Grant's Review Corner: Volume 5

I would like to open this review by confessing a horrible secret. This secret is so horrible that I suspect most of my readers, and an even greater percentage of my real-life friends, will desert me upon learning it. Nonetheless, at the risk of alienation, I believe it is necessary to get this out of the way.

I genuinely like Justin Bieber's song "Baby".

I dislike it when people solicit me to review their things, particularly things which I believe to be crap. Who am I to judge what is crap, after all, when I enjoy listening to the song that, prior to the release of "Friday" by Rebecca Black, had the highest number of "thumbs down" votes of any YouTube video? If there's anything to be gathered from the prevalence of computer-generated Sudoku puzzles, which I believe to be artless crap, in the highest-selling publications, it's that my opinion even on those things which I am most passionate about don't always align with the vox populi. Don't get me wrong; I believe that computer-generated logic puzzles are capable of being legitimately entertaining, which is why I have been on occasion obsessed with getting high scores in Everett Kaser's Sherlock, and why I am a fan of Link-a-Pix by Conceptis, the company with perhaps the most consistently high-quality pictures in their picture-forming logic puzzles. I do, however, find that handmade puzzles are generally more capable of being masterpieces as opposed to entertainment. Additionally, there are times when I think that something is crap (or occasionally the opposite), but am unable to articulate why in any way that resembles a proper critical review.

Thus, I wasn't eager to get e-mail last month:

Grant,

I came across your site a long time ago, but have only recently gotten around to really exploring the world of puzzle websites.

I have invented and patented a new kind of math and logic puzzle called the OCTO.  I have attached some sample puzzles and a detailed set of instructions (not necessary to get started, but they may be helpful if you get stumped).  More OCTO puzzles and an interactive version are available at www.octo-puzzle.com.

I am interested in your opinion of the OCTO puzzle.  Any feedback is welcome...

Thanks.

Sheesh. Ever since Sudoku became the rage, everyone's been trying to create the next puzzle to dethrone Sudoku. In fact, I had first encountered OCTO well over a year ago, and unfortunately, I found it just as bland then as I do now. To the creator's credit, though, the puzzles seem uniquely solvable, which is more than I can say about [name redacted]. As such, I have decided to give the name and URL of the puzzle in case a reader who disagrees with these opinions decides to pay for OCTO puzzles.

I don't feel up for writing a detailed solving analysis for an OCTO puzzle, so I'll cut to the opinions:

1) I feel that OCTO is lacking in elegance. Sudoku was similarly lacking in elegance until Nikoli discovered that you could arrange the givens symmetrically, use fewer of them, and make puzzles of a wider range of difficulties than Dell was offering. This is also, I believe, true of Calcudoku (also known by the trademarked name KEN-KEN); the puzzles generated by computers often lack a certain je ne sais quoi, and I didn't care at all for Calcudoku until I saw that Thomas Snyder was making Calcudoku puzzles with visual themes and solving themes. When I see an OCTO with symmetrical givens in an extremely pleasing pattern, I might take it more seriously as an art form and not just a diversion.

2) One thing about OCTO is that the left number and right number of an octagon cannot be distinguished from each other unless one of them is a given, nor can the top and bottom number of an octagon, the top-left and bottom-right of the top-right or bottom-left octagon, or the top-right and bottom-left of the top-left or bottom-right octagon. This results in 36 pairs of cells in which at least one of them must be a given. I fear that this might restrict the ability to construct elegant puzzles where the givens can be arranged in a wide variety of fashions.

3) The demo at http://hosting.octo-puzzle.com/ could use some serious work. While the solving interface is mostly everything I could ask for in such a puzzle (you can type in candidates), why do no scroll bars appear on the right-hand side, forcing me to view the page in full-screen on a higher resolution to see everything, such as the timer and the "new game" button? Furthermore, why does it take a few seconds to load each puzzle when it's not generating puzzles on the fly? On that note, if the puzzles aren't generated on the fly, why not just let me pick whatever puzzle I want from the certain number of puzzles available? Why not have a "Select puzzle" button instead of a misleading "New Game" button that suggests the puzzles might be made on the fly? Also, even when the option to highlight wrong numbers in red is off, an octagon will only get a green check mark if the numbers are all correct, and not merely a permutation of the correct numbers. I'd prefer the option of a system that highlights conflicts (matching numbers where they shouldn't be, diagonals or diamonds with the wrong sums) rather than correctness.

4) What's up with the intellectual property? Why is there a patent? Perhaps there are valid reasons to have one, but I am a huge proponent of seeing the collective creativity of the puzzle community, not stifling it. Trademarks like OCTO do not stifle creativity; the patent, however, does, because if someone actually manages to do something elegant with the rules that the original creator hadn't thought of, then the patent could lead to lawsuits and stuff. Plus, for some reason, it reminds me of patents on things like the "tap" mechanic in Magic: the Gathering or double-clicking, or the time the Grabarchuks got in a fight with Conceptis over Chain Sudoku / Strimko (which I believe is more visually appealing than OCTO, because it doesn't have so much going on, although as a puzzle it is, as painful as it is to say, less innovative than OCTO). I've seen more patent-worthy innovation in a device that lets you hang bacon on it when microwaving so the fat drains away than in OCTO. The only legitimate reason I see for a patent is to prevent a program that generates millions of these puzzles on the fly from threatening his business. The logo looks bland and generic; I would change the logo to something that uses an OCTO-style octagon as a letter O, because that would make the brand more distinct. Since the author deems it fit to patent OCTO, I deem it fit to demand US$1,000.00 if he incorporates my idea into the logo.

5) Man, I wish I could sell my puzzles in local bookstores.

While I like to believe that my logic puzzle solving and constructing abilities are closer to those of the world champions than the average three-year-old's chess playing abilities are to Kasparov's, I nonetheless feel that no critical review of OCTO (be it positive or negative) would be complete without opinions from said world champions.

Thomas Snyder, a multiple-time World Puzzle Championship champion, apparently got solicited to review OCTO, too, and stated, "I simply ignored the email.  It is a bland and uninteresting puzzle, with inelegantly large summation clues and much less of the feel of rushing to a solution as a sudoku.  And 100% computer generated.  The only thing I can say is this guy went the expensive route to actually patent his idea (http://finance.yahoo.com/news/The-OCTOR-a-New-Math-and-prnews-2757929161.html?x=0&.v=1).  Not sure how I feel about that." As stated above, I'm not sure how I feel about it, either. He went on, "I used to try to respond to these sorts of emails, but then after winning more championships and such I get too many (maybe one every week or two) to bother to respond to.  Most people just getting into puzzles don't recognize the number of types that are already out there, the depth of construction needed to make a puzzle elegant, etc.  They instead think their idea is a beautiful unique thing.  Which it never is.  So I stopped trying to crap on their parade.  I'll let them waste time in an unfriendly marketplace without my advice." As a constructor who has earned Thomas Snyder's respect, I know all too well how unfriendly the marketplace is, and am only motivated to write this Grant's Review Corner by one recent development which I will reveal at the end of the article; aside from that, I plan on following Snyder's lead regarding advice to aspiring authors.

Palmer Mebane, a man who has left a deep impression on me with his seemingly overnight transformation from amateur constructor to holy-crap-this-is-incredible constructor and World Puzzle Championship winner, responded to my thoughts, "Sounds pretty accurate to me. There's basically not much fun to it, for the same reasons that inelegantly made puzzles get boring fast and Sudoku larger than 9 by 9 are a chore. In this case I think even a good human constructor would have trouble packing elegance into this. Your point 2 is a glaring idiocy, and shows the guy behind this type is not a very big puzzle enthusiast." He then suggested potential improvements to OCTO; I refrain from mentioning these ideas here, lest, in his words, I "get a patent claim shoved up [my] ass". (Seeing that Palmer Mebane is more reserved with invective than the Angry Video Game Nerd or The Nostalgia Critic, I think he means business.)

Finally, Thomas Collyer, a Sudoku champion from the UK, wrote, "I’d not come across octocube [sic] before.  The first couple of solves seemed nice enough, I liked how there were different elements of the puzzle to consider.  And then I started noticing that some of the interactions felt more contrived rather than natural.  Perhaps this is a reflection of the guy writing the puzzles rather than octocube itself.  Although I’d agree the whole left/right and top/bottom given thing needs fixing.  Preferably in a way that also didn’t feel contrived. . . . I could see this in current form being published in a newspaper puzzle page a la kenken, but I’m not sure if you could make any more of it.  I think you probably could actually, but I can’t put my finger on what would give it a bit of sparkle." Huh. Apparently people from the United Kingdom are pretty forgiving. (They've certainly forgiven the United States for having declared independence from England, which is now one of our allies in the international community!)

My verdict is that OCTO is a solvable diversion, yet a bland one, in much the same way that computer-generated Calcudoku puzzles are mostly solvable and bland. It tries to bill itself as the next best thing to Sudoku, but fails. Sadly, with any art form where one stands to make lots of money, there will be legitimate artisans, and there will be hacks. For every Hitchcock creating innovative movies, there's a Seltzer and Friedberg appealing to the lowest common denominator, yet earning the ridicule of many others. For every Will Shortz, whose keen eye for what makes a crossword challenging and delightful brings joy to millions of New York Times readers, there is, quite frankly, a Will Shortz, who attaches his name to Sudoku puzzles without symmetry or elegance and, from what I've read (as a snide Nikoli snob, I refuse to try the puzzles on principle), occasional multiple solutions(!).

If OCTO is so bland, though, why have I bothered to "review" it here? Three words: Logic Masters India. LMI is responsible for selecting the people who will represent India in the World Puzzle Championship, and frequently hosts puzzle tests whose main dissimilarity to the ones used to determine the outcome of the WPC is that they don't determine the outcome of the WPC, and you can start the test whenever it is convenient (but after you're given the password to unlock the PDF file, the time limit starts). If you click the link, you will see that LMI is slated to have an OCTO test later this month. In Palmer Mebane's words, "Oh geez." Palmer and I are actually hard at work on creating our own test for LMI; I was already planning to make a kick-ass test, but seeing OCTO on LMI only motivates me further to prove that we two humans are just as capable of our craft as computers are, if not more.

I would also like to issue a challenge to Thomas Snyder: if he enters the OCTO contest and gets one of the three highest scores, I will pay him $18, the retail value of the two OCTO books being offered as a prize for the top three finishers. He can keep the books, or give them to me, or do whatever he wants with them for all I care, but he will get money regardless. Thomas Snyder might easily choose principle over money; I can't say I'd blame him, as I've had to do the same thing when people offer to pay me to sully my pleasantly ad-free blog. However, I would pay money for the privilege of seeing Thomas Snyder kick this test's ass after his comment that OCTO has "much less of the feel of rushing to a solution as a sudoku." (Edit: Thomas Snyder has declined the challenge. Good for him!)