Not Really a Meme

[dr_whom]

I know I don't usually make contentful LJ postings, but it seemed like a good opportunity to make some post-Mystery Hunt comments, both on this Hunt and on my opinions on various aspects of Mystery Hunt theory in general. And really, there's enough of this going around that it practically counts as a meme anyway.

First of all, congratulations to the Bombers overall for an excellent Hunt that really hung together: challenging and interesting puzzles, well-developed theme, lots of innovative concepts. Nice job, thedan, wesleyjenn, and everyone else. I was especially impressed by the thematicity of the whole thing—by which I mean, not the theme itself (though that was just fine too), but the way nearly all the puzzles hung together and related to the theme and especially to the subthemes of the different rounds. All the Round I puzzles tied in to performing arts, Round II to sports, and so on, while still having enough within-round variety in puzzle types and subject matter not to get tiresome (e.g., Round II puzzles weren't all about sports); and at the same time, the Deadly Sin–tagged puzzles were thematic for their respective sins. I'm pretty impressed by this—we made little enough effort last year to tie the individual puzzles in even to the general secret-agent theme, let alone to Boston, Cambridge, Washington, and so on. Because of this this year's Hunt felt more like a unified Hunt and less like just a collection of puzzles.

Next, compliments on the way the sin event puzzles fit into the overall Hunt structure. Last year SPIES learned how easy it is to screw up the relationship between scheduled events and ordinary-type puzzles; congratulations to the Bombers for finding an innovative way of handling this issue.

Also, the dodecahedron metameta was the Best Thing Ever. The only way it could have been better, you know, would have been if it could be completed without the Round VIII meta.... But actually, this relates to one of my hobbyhorses: I think this Hunt had too many potential roadblocks. It couldn't be completed without all ten metas, none of which were backsolvable; and most of the Hell puzzles couldn't be backsolved around either. In my opinion, the ideal hunt should be one in which no one puzzle (except for the metameta or location-of-the-coin puzzle or whatever) is absolutely indispensable to win; it leaves the Hunt as a whole too vulnerable to a single broken puzzle, or can leave an otherwise strong team too vulnerable to a single mental block. The 2003 Matrix Hunt did an admirable job of this, and it's what we aimed for and almost hit last year with SPIES. (The only individual indispensable puzzle apart from the metameta and runaround was the Paris meta; the metameta could have been—with difficulty—solved with any one other meta missing, and the remaining meta backsolved.)

On the other hand—the last three Hunts would all have probably been over by 8pm on Saturday if not for either an unexpectedly obscure meta (Normalville orange star, Hell Round VIII) or a poorly-thought-out round structure (SPIES Buenos Aires). So maybe Noah's right that teams have gotten too good. But does that mean that Hunts will only be long enough from now on if there are frustrating roadblocks, or for the last three years have we just all been overcompensating to avoid another 9am-Monday Time Bandits ending?

I agree with Noah in regarding "pure" metapuzzles—that is, metas where the only data is the puzzle answers themselves, and possibly an ordering on them—as more elegant and appealing than "shell" metapuzzles, where there's additional structure given aside from the puzzle answers. But with four of the last seven Mystery Hunts having had almost exclusively "pure" metas, I can see how it would begin to seem like ground that's been thoroughly picked over—shell metas obviously allow a lot more flexibility in metapuzzle content and style. And I thought the Bombers did a pretty good job at giving their shell metas a consistent structure and style by means of the video gimmick. Also, I'm pretty proud of Metaphysical Plant for the amount of backsolving we managed this Hunt, despite the generally high backsolve-opacity of shell metas relative to pure metas.

One sort of fiddling little criticism I have about this Hunt was that individual puzzle types didn't seem to be spread out very well. What I mean is, for example, the first two crosswords in the Hunt were pretty much back-to-back ("Thinking Outside the Box" and "Pyramid Scheme"), if I'm not mistaken. "Squad Car" came hot on the heels of "Unsound Effects", both cryptograms without word breaks. Late in the Hunt, there were two diagramless crosswords close together. I worked on the first of each of these pairs of puzzles, and then in two out of three cases when the second one showed up pretty much immediately afterward, I was too fatigued by that puzzle class to work on the second one. In the SPIES Hunt we tried to put puzzles of the same general class at least a couple of rounds apart. But this is a minor detail, of course.

Coming up: a few comments on individual puzzles!

Comments

[thedan.livejournal.com]

Thanks for praising the themed rounds... this was actually one of the more important things in the Hunt for me. I feel that if each round has a different "feel" to it, it makes it exciting to reach a new one, which is a nice motivator. I didn't think we were going to be able to make the sin-linked puzzles fit the sins, but it turns out that Mike's a genius.

As for separating similar puzzles, we tried to do so as much as we could, but as you noticed, it worked better in some places than in others. In general I think we did well with keeping pairs of the same puzzle type out of the same round, but proximity beyond that was hit or miss.

[gymble.livejournal.com]

I enjoyed this Hunt. I was a little more involved this year than normal and contributed to solving at least a few puzzles. My one quibble was that there never seemed to be many puzzles in play at any given time, but that might be more due to our team's size and skill than anything to do with the Hunt design.

I am frustrated by the tendency of our team in particular to get hung up on a single meta - there should be a better way around this.

[dr_whom]

Well, sometimes it's our fault (like the Foxtrot Delta Bravo meta on the Monopoly hunt) and sometimes it's the meta's fault (like this one and the orange-star Normalville one). I'm not sure there's much we can do about the latter case.

[novalis.livejournal.com]

I really enjoyed the creative tasks. I created on the Illuminati card, appeared in the movie, and added some bits to the declaration of war. These were fun.

I strongly disliked 9.3, Manipulating The Masses, because it involved unmotivated anagrams.

I liked the DVDs a lot.

[dr_whom]

Unmotivated anagrams? I mean, it was an anagram puzzle; it wouldn't be a puzzle if they told you what order to put the letters in. The anagrams were motivated by the fact that each phrase was an anagram of the name of a chaplain. Unmotivated anagrams are when you extract a series of letters from some data set, and then have no indication of what order to put them in to achieve the final answer.

[novalis.livejournal.com]

I didn't see any clues that the phrases were anagrams -- and even if I had, I don't think I could have un-anagrammed them. That was what bothered me.

[dr_whom]

Anagrams per se weren't clued. It was clued ("masses", "preach", and the symbols) that the phrases clued religious leaders somehow. At that point, it was a bit of an intuitive leap to look up the list of MIT chaplains, but at that point they could be found to be anagrams by inspection.

I'm not by any means saying the puzzle was easy; my team never solved it (though I don't know how far we got on it). But if something needed better cluing, it wasn't "these are anagrams" but rather "look up the list of MIT chaplains".

[(Anonymous)]

The problem with that puzzle, imho, was that the symbols were irrelevant to all but the last step. The fact that it involved anagrams, not so much of a problem. They look like anagrams, and they almost all have V's which tells you that something is up.

--noah

[easwaran.livejournal.com]

I spent a while looking at them, figured they were anagrams, but couldn't get anywhere. I didn't think to anagram proper names together with titles. But perhaps an MIT person would recognize some of those names and thereby get started.

[(Anonymous)]

Regarding the "teams are too good" issue -- I think that the Bombers actually found a brilliant solution to that by letting the hunt keep running until 3 PM Sunday. Having the sin meters and the spiral that showed all the puzzles was good in this regard, because it meant that there were a number of clear goals for teams to achieve other than victory.

I think that this is the solution for the future. Improvements in Google, Wikipedia and database technology all favor solvers. If writing-teams try to write hunts that go until mid-Sunday for the best teams than they will other exhaust themselves or write broken hunts. But I don't see anything wrong with writing a hunt that the best teams complete in 30 or even 24 hours as long as the other teams can still have fun into the middle of Sunday.

DavidS

[cananian.livejournal.com]

It's worth thinking about whether huge teams are a "good thing" or not. The fact that some teams "didn't have enough puzzles to work on" could reasonably be read as "some teams were too large". I, personally, don't like hunts with huge fan-outs, because they reward the bazillion-member teams while penalizing smaller teams.

One obvious way to make the hunt run longer is to reduce the fan out, which might also have a side-effect of reducing team size. However, the large teams would probably bitch and moan mightily (and, for that matter, even the small teams might then find themselves more vulnerable to getting stuck on a single "killer" puzzle).

[(Anonymous)]

Metaphysical Plant had "only" 45 people. Although that's definitely not small, it's also not huge. And cutting it down any smaller would mean that people didn't get to hunt with their friends, and that we wouldn't be large enough to write a good hunt in the event that we won. (Yes, I know some teams can write hunts with fewer people, but we couldn't.)

Furthermore, having hunted with the giant Random team in the past, there's a certain excitement in the large team experience. Hunting with a large team can be a lot of fun, even if you're not that into puzzles, or if you don't know how to solve them. I think that's important, both to make Mystery Hunt a fun event for as many MIT students as possible, and because people get lured in by the general excitement, and then over the years come to love puzzles.

--noah

[cananian.livejournal.com]

I agree for the most part, as I've solved on both large and small teams, and enjoyed both. This year ACME never exceeded 25 people in the room, and had as few as 3 during overnight shifts. Yet we made substantial progress. Setec and the Bombers were both infamous for being "extremely small" and yet very competitive.

Still, I think it's a "theory of the hunt" issue worth considering. It seems the only real limit to team size is social: groups tend to ex/implode once they exceed a certain size. I wonder if there are (or ought to be) hunt-structural factors that moderate team size (to whatever measure of team size you personally feel to be 'optimal'). A "reasonably-sized" team of 45 people is still more people than one (or a small number) of hunt leaders is likely to know personally, making it hard to form teams and/or hard for new teams to be competitive. But perhaps that is a good thing... I'm just suggesting questions, not defending answers (yet, at least!).

[cananian.livejournal.com]

Two years later: "And cutting it down any smaller would mean that people didn't get to hunt with their friends" -- Hm. I already have more friends that I can possibly solve with on a single team. One's set of mystery hunt friends only grows with time, as far as I can tell. Any limit on team size (and there must be *some*, unless someone writes a cooperative hunt we all solve together) will eventually be "too small" from the solve-with-friends angle. I think you pretty much have to make an arbitrary choice on team size, based on the number of people you can comfortably accomodate in your solving room(s), and work from there. You wouldn't want to write a hunt that necessitated more people on a team than MIT's Mystery Hunt room allocation can accomodate -- although, given advance notice, I think it might be interesting to write a hunt optimized for smaller teams (say, "one classroom" teams instead of "two and a half classroom" teams). You'd want to give advance notice so that the larger teams can split up into "Large Team A" and "Large Team B" subteams in advance, and not arrive at the hunt to find that half their members have nothing to do at any given point in time. No doubt social drama would ensue, but it would be an interesting hunt experiment.

[cananian.livejournal.com]

It may seem in retrospect that the Matrix hunt might have done an "admirable job" of preventing single-puzzle roadblocks, but when we ran the hunt we were certainly kicking ourselves because a number of difficult puzzles at the entrance to the "matrix" (R puzzle) maze was starving some teams and preventing progress. Of course, the fan out was high once you could get past the roadblock. In retrospect, we should have either deliberately made the opening puzzles very easy, or provided multiple entrances.

I don't think that "teams have gotten too good" -- the Matrix hunt lasted until Monday not because it was broken (necessarily), but at least partly because there were a large number of difficult unbroken puzzles (although it's likely a few were broken). "Gnireenigne Lab" may be one example: the puzzle was extremely straightforward reverse engineering (you were given everything you could possibly need, manual to every piece, etc), but in retrospect it's crazy how much time it consumes to go through the process. I could/should have given full commented source code and a working example of the mechanism and it would still have been a reasonable puzzle, but with a much smaller required-time-to-complete. ACME test-solved all of its puzzles, but didn't do a good job summing the combined solution times and accounting for the fact that other teams weren't ACME, and thus were likely to have trouble in different places. Calibrating puzzle difficulty is difficult. The Matrix had 104 puzzles, compared to 96 puzzles in this year's Evil hunt, but lasted until Monday morning.

I think the reasons hunts have been shorter is that constructing teams are afraid of repeating the 2003/2004 marathons, and have chosen to err on the side of caution. I have no problem with that!

[easwaran.livejournal.com]

Weren't there 88 regular puzzles, 10 metas, 7 sins, and a metameta? That adds up to 105. But maybe I just counted the regular puzzles wrong.

[cananian.livejournal.com]

I think I was just counting the number of "regular puzzles", and I counted by hand, so I could be slightly off. I didn't count events or metas, since they seemed to be structure-dependent and thus slightly incomparable. Regardless of the exact numbers, my point was that there weren't twice as many puzzles in the Matrix hunt, even though Matrix ran close to twice as long.