon for those who might appear in it.

Finally, my Thomas Snyder Outdoes the NYT KenKen posts just gained the attention (by this I mean a posted reply) of one of the key people behind the NexToy KenKen(TM) line of puzzles. Its worth a read, at least for a different view from inside a company that makes money by supplying the puzzles en masse to book and magazine publishers, iPhone app developers, ..., who want hundreds to thousands of puzzles at a time. In a given year, most solvers could probably deal with 1 or 2 KenKen a day, so the argument that hand-crafting is impossible given the market fails a little bit there to me, but then my market is puzzle solvers and NexToy's (direct) market is puzzle publishers (and they earn more by spreading the wealth around to competing companies, fueling the "craze"). The fundamental question to me is do we need dozens of KenKen products, more than even I could stand to solve, if the quality is not what it could be? My answer is no (as is the case with Sudoku too, but you know that already). Anyway, davidlevylondon's comments are worth reading as a counterpoint to my own so I point you back at the discussion on the math of calcudoku that earned his reply.

Tentatively, 360 without Wolves and Sheep in Fences (spent last 15 minutes trying to get it, had what I thought was an answer, entered it, checked my work, saw a mistake, undid, ...) and Sweet Sixteen (I hate math puzzles on new grid shapes). This was indeed a hard test with a lot of challenging but solid puzzles like the Inside/Out Corral, the Four Square, .... I expected a list of USPC solvers with R's in their names for Magic Puzzle'Rs. I got something close - puzzle authors and me.

Full Report:
( Read more... )

So, I'll just "live blog" my thoughts as I read through the instructions for the first time. With the usual password fun9ess appearing even for the instructions, I expect another great password on the lines of the key lime pie one for tomorrow's test.

1. 23 puzzles (last year 20), total of 400 points (last year 365). I've never been over 400 so this will be the year to get to there with my highest score to date. The addition of low-valued puzzles seems to account for the change. The old standard of adding bonus points to puzzles to encourage their solution seems to have really fallen out of favor.

2. The Nikoli addition does mean nikoli versions of many of the common types (so US designer versions of "variations" on those, since Nikoli prefers the pure form of their puzzles with as few instructions as possible). So, Sudoku, Corral, Masyu, and somehow a 25-point Yajilin will all be on the test from our Japanese friends.

3. I am much less excited, as any reader of this blog knows, by having an "official" Calcudoku (#3). The "non-associativity of subtraction/division" rules are in play so, yeah, don't get confused by my missing operation puzzles where I recognized non-commutativity is the actual problem and this two-cell constraint is silly.

4. No word search (my favorite type is gone!) but several word fill-ins including the now ever popular Magic R version of a word fill-in that's been on Turkish and German tests too (although with no touching constraint and a single diagonal direction allowed as well).

5. A counting "puzzle" (screams NO!!!!!!!!). The author doesn't even want to be known - for shame! Until a name is given, I'm assuming Baxter.

6. A Craig Kasper puzzle that involves diagonal information, like his past Atomic Fusion that "spoiled" my 2006 (screams DI-AGONY!!!!!!). His other contributions (word fill-ins/criss-crosses) should be fine.

7. The Friedman puzzles seem approachable (I say this at my own risk of foot in mouth disease). Coordinate pairs looks like dot triangles but only has line segments so I bet I can do it. C Notes is exactly like the 100 puzzles that have occured on WPC Bulgaria and Belarus. The 2-3 Maze is probably the unknown variable, but I deal rather well with logic mazes.

8. Eminent D'OHmain looks like fun (and it looks like Adam, if I didn't already know authorship, enjoys capitalizing parts of his puzzles titles given last year's SuDON'Tku and this year's SuDUOku). SuDUOku seems another tough twist on last year's version. I have a lot of surplus/deficit sudoku experience from working with Wei-Hwa on these kinds of puzzles, so I think I'll do alright. I won't advise how to solve this puzzle yet though until I make one as I know last year I threw people off with SuDON'Tku.

9. Triangular Skyscrapers amazingly uses a tiling constraint I put into my "Mosaic Sudoku" in the upcoming Sudoku Masterpieces where each triangle orientation has only one of a digit. I also put in two colors with a labeled set of pieces so you'd know which ones of a particular triangle shape may be a particular digit but I doubt that additional use of color will go on here. Since I got that to my publisher before Aziz's puzzle here and since he would have no way of seeing my puzzles, I think we tie for the concept of a triangulated square grid that works this way. The Four Square variation looks like fun too (although here is The Typo(TM) in the instructions, as Skyscrapers is #8, not #11, since I definitely remember what #11 is). While I realize the dark squares indicate the turning snake, did anyone else initially wonder if primes would be important for this snake before reading the instructions?

10. Tuller follows up his fabulous Sheep in Fences last year with the addition of wolves (next year, obviously wolves dressed as false sheep!). Now the inside/outside shading rules of slitherlink that tend to be underused will be very important. His inside-outside corral will also be a good challenge, and I really look forward to it.

So, overall strategy/goal is still obviously to finish the test, but I'll be banking the big pointers early. Last year I did Spot The Differences back to the front and then did the hardest puzzles going forward. I imagine the same here. I'm sure I'll skip over the -5 pointer #11 until the end. Best of luck to everyone.

Yesterday someone pointed me to the French WPC Qualifier as additional practice. I took a quick peek and was pleasantly surprised to find that there were more Hashiwakakuro and Total Masyu on there in the list of puzzles. Then I looked at the puzzles and saw that my blog examples had been reused in entirety (without credit given) and authorship given in the solutions instead to Jean-Louis Legrand. He stole my mystery KenKen too. A little research suggests he has copied my styles before (he did a Mastermindoku in 2008) which is fine with me. Copying styles is gray area, and should give some credit to the originator, but I welcome any new puzzles in ideas I've shared here. Copying my exact puzzles is not! I'm seriously upset by it, disappointed in the French organizers as it compromises their championship to have 5 puzzles that were already out there in the public, and while I've sent email along the proper channels to get the details behind this matter (and will get the authorship issue fixed), I felt the need to vent my frustration here as well.

Anyway, for today's Friday puzzle, when everyone will want to be discussing the rules for Saturday's test soon, I thought I'd do a variation on two types that seem to be on the USPC every year in recent memory: Yajilin-Battleship.

In these two puzzles (the first very easy, the second less so), place the fleet of ships into the grid so that no ships touch, even diagonally (note this is different than the yajilin black square rule where diagonal touching is allowed), and so that each numbered arrow gives the number of ship segments in the indicated direction. The remaining empty white squares must form a single closed loop. Note: In the second puzzle two ships and two seas are given to you in the grid; the "seas" are necessarily in the loop.

Next Saturday is the United States Puzzle Championship. On today's NPR Sunday Puzzle, Will Shortz announced the upcoming competition and, as the past champion, I got to announce the prizes for this week's player. Anyone looking for me saying "KenKen" for some kind of music remix can find it there, as well as my favorite book plugs from Liane Hanson: "Battleship Sudoku ... I can't even imagine" followed by "Mutant Sudoku ... even better".

Next weekend is also the second edition of the Sudokucup. Like last time, when it was during the MIT Mystery Hunt, its scheduled over a busy weekend so I'm not sure I can compete but it comes highly recommended. The test can be started anytime over a 48 hour window on Saturday the 20th or Sunday the 21st and then answers returned within 2 hours. Czech sudoku champion Jan Novotny has put together all the puzzles this time so I look forward to seeing his offerings.

Been doing a lot more Nikoli paper puzzles recently (online solving is not the right practice) as well as test-solving some of the German finals puzzles for next weekend. Today, at 10 AM, with my normal musical cue lead in, I took the German Qualifier from a month or so ago under USPC conditions. It was a great test. Comments and times on puzzles after the cut:

( Read more... )

There are a lot of silly, nonsensical sports cliches that need to die. A couple years ago, when my hometown actually had a hockey team that made the playoffs, one such overused expression that bothered me during hockey's broadcasting was that a 2-goal lead was "the most dangerous lead in hockey". This never made sense to me on logical grounds as a single goal lead must be less secure, but the "logic" went that a team was more likely to get "too comfortable" with a 2-goal lead and squander it.

Anyway, this year we have fallen victim as a nation to the overuse of the cliche "the series hasn't started until the home team loses". According to this cliche's logic, tonight we have the chance to witness Detroit winning the Cup in a series that was "over before it even started" or Pittsburgh raising the Cup, ending the series when it was "just getting started". Either way, I hope the cliche is put to death based on its clear failure here. Personally, given the advantage of last-change on lines, I'm voting for "over before it started" happening tonight, against personal preference.

As I'm now in a much more relaxed month than May, and I miss sharing puzzles here on my blog, I've decided to begin a new series where every Friday(ish) I will post a puzzle here to solve.

Many of the types will likely be sudoku similar to those in my books, including Mutant Sudoku and Sudoku Masterpieces where my co-author onigame has already shared some of his "extras" online in what he's started to call the "too hard to be published" series, maybe because I often give that advice when editing his designs.

Other types I write may be common puzzles or just a new random variation I want to experiment with. Regardless of the type, while the daily regimen in the KenKen project was unsustainable given my day job, providing one puzzle a week will not be, so I should be able to sustain this for awhile.

Today's puzzle is from a type I conceived of 2+ years ago alongside my first version of Paint-By-Number Place. Basically, I dislike 16x16 sudoku and strongly dislike 25x25 sudoku (why Nikoli why) but the larger grids can have some interesting properties. What if you only needed to enter 1-9 on those larger grids? The result is Extra Space Sudoku.

Instructions: In the puzzle below, fill in the white cells so that 1-9 appears exactly once in each row, column, and bold region. The colored cells will not contain numbers. You can find more Extra Space puzzles in Mutant Sudoku, coming out in less than three months.

So last weekend I took part in Decathlon 5, with a team including my best friend, currently in Boston, and some Stanford buddies. Aside from me, the rest were Bay Area event rookies. This meant I expected we would do okay, but likely not win. While this prediction proved true, I had a ton of fun that weekend.

( Read more... )

Dear NBC Sports,

I am writing this letter because today you went from being on notice to being dead to me. It is clear to me that you understand nothing about sports and should no longer be considered a relevant national network for the televising of sport. To refresh your memory, being on notice was a status you earned after the hypocrisy of tape-delaying the Beijing Olympics for the west coast after specifically requesting the Chinese organizers move finals of events to the morning so they could be live in the US - not that I didn't already resent you for pushing the overtime of a Sabres-Senators NHL playoff game to Versus for pre-race horse coverage. However, today you showed a complete disdain for the meaning of sports and so you are officially dead to me. Simply put, while further "tape-delaying" of sporting events is a continuing insult with your coverage, it is beyond unacceptable to completely cut out relevant sections of sporting events because they do not fit within the comfortable time frame you set so that everyone could see Martha Stewart at her normal time. I should not, after watching a first set of tennis 5+ hours after it occurred at Roland Garros, be teleported magically through time to a third set of tennis 4+ hours after it occurred at Roland Garros with a commentator stating that there is simply not enough time to bring us the second set but something exciting happened in it towards the end (with no video to prove or disprove this notion). You specifically paid for the rights to televise this event - if it is such an inconvenience to you to actually televise sport both live and unedited then save yourself the money of buying such rights.

Now, the other companies that are dead to me (Verizon, UPS) are easy to deal with as I can both stop using their products and can explain to others exactly what "brown" can do me for me. Trust me, they can hear me now. As a fan of sport, it would be a hard pill to swallow to not watch future sporting events that accidentally happened to air at some time on your set of networks (whether pre-empted by Heidi or not). So, your punishment must be more tailored to your crimes. Since your business profits from your advertisers, I will specifically never use a product from any company that ever advertises during your sports time and I will make that known to anyone who asks. Can I somehow live without Franklin Templeton Investments, Geico, and ED pills? Only time will tell...

Did Shinteki Decathlon 5 today with my best friend from college and a connected group of other Stanford friends. Since the event is running next weekend, I'll post the madlibs version of it now. It was [adverb] [adjective]. [Pronoun] [verb] [noun]. [Emoticon]!!!

Big task tomorrow is to finishing some editing of my current sudoku book with onigame. The experience writing that book was also [adverb] [adjective], but in different ways.

I always carefully check the updates Nick makes to the wpc.puzzles.com page, including the rules, to see what I can learn about the test before it occurs. Well, this year, on the day the announcement of the test was pushed, the text thanking Nikoli changed on the website as well. It now reads: "Selected puzzles for the USPC and the Sudoku National Championship are provide by Nikoli, Japan." where it used to just mention the Sudoku National Championship. My guess is this means some authentic Yajilin and Masyu if not other types this year. I'm excited (because, at the very least, this means the way I normally solve these puzzles will work unlike the computer-generated Masyu in 2007 that tripped me up a bit).

Worst performance so far, at least in terms of finishing puzzles. I think the 7:30 AM versus 9+ AM start time had a lot to do with it.

I liked the types on previously on OAPC but spent about a half hour on them and only got the first 3. Wasted time.

Tripod sudoku went better than last time I saw it and I got back on track.

Tried to do the first Akkoy, kept breaking things, realized I should have read instructions, etc. Skip.

Got the first three psycho killers done and then skipped the last after proving I was making lots of mistakes on it.

Half-Life was pretty fun, although the linkage between grids was unusual to say the least. The Snake from last time was much more awesome in comparison, but I love classic types so I don't mind seeing this.

Magic R was my favorite of the test, but then I really love word-filling puzzles which correlate well with WPCs I do well at versus WPCs I do poorly at. Observational puzzles and word puzzles are good for me and I'm glad the Turkish designers have been experimenting with these types.

Toplamatik was ok but seemed to heavily reward good guessing if you spotted a good set of interlinking constraints. Got the first three done (and almost the last as time expired).

Wiring is a variant of one of my least favorite puzzles (Spokes) so I never looked at it, and didn't have time anyway. I wonder if it was really overvalued though.

Clock was running down, so I did the high-value Persistence of Memory snakes (3,4) then did the highest value Ikebana. Worked on the unfun optimizer - got something that was 27 away (probably not good enough for top 3 scores but still worth the chance as it could be 18/16/14/11 points, then took 15 minutes that I had left to salvage the test a bit.

This meant getting both remaining snakes (2-3 minutes total), the last psycho killer (finally - another 3 minutes), and the last 2 Akkoy's including the grossly overvalued 7.6 pointer. Got close a couple times to the last Toplamatik but didn't finish before extra time started and I don't like submitting answers after extra time anyway. I did add comments there this time - lets see if that affects the score.

So, a lot left undone, some types never looked at. I think I never got into a rhythm in part because I woke up mentally with about an hour left in the test. This was the smoothest test so far on the admin side (no errors, started immediately on time with the new server), but somehow was the least satisfying to me almost entirely because I had 11 puzzles left unsolved at the end of time and that's not happened before.

[ETA: Reviewing my answers, I left out the last island when typing out the last column of Half-Life. It was given (a 2) which might be why I did WBBBBWB and not WBBBBWBW. That will hurt if graded sternly. I also didn't read the instructions on answer entry for Magic R well. I have all the first letters for horizontal words, then vertical words, in top-left to lower-right order always. This means my left to right order is not their's in the second part because I started at the top-left and went across and down and not at the left and go solely right. KRM, AKTRORU for example (my answer) can clearly be better read as KRM, RKUTRAO in their format. Stupid! Thank goodness USPC is at 10 AM as this was super sloppy on my part.]

[EFTA: I have a different answer for Step By Step that looks fine to me on paper and puts a 40, not a 41, in the grey cell in the fourth row. I've rechecked it twice so its good]

Never blogged about BANG 21 from the Burninators which I test-solved and helped run GC for a couple weekends ago but it was an incredible event. Well-themed puzzles (all to 21 in some way like the Quiz Show, blackjack, the legal age to drink, ...), and for GC/scoring purposes, good use of an Android app to log-in teams and hints. It will set a good standard for what we try to accomplish in BANG 2X whenever we get around to finishing it.

Nikoli has decided Number Link is the next style to go online sometime in July. I like these puzzles, but cannot conceive of them working as well at large size online as opposed to paper as you want to "guess" a ton in these puzzles to intuit how the paths can arrange themselves. I need about 3 shades of pencil lead to be my fastest, but I guess we'll see. I would have still preferred Fillomino or LITS or even Spiral Galaxies but my votes have failed each time.

This weekend is OAPC 5. I learned in Zilina that winning one earns you a t-shirt, so I have 3 OAPC t-shirts here with me now and will go for another this weekend.

The weekend after that I'll be doing Shinteki Decathlon with some friends around Stanford and a couple weeks after that is the USPC. Good times.