2011 Mystery Hunt
Jan. 24th, 2011 05:08 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
tahnanI should probably do that Mystery Hunt recap/reflections thing, not that I think I've done so for several years now, but I might as well get back into the habit. (Fair warning: this won't really be a recap; I'm not going to say "and then I did this! and then that!".)
It's difficult, though, synthesizing everything that happens during the Mystery Hunt. My impressions of, and feelings about, the Hunt have to be reinterpreted after the fact—did I really hate Puzzle X, or was I just too tired to realize that I was just too tired to like anything at that hour? Did I actually like the structure, or did I only like what I thought the structure was, and now that I can look at everything at once, do I realize the structure was different than I thought? That sort of thing. Also, as a team captain—at the very least, nominally, because let's be clear, our team's Mystery Junta does more leading than I do, and really, most of the non-Junta members also do more leading than I do—I feel responsible for the team, which means that I can't be sure how I felt about the Hunt until I'm sure how my team felt about it.
So now that I've had a week to assimilate and catch up on sleep, what do I think? I think that, overall, this was a great hunt, not unflawed but entirely enjoyable, and I really wish I'd gotten more sleep beforehand.
I sent mail to a couple of the team captains to say at least some of the following, but the first thing bears repeating: the infrastructure was fantastic. Metaphysical Plant did a really terrific job with all of the technical details of calling in answers, adding things that seem so obvious to have in retrospect (an answer log, a history of the answers you've tried, typing in answers so you don't have to spell things over the phone, a "contact HQ" link so you don't have to physically dial and redial to get through to HQ for things like "we need you to deliver the objects" or "I think there's an error in here" or the like). The achievements were an inspired addition; their unlocking mechanism seemed well-engineered and, just as importantly, transparent, so that we were always aware of how many more puzzles we'd need to solve to unlock new puzzles, or at worst how long we'd be waiting. And kudos to them for having a replacement puzzle onhand in case something broke, which indeed something did.
I thought the structure of the hunt was excellent; the metas felt well-designed, my favorite being the overall structure of the Megaman round. I also felt, though, that the puzzles didn't always measure up to the standard set by the metas, and this was particularly apparent in the Zelda round. While I appreciate a certain amount of innovation in each puzzle giving three answers, I know that our team ended up more than a little confused—and frustrated—by the way that the puzzles worked so differently, so that doing the work on "Counting the Ways" got you all three answers more or less at once via the same technique, whereas there were three different independent extraction mechanisms in "The Word", and in "Execution Grounds" there were three non-independent extraction mechanisms. As a consequence, it wasn't clear that we weren't getting any of the answers to the latter because we hadn't gotten the first answer (something that wasn't necessary for "The Word").
(I also thought some of the puzzles themselves in that round weren't very good. "The Word" was frustrating insofar as we found three pieces of information, but the second and third weren't the way to get the second and third answers, they were the way to get the second answer; to some extent, that's our fault for mis-solving, but it's something that happens when there are three independent answers in a puzzle. More importantly, the clue phrase "SHIRER BOOK" is incredibly non-specific. "Counting the Ways" required knowledge of a particular album, and if you didn't know that album, it seems to me that there's very little way of getting there. Compare that to, for instance, Unscrambled Cable Porn and Seven Days, where (a) there are clues in the title ("Seven Days" = "One Week") or in the flavortext ("3-2-1 Contact" ≈ "The Bloodhound Gang"), (b) the songs were well-known (both had been on the Billboard 100 chart), and (c) the pictures were readily identifiable, so reasonable guesses as to what they were would be around 2/3 right, and once you had enough of them you could either recognize the song or dump the words into Google. None of that was true of this puzzle, especially because the pictures were so vague—we could, and did, describe the pictures in any number of ways, and I feel like your chances of hitting on a phrasing close enough to get a Magnetic Fields lyric when you Google are incredibly small. Even "Zelda and Zebra" turns up enough hits on a children's book to swamp the few hits for the lyrics. Meanwhile, we were finding all kinds of other things that looked reasonable—"bereft beret", "reaper reader", and so on, more or less endlessly. In fact, we solved it once someone who knew the album happened to look at the puzzle...but not until we'd put in hours of utterly useless work. [EDIT: a little more on the nature of "useless work" in my comment below.])
The saving grace of the round was that we found one of the metas to be incredibly backsolveable—we called in five correct backsolves (and only two or three wrong guesses) in half an hour, once we had the meta-answer. Although even then...well, my feelings about backsolving are documented elsewhere (and again, many thanks to Dr. Whom and his compatriots for making clear up front the value of calling in backsolves, and for tracking backsolves). We found the "Fellowship" meta in Zelda fairly backsolveable, because it relied on wordplay from a pretty limited list, but that wasn't true of the other two, where we didn't really feel like calling in every choral dance, or everything that might be associated with an angel, especially because there was no way of knowing which puzzle those answers belonged to. In a way, then, we were lucky that we'd found four Fellowship answers (giving us five backsolves to call in) and six Holiness answers, rather than the other way around. (Yet another irritating aspect of the Zelda round: once we'd backsolved part of a puzzle, there was a notable disincentive to look at the puzzle again, lest you put in an hour's work only to get out the answer we already had.)
I'm lingering on Zelda. That's because we lingered on Zelda; more than anything, that was where we stalled, after making pretty solid progress on puzzles from the first two rounds, as well as the metas from the first two rounds. Backsolving as much as we did, once we'd finally finished off the Zelda metas, actually did serve to re-energize us, but I'm pretty sure our longest gap without calling in an answer occurred with two of the three Zelda metas finished, and that was before we got into those late-night, no-one's-awake hours. I'd like to spend more time discussing Civ and Katamari, but I felt like I, at least, spent a lot less time with them.
There were certainly puzzles I didn't like, other than the ones listed above. There were also puzzles I liked a lot (and probably puzzles I would have liked, if only I had seen them during the Hunt).
It's difficult, though, synthesizing everything that happens during the Mystery Hunt. My impressions of, and feelings about, the Hunt have to be reinterpreted after the fact—did I really hate Puzzle X, or was I just too tired to realize that I was just too tired to like anything at that hour? Did I actually like the structure, or did I only like what I thought the structure was, and now that I can look at everything at once, do I realize the structure was different than I thought? That sort of thing. Also, as a team captain—at the very least, nominally, because let's be clear, our team's Mystery Junta does more leading than I do, and really, most of the non-Junta members also do more leading than I do—I feel responsible for the team, which means that I can't be sure how I felt about the Hunt until I'm sure how my team felt about it.
So now that I've had a week to assimilate and catch up on sleep, what do I think? I think that, overall, this was a great hunt, not unflawed but entirely enjoyable, and I really wish I'd gotten more sleep beforehand.
I sent mail to a couple of the team captains to say at least some of the following, but the first thing bears repeating: the infrastructure was fantastic. Metaphysical Plant did a really terrific job with all of the technical details of calling in answers, adding things that seem so obvious to have in retrospect (an answer log, a history of the answers you've tried, typing in answers so you don't have to spell things over the phone, a "contact HQ" link so you don't have to physically dial and redial to get through to HQ for things like "we need you to deliver the objects" or "I think there's an error in here" or the like). The achievements were an inspired addition; their unlocking mechanism seemed well-engineered and, just as importantly, transparent, so that we were always aware of how many more puzzles we'd need to solve to unlock new puzzles, or at worst how long we'd be waiting. And kudos to them for having a replacement puzzle onhand in case something broke, which indeed something did.
I thought the structure of the hunt was excellent; the metas felt well-designed, my favorite being the overall structure of the Megaman round. I also felt, though, that the puzzles didn't always measure up to the standard set by the metas, and this was particularly apparent in the Zelda round. While I appreciate a certain amount of innovation in each puzzle giving three answers, I know that our team ended up more than a little confused—and frustrated—by the way that the puzzles worked so differently, so that doing the work on "Counting the Ways" got you all three answers more or less at once via the same technique, whereas there were three different independent extraction mechanisms in "The Word", and in "Execution Grounds" there were three non-independent extraction mechanisms. As a consequence, it wasn't clear that we weren't getting any of the answers to the latter because we hadn't gotten the first answer (something that wasn't necessary for "The Word").
(I also thought some of the puzzles themselves in that round weren't very good. "The Word" was frustrating insofar as we found three pieces of information, but the second and third weren't the way to get the second and third answers, they were the way to get the second answer; to some extent, that's our fault for mis-solving, but it's something that happens when there are three independent answers in a puzzle. More importantly, the clue phrase "SHIRER BOOK" is incredibly non-specific. "Counting the Ways" required knowledge of a particular album, and if you didn't know that album, it seems to me that there's very little way of getting there. Compare that to, for instance, Unscrambled Cable Porn and Seven Days, where (a) there are clues in the title ("Seven Days" = "One Week") or in the flavortext ("3-2-1 Contact" ≈ "The Bloodhound Gang"), (b) the songs were well-known (both had been on the Billboard 100 chart), and (c) the pictures were readily identifiable, so reasonable guesses as to what they were would be around 2/3 right, and once you had enough of them you could either recognize the song or dump the words into Google. None of that was true of this puzzle, especially because the pictures were so vague—we could, and did, describe the pictures in any number of ways, and I feel like your chances of hitting on a phrasing close enough to get a Magnetic Fields lyric when you Google are incredibly small. Even "Zelda and Zebra" turns up enough hits on a children's book to swamp the few hits for the lyrics. Meanwhile, we were finding all kinds of other things that looked reasonable—"bereft beret", "reaper reader", and so on, more or less endlessly. In fact, we solved it once someone who knew the album happened to look at the puzzle...but not until we'd put in hours of utterly useless work. [EDIT: a little more on the nature of "useless work" in my comment below.])
The saving grace of the round was that we found one of the metas to be incredibly backsolveable—we called in five correct backsolves (and only two or three wrong guesses) in half an hour, once we had the meta-answer. Although even then...well, my feelings about backsolving are documented elsewhere (and again, many thanks to Dr. Whom and his compatriots for making clear up front the value of calling in backsolves, and for tracking backsolves). We found the "Fellowship" meta in Zelda fairly backsolveable, because it relied on wordplay from a pretty limited list, but that wasn't true of the other two, where we didn't really feel like calling in every choral dance, or everything that might be associated with an angel, especially because there was no way of knowing which puzzle those answers belonged to. In a way, then, we were lucky that we'd found four Fellowship answers (giving us five backsolves to call in) and six Holiness answers, rather than the other way around. (Yet another irritating aspect of the Zelda round: once we'd backsolved part of a puzzle, there was a notable disincentive to look at the puzzle again, lest you put in an hour's work only to get out the answer we already had.)
I'm lingering on Zelda. That's because we lingered on Zelda; more than anything, that was where we stalled, after making pretty solid progress on puzzles from the first two rounds, as well as the metas from the first two rounds. Backsolving as much as we did, once we'd finally finished off the Zelda metas, actually did serve to re-energize us, but I'm pretty sure our longest gap without calling in an answer occurred with two of the three Zelda metas finished, and that was before we got into those late-night, no-one's-awake hours. I'd like to spend more time discussing Civ and Katamari, but I felt like I, at least, spent a lot less time with them.
There were certainly puzzles I didn't like, other than the ones listed above. There were also puzzles I liked a lot (and probably puzzles I would have liked, if only I had seen them during the Hunt).
- Squared Key: This was a very satisfying way to start the hunt, with a good mix of progress, stalling, and more progress.
- Pipe Dream 2: I didn't seem to mind this as much as people on other teams have been reporting. It was somewhat grueling to push through, but once it got underway it was wholly solveable. (As a quibble, we weren't really getting anywhere until we looked up the answer to the first Pipe Dream, which seems to me to be a little unfair to people who wouldn't know that there was a first Pipe Dream.)
- Everybody's Got To Be Somewhere: I didn't work on this at all; I just wanted to take a moment to observe Dr. Whom's ability to write a musical theatre puzzle (and I had no doubt, looking at it, that it was his) to which I could contribute more or less nothing. Seriously, looking at the answer, I know exactly four of those twenty-eight songs, and two of those I don't even know very well.
- Timbales: hilarious. Fun, and really really funny.
- Expletive Deleted: I felt like this was a good idea that wasn't quite carried out well. "And" and "not" are easy to skim over; while a caret certainly indicates "power", some of the other symbols could be power symbols as well, like the starburst-H and the lightning-bolt-Z; the "smoking" connection is similarly loose; if "shock" indicates a symbol, "stunning" looks like it should too...we ended up spending about twice as long on it as we should have, thanks to some of the weirdness, and the fact that every time I looked at Mork from Ork I had to remind myself that it wasn't the Geico caveman.
- The Writing on the Wall: as with many puzzles, I'm looking forward to seeing the solving stats. The basic insight seems to me to be very hard to have if you don't happen to draw the graph the right way, and while the flavortext specifies that these things are "hallways", it's hard to imagine not drawing the graphs as dots with lines between them, instead of lines with dots between them. Meanwhile, there was so much you could do with the information—given time-travel books, trying to order the hallways by their occurrences in the books, or the actual timelines, seems awfully reasonable. The only other clue to what's going on, besides the graphs, is the way the black and white backgrounds change, but since some change too quickly to track and some hardly change at all and some change, well, arbitrarily enough that our solvers thought it was the order they were clicking. Once again, I'm glad this was fairly easily backsolved.
- Hints, with a bit of love!: also terrific; a lot of fun to put together.
- Painted Potsherds: A few of them seem a little dicey to me, but overall definitely enjoyable.
- Part of Speech: I almost thought this was a little too straightforward, but of course, that's going to be my own bias coming through.
- Laureate: oh why god why did there have to be a British cryptic. I picked it up out of obligation, and pounded on it as long as I could bear, with a lot more wordlist-searching and accepting things without understand the clues than I usually do in a cryptic, before handing it to someone else to finish because I couldn't take any more.
- Plotlines: A lot of fun, even though I didn't really have any useful insights. The worst part is having it to cross it off our team's list of puzzles to write once we win.
- Unnatural Law: the last thing I did before going to bed Sunday morning was to see what was going on—both parts—and hand it off to someone else to deal with. Which felt good, actually; it was nice to more or less end the hunt experience on a positive note (this was after the coin was found, and I woke up about an hour before we struck our headquarters.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
(no subject)
Date: 2011-01-24 10:55 pm (UTC)(no subject)
Date: 2011-01-25 12:24 am (UTC)(no subject)
Date: 2011-01-25 02:22 am (UTC)(no subject)
Date: 2011-01-25 02:43 am (UTC)(no subject)
Date: 2011-01-25 05:06 am (UTC)No, I'm pretty sure people had just heard the album. Wow, I can't believe it came out 12 years ago...
(no subject)
Date: 2011-01-25 01:43 pm (UTC)Something I think, though I don't think I made the point in the main post, is that if a puzzle is going to be unsolvable without specific knowledge, it'd in general be nice if that were apparent up front. If you opened the bridge puzzle, you weren't likely to spend much time on it if you didn't know bridge. Even a list of trivia that turns out to rely on a common theme won't case this problem—you'd either at least get some of the useful legwork done, or decide you weren't going to be able to do anything. The problem with "Counting the Ways" was how much work it was possible to put into it that wouldn't be remotely helpful later, without that being at all apparent. (A perhaps-obvious comparison, too, is to last year's Confirmed Machine 925: if you spent an hour coming up with phrases to describe those pictures but never had the necessary insight, your work would at least be useful to the next person to pick up the puzzle. It seemed like the same would be true for this year's "describe these 26 random pictures"—and speaking of red herrings, 26 pictures, and one has two obvious "Z" words? Sheesh—but I strongly suspect that once Sarah realized they were song lyrics from a specific album, all the previous brainstorming was wholly unhelpful.)
(no subject)
Date: 2011-01-25 03:59 am (UTC)We did catch the "not", but there was a huge red herring on that one: one of the symbols is, well, a knot.
(no subject)
Date: 2011-01-25 04:02 am (UTC)(no subject)
Date: 2011-01-25 04:18 am (UTC)Codex never did get the other two clues, despite even noticing the state/number thing and several Colbertisms. (And I've watched a lot of The Colbert Report, so it wasn't a case of not having the domain knowledge!)
(no subject)
Date: 2011-01-25 04:32 am (UTC)(no subject)
Date: 2011-01-25 09:27 am (UTC)(no subject)
Date: 2011-01-25 09:28 am (UTC)(no subject)
Date: 2011-01-25 10:02 pm (UTC)Interestingly, I think Googling "Zelda" and "zebra" was a lot more effective at getting one to 69 Love Songs two months ago than it is now. I don't know why that would be the case, but I agree with you that when I do that search now I don't find the song until like the third page of hits. But during the writing and testsolving of that puzzle it was a lot more immediate. I do think I disagree with you on the comparison with "Seven Days" and "Unscrambled Cable Porn", though—i.e., I don't see that the flavortext hints on those would be any more helpful than the title "Counting the Ways" was to somebody who didn't already know the relevant songs.
As I think I mentioned to you elsewhere, at least in my opinion the fact that the Zelda puzzles had different relationships between the three answers was a feature, not a bug. I found it kind of disappointing that the Orbital Nexus puzzles from Zyzzlvaria all extracted their four answers in the same way (which was perfectly appropriate for the theme and structure of the round in context, I want to add!), and I enjoyed the fact that we used the Zelda round to explore the different types of ways a single puzzle could produce multiple answers. Maybe we would have helped a bit if we'd front-loaded the type (3) ones a bit more, since those were the subtlest?
In retrospect, we should have put a diagram on Pipe Dream 2 demonstrating the way the pipes worked, as the first Pipe Dream had. I didn't want to do that because I couldn't think of an elegant way to do so without giving away which pipes used the warp zones, but it probably wouldn't have been too hard. And then a testsolver who hadn't seen Pipe Dream 1 before solved the puzzle (although he did infer from the title that Pipe Dream 1 existed and went to look it up), so we decided it wasn't necessary.
Everybody's Got to Be Somewhere: Hee hee! A large part of my time writing that puzzle was spent just looking down the song title lists at castalbumdb.com hunting for placenames. That said, "The Streets of Dublin" and "14 Dwight Ave., Natick, Massachusetts" are amazing songs, and you should definitely give them a listen if you get a chance.
I'm very glad you had both insights on Unnatural Law; I was afraid that in general people would just get stuck on whichever they saw first. That happened a lot in testsolving.
(no subject)
Date: 2011-01-26 01:51 am (UTC)Regardless of which, the main difference is really (c). How much of a clue there was in the title/flavortext is much less of an issue than the fact that it was (IMHO) prohibitively hard to have the right insight at all. The lack of cluing for it was a supplemental problem.
I don't know what it was in Unnatural Law that clued me into both observations. Someone had looked at it and thought it was a very complex logic puzzle, so I know that I skipped to the end to get a sense of the rules without going through the typical logic-puzzle setup, which meant that one of the first things I saw was the forced quartering of soldiers, which immediately suggested "third amendment"; and then while the one above it had a clear fifth-amendment violation, there was just too much text there to clue that, but "he couldn't do it perfectly reliably for any interesting fact about the robot" was a really clear expression of Gödel's Incompleteness Theorem. Or the Halting Problem. Or, well, it turned out to be something I recognized as a mathematical law, even though I'd never heard of it. Actually, I'd never heard of most of them, and I think most of my preliminary notes were wrong (I know I had "ideal gas law" written down, though possibly I did have "Church-Turing law" right), but it was still enough for someone smarter than me to figure out. Anyway, I liked it.
(no subject)
Date: 2011-01-28 05:40 am (UTC)I am simultaneously all giddy over a Magnetic Fields puzzle and very aware that I was the only person on a 40-person team that knew the album. I'm not sure how better to clue it, though. We got very attached to a few wordplay shenanigans ideas early on, which meant we weren't on the track of songs at all.
(no subject)
Date: 2011-01-28 05:59 am (UTC)In a sense: some puzzles take twenty solver-hours because there's nineteen hours of work, plus forty-five minutes of blank staring to get the insight and fifteen minutes of denouement. When the work is filling in a grid or Googling for answers, you can feel good about it even if you aren't the one who gets the final insight. But in the case of this puzzle, the nineteen hours of work didn't really contribute anything (I assume--I mean, I'm guessing you got "I was an army officer and you were a rockette" from the picture, not from previous brainstorming), and I think that's just frustrating for solvers.
(no subject)
Date: 2011-01-29 07:00 am (UTC)*shrug*