Recap: P&A 69 (aka Damn You, Bender)

(P&A 70 was released today, which means the answers are now online for P&A 69; so here’s my spoiler-laden recap of our experience solving the last issue, themed around Isaac Asimov.)

I’m writing this report a week after the fact, but I jotted down solve times, which should help jog my memory. But before I recap our solve path, let’s have a little chat about QR codes.

QR codes seem to show up more and more frequently in puzzlehunt puzzles, and I haven’t heard anyone else describing them as the bane of their existence. I’ve probably run into half a dozen puzzles where I figured out how the puzzle was supposed to be interpreted as a QR code, but the actual interpretation seemed impossible. The things that drive me crazy are:

1) Unless the black squares are perfectly black and the white squares perfectly white, scanners don’t seem to work.
2) All of the scanning apps I’ve tried don’t indicate any difference between “I don’t see a QR code in front of me” and “This QR code is invalid.”
3) Despite the fact that there are a million apps that scan codes and give output, there seems to be no place online to manually enter a QR code grid and convert it (or explain why it doesn’t convert).

To skip to the end of the report, I figured out what to do with Bender the day before I actually solved it, because I could not get the %$#% QR code to translate. When you know exactly how a puzzle’s supposed to be solved, but you can’t solve it for technical/transcription reasons (and you’re not even sure which)… that’s not much fun. Fortunately the rest of this issue was more enjoyable, we started when the issue was released, at 1pm ET. Jackie solved with me, but joined about 15 to 20 minutes late.

My printer prints documents in reverse order by default (which is sometimes annoying for logic puzzle competitions, in which the light warmup puzzles tend to appear first), so Watson was the first puzzle off the presses. It was a fairly straightforward break-in, and I solved it solo and submitted the answer at 1:08. Checking the stats, I saw that solvers were getting AquaCom quickly, so I broke into that and solved it at 1:25. I translated some of the Call numbers from phonespell without knowing what to do with it, and then worked on Rosie. Once I had the gimmick for that, I was able to extract an answer at 1:41 with only about half the grid filled.

Meanwhile, Jackie had joined and had been working on Th Stpfrd Wvs, occasionally calling out a crossword clue she needed verified. (I’m old enough to have actually read the “Cheers and ____” column in TV Guide.) I came over as she was extracting letters, and was able to wheel-of-fortune out the answer at 1:44. I think at this point we worked on Bubo together for a while and got the clue phrase KINGLY FOURTH; we saw that several of the winged creatures/items had “king” varieties and assumed we had to take fourth letters of these. This didn’t lead anywhere, because it was very very wrong. (We didn’t correct that thinking for a long time.)

After some frustration, we switched to Astro Boy. Jackie had the insight that the names could be spelled by tracing the graph edges, and I suggested copying the same letters to the right side, since the circles were in the same place. Together, we got to the answer at 2:16. Actually, at some point during that time, I think I split off and looked at Tik Tok while Jackie was getting the Astro Boy aha, because I finished Tik Tok four minutes later (at 2:20), and I know the deciphering for that took more than four minutes.

Once I looked at Optimus Prime, it was pretty straightforward and we finished it at 2:31. Meanwhile, Jackie was solving more phone numbers in Call and had the insight that they matched in pairs. I joined in and figured out what to do with the area codes, and we got the answer at 2:39. Then I looked at R2D2 for the first time, which fell by 2:53. Jackie bit the bullet and started chipping away at the cryptograms in Vision. She read me some of the clues, and I not only recognized who they were but was pretty sure they all had eyepatches; since Jackie had actually read the flavortext, she figured out that left vs. right would be important, and we submitted the answer at 3:32.

From roughly 2pm on, we’d also been looking periodically at the meta with no luck, and we weren’t yet sure whether we should be filling things into white boxes or gray boxes or both. Even with ten answers out of twelve we weren’t sure what to do. I figured out that the letters in Bender translated to either BLACK or WHITE, but then struggled with making the resulting QR code scannable. (I colored in the BLACK squares with pencil… not good enough. Colored over it in sharpie… not good enough. Then I manually filled in a grid in Excel… still no, probably because I had a couple of squares wrong.) Meanwhile Jackie had what seemed sure to be a useful insight, as it was possible to fill an answer in the gray boxes reading across one row and an Asimov title (I forget which one) across the corresponding white row so that the letters matched. But after pursuing this further, we couldn’t find any more instances that worked. I went back to QR codes, and Jackie figured out that the first words of our answers (with two missing, of course) seemed to fit into the top boxes, so we should be generating something else for the bottom boxes.

We were already running late for yet another Revolution home game, so Jackie was packing up the car when I yelled (not literally) “Eureka” from downstairs. I had already pictured Silver Surfer in my head when we got WEB SURFER as an answer in a robot-themed issue (I know he’s not a robot, but he looks like one) and I got lucky enough to get the phrases “Silver Surfer” and “platinum blonde” in my head at the same time, allowing us to submit the meta answer at 5:00, make the top ten, and have me be much less cranky at the Revs game than I would have if I were still thinking about the meta. In the car on the way to the game, I was attempting to backsolve the remaining answers, and I realized what the Bubo answer was and why without the puzzle in front of me. But after numerous attempts to force the Bender answer (my most promising option was POTATO BUG), after the game I resigned myself to translating the entire grid into letters and automatically translating it rather than coloring by hand. The resulting grid was only a square or two off from my previous Excel file, but it scanned instantly, and the resulting challenge (resulting in a complete for the issue) was trivial compared to the QR code-grappling that preceded it.

I’m curious to know if I’m the only person crazy enough to be trying to work out the Six Authors meta-answer with five out of six issue answers. I have some incredibly vague theories, and while I don’t think cracking it would allow me to guess the sixth answer in advance, I think it might give me a bit of an advantage on the November issue… [NOTE: I wrote this paragraph in September, and in this month’s issue, Foggy revealed that there is no pre-written meta for the six answers, but he’s inviting readers to create a meta they feed into, Spaghetti-style.]


7 thoughts on “Recap: P&A 69 (aka Damn You, Bender)

  1. Relieved to know I wasn’t the only person to have problems with the QR code. I ended up making a grid in something (Adobe Illustrator?) and copying over all the squares to be solid black or solid white. Then made that into an image and found something online to scan it. Tedious. (And, of course, spent awhile before this trying to get the original puzzle QR code to scan in some manner thinking “I can’t really be expected to copy all of this over… right…?”)


  2. Didn’t do the Asimov P&A, but I’m beginning to hate solver-generated QR codes as a puzzle mechanic for many of the same reasons. WarTron: “Oh, this looks like it’s turning into a QR code! Let’s put it into a spreadsheet and…won’t scan. Export as an image? Won’t scan. Play with the contrast? Won’t scan…” Same thing at Mystery Hunts. Same thing elsewhere. I helped testsolve the Microsoft Puzzlehunt one year, and there was an honest-to-goodness QR code meta. Took us all weekend to get that. Our spreadsheet wizard was massaging the available data forever until he broke through. I mentioned this to one of the editors, and they said, “Well, QR codes are designed to have some error correction, so that shouldn’t have been a problem…” I know QR codes have error correction, and yet it *was* a problem. It seems like a clever way to turn a grid into an answer in theory, but in practice it is annoying and unreliable.


    • It shouldn’t be too hard (if one understands the QR code algorithm) to write a decoder in Excel VBA. (It may even be possible to create one using spreadsheet functions alone.) It would be kind of annoying to actually write the decoder unless one likes doing that sort of thing, but something of this sort absolutely wouldn’t require any fiddly formatting work to finish things off.


      • It’s certainly possible to write a spreadsheet-based QR decoder, and perhaps even a good idea, but for me it really comes down to one question: is decoding a solver-generated QR code fun? Maybe 6 years ago when it was fairly fresh, but not anymore.


      • To me, it’s just bizarre that it doesn’t already exist. Every reader on the market has to scan and then interpret, so the interpretation code has been implemented hundreds of times. Why has no one bothered to let users join in at that stage of the process? The puzzlement reason is obviously way too niche to warrant it, but I could imagine a situation where you want to interpret a QR code and don’t have access to a camera (that’s less and less likely these days, but maybe you’re on your computer and your phone is dead?).


  3. For the QR Code puzzle, I actually started the solve in Excel and only realized it was a QR code when I used conditional formatting to identify patterns. I can see how this type of puzzle would be difficult as a paper/pen solve. But using excel, and resizing the grid cells to be square, resulted in a fairly easy scan for me. Not stating my opinion on this type of puzzle one way or another, but if anyone is tackling this type of puzzle with excel in the future, double check that the width and height in pixels match, that should clear up scanning problems.


  4. It’s good to actually get this feedback on QR Codes. In defense of the QR code, I do feel it’s as valid (or invalid) as Morse code and Braille in puzzles…it better be used sparingly if at all.

    As far as the technical side, prior to publication I emulate what is one possible way to generate the code: copy the text, put in Google Docs or Excel, and format to get black and white squares. I then scan the result with my phone and see if it works.

    I’m not sure when the last QR code puzzle I did was, but I’ll try to avoid them for the next few years.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s