Mystery Hunt 2017, Part 4: Team Size and Hunt Length

(Note: The 2017 Mystery Hunt was created by Setec Astronomy, a team of about fifty people. I did a small fraction of the work and was not in charge in any way. Any opinions or perspectives below are my own and don’t necessarily reflect that of my teammates.)

Early Sunday morning, a social media post was forwarded to several members of our team, which declared that Setec had intentionally made the Hunt shorter to “make a point,” and that we should have warned teams that we would be presenting a “mini huntlet” so they could plan for that when making travel/hotel arrangements. Said post is not publicly available, although I later realized I could see it because the author was, at the time, friends with me on Facebook. (Incidentally, if you consider 143 puzzles and 14 metapuzzles, most of which had an average solve time of over an hour, to be a “mini huntlet,” you may want to dial down your expectations, or you’re in for a lot of future disappointment. Just saying.)

I’ve said this elsewhere, and I think people know it by now, but I’ll say it again here: We did not intentionally make the hunt “shorter.” We certainly wanted parts of it (particularly the character rounds) to be accessible to small/casual teams, but nobody on our team, many of whom have experience writing previous Hunts, believed the coin would be found before Saturday evening. In fact, the question of what we would do if a team finished all the metas before the fourth event didn’t even come up until an e-mail thread three and a half days before Hunt, and in that thread, I said, “I’ll note again (as others have) that if no one finishes all the metas and most of the character puzzles by 6:30pm Saturday, and it would be unprecedented if they did (I think? when did Sages finish the metas in ’12?), none of this will be relevant. So in all likelihood the contingency plan won’t be used.” So at the time, we thought a team finding the coin by 6:30pm was unlikely; a victory fourteen hours earlier than that was not even on the radar. (The realization that the top team might finish before THREE of the events didn’t come until Friday night; we discussed contingency plans before I went to sleep, and by the time I woke up they’d been carried out… kudos to the late shift team for managing the situation effectively.)

One of my personal pet peeves is the statement, “The Hunt ended early.” The coin was found early, but the Hunt ended Sunday at 6pm. When I first starting hunting, everything shut down as soon as a team won (I remember we were doing the endgame during the Matrix Hunt, and they just stopped us when the coin was found), but the SPIES folks had the great idea in 2006 of continuing to manage HQ for the rest of the weekend so that other teams could continue to enjoy the experience. The degree to which things were operated have varied from year to year, but in 2014 and 2017 we made it a point to continue to run full endgames for teams that got there, so that if a team hung in and kept solving, they still got the reward they deserved if they finished in time. One of my least favorite Hunt memories was when somebody on MY TEAM (in 2010 when I was solving with the team that would become Alice Shrugged) saw me continuing to solve a puzzle after the winner was announced, and derisively said to me, “Didn’t you hear the Hunt is over?” That year a lot of our team members gave up, but about a dozen of us persevered and finished the Hunt, and we convinced more people to stay in it for the long haul the next year.

As a side note, the aforementioned team definitely knows how I feel about this sort of thing; this year, as We’ve Made A Huge Mistake, their HQ had a room called the Things Dan Katz Hates Room, in which work tables were named after things they were pretty sure that I hate… including “Teams That Stop Hunting When The Coin Is Found.” As a side side note, I played Mystereo during their character endgame and made it a point to add in the line, “And if there’s one thing I hate, it’s special abilities!” To which Erin Rhode, on cue, responded, “Make another table!”

As a Hunt constructor, a philosophy I have repeatedly given, received, and applied is to “write for the middle teams.” We had 90+ teams this year (though a couple dozen dropped out without solving a puzzle) and once you’ve run a Hunt, you realize that you want to try to keep them all happy. Most of the Hunts I’ve co-written for middle teams (the four Setec Hunts, the two Evil Midnight Hunts, and Alice Shrugged) have been well-regarded by most teams, modulo the occasional underclued killer meta that artificially slowed down the top teams. What’s changed over the years is that the spread of team size and ability has become massive. To put it bluntly, when you have superteams of 150-200 power solvers alongside casual teams of a dozen MIT students, there is no way to avoid the former finishing quickly and simultaneously avoid the latter becoming overwhelmed. This year was certainly atypical, but I also think it’s really cool that seventeen teams got to complete the Hunt. On the one hand, I think a lot of us would have preferred for the completion times to be pushed back about twelve hours (I’m sure the top teams feel that way in terms of entertainment time), but if that had happened, a bunch of teams that finished probably wouldn’t have. Where exactly is the “sweet spot”? It’s debatable, and I’m not sure what I think myself.

Some historical context for people relatively new to the Hunt: In the 2000s when I started hunting, there was a bit of an arms race between small “outsider” teams and larger “MIT” teams. After a bunch of Hunts were run by smallish teams like Setec Astronomy and Palindrome, two of the more passionate MIT-based teams (from Third East and Random Hall) became larger than previous Hunt teams had been in order to compete; but when each team won, they split up into smaller teams in the future. But those big teams had a lot of very casual solvers, whereas both of the teams that finished the Hunt super-early this year had 150+ members AND a nontrivial number of power-solvers. In addition, the team that ran the Hunt last year splintered, and there were multiple members of that team on BOTH of these teams.

Having said all that, there have been large skilled teams in the past as well, so why DID the coin get found so early? I think there were three key reasons:

1) Quest metas weren’t as hard as we anticipated.

A lot of metas got solved by top-level teams using fewer than half of the puzzle answers. We didn’t expect this; an intentional design of the structure was that if teams focused on character puzzles, at some point they would get stuck and not be able to open anything else until the first event happened, or they bumped themselves up by solving a quest meta. Since we wanted that first event to be enough of a bump to the character-focused teams, solving your first meta also had to be a significant bump, as well as the second. When the power teams sliced through the first few quest metas like a hot knife through butter, they were off to the races; we wanted teams who weren’t that good at our metas to make steady progress, and that means the teams that were excellent at them swamped the map.

One possible explanation for why the metas ended up easier than expected is that we tested them using Puzzletron (when I wrote in 2014, individual puzzles were tested in Puzzletron but metas were generally tested in group sessions). That means a lot of them got cleanly solved by one person, and when 100+ people are all staring at it at the same time, a meta is going to fall more easily. Again, I think we were more concerned about solvability and elegance than ensuring that metas were sufficiently hard, which has rarely been a problem in the past, and this time, for some teams, it was.

2) Quest rounds had low-hanging fruit.

We had a strict rule that character puzzles had to be fun and accessible (a 1 or 2 on our difficulty scale of 5). While quest puzzles were intended to be 3/4/5, if they tested as 1 or 2, that didn’t bother us too much, because occasionally running into an easier puzzle is nice, and we didn’t have the time to totally revamp many puzzles. But as noted above, some teams could solve quest metas with only a few puzzles, and if those few puzzles weren’t too hard, the round didn’t put up much of a fight. This might have been less of a problem if the unlock requirements for individual puzzles within the quest rounds had been higher (those numbers were chosen by me, so that’s entirely my fault). Ultimately we were more nervous about bottlenecks than overconnectivity; until this year, I’d witnessed lots of team frustration about being stuck and very little about accessing things too quickly. After this year, the latter may be more of a concern for constructing teams.

3) The Cleric meta was super-backsolvable.

I thought The Cleric was going to be our hardest character meta… In practice, most of the top teams cracked it first. (I’d say The Linguist was the character meta that was frequently solved last, though that’s from memory, not from stats, and it might not apply to the top-level teams.) The thing about the Cleric meta is that once you know the answer, you can figure out pretty much all of the feeder answers. So for the teams that solved it with three or four answers, they got six or seven future answers for free. We intended solvers to have to keep moving back between solving quest and character puzzles, and having a bunch of character puzzles that could be trivially backsolved interrupted that flow.

So in the long run, what do we (the Hunt community) do? The simplest solution is for solvers to accept the status quo and act accordingly; if you choose to be on a team that’s really big, there’s a good chance you’ll have a shortened Hunt, and everyone can decide that for themselves. (The social media post I complained about above noted that teams can’t be expected to police their own size… I apologize that the people writing you 100+ puzzles for free don’t have time to manage your team politics.) The main problem with this is the tradition that the winning team writes the next Hunt, which means that to “control” the Hunt, you’d have to opt into the big team “division.” I’m not sure I think the Hunt should exclusively belong to the people willing to be on giant teams.

There’s also the idea of team size limits, which has come up multiple times over the years. I know of two main reasons that it’s usually been dismissed. One is that it’s easy for teams to break that rule, intentionally or unintentionally (as it is, most big teams said they had more members than they estimated when they registered), and another is that the Mystery Hunt originated as an MIT social event, and limiting team size interferes with that. Having been an MIT student, I wanted to be able to tell freshmen they should come by and solve some puzzles, and I would hate for people to have to decide whether a new member should be welcome because they’ll count toward a size limit.

I made a novel suggestion at lunch after the wrap-up, which I’m not even sure if I like, but I’ll say it for brainstorming purposes. In the spirit of events like BAPHL and DASH, you could change the rules from “winner runs the next one” to “whoever’s due runs the next one.” For example, the next Hunt could be written by the earliest-finishing team that doesn’t have any members who have run the Hunt in the last three years. But having the winning team run Hunt is a hallowed tradition, and there’s no way you could get consensus from the Hunt community to permanently change that tradition. Not to mention the question of what happens if a Hunt runs long, and the team that finally finds the coin Monday at 2pm is ineligible to write…

Regardless of whether or how the team size epidemic is dealt with, I want to make one last plea to Death & Mayhem. You’re a giant team. Don’t write a Hunt just for giant teams. There are thousands of people who participate in Hunt and are looking forward to what you’re going to give them, and you’ll have a much more satisfying weekend if you prepare for all of them, not just for the team that’s going to win. In 2012, Manic Sages won as probably the biggest Hunt team in history, and I think they wrote a Hunt designed for a team like them… since that team didn’t exist, the 2013 Hunt set a record for length, and only ended because tons of puzzle answers were given out for free, and the meta requirement was relaxed to five out of six. There are many things in that Hunt that many people liked, but it would be hard to argue that they didn’t overshoot their target (and as a result, few teams saw their super-creative endgame).

D&M has been in contention for a long time, and it’s nice to see you guys finally win, but as the latest mega-team to write the Hunt, just remember that with great power comes great responsibility. 2018 will be the 20th anniversary(!) of my first participation in Mystery Hunt (as a high school junior) and I look forward to celebrating it with you, even if my team’s done in fifteen hours.


52 thoughts on “Mystery Hunt 2017, Part 4: Team Size and Hunt Length

  1. I loved this hunt and thought the Metas were especially strong, probably my favorite ever. The relative brevity was disappointing only because it means less puzzle solving and hanging out with many friends you only get to see once a year, not so much because the coin was actually found. (Oh and coming in second if you’re on Palindrome when you feel you’re so close to winning you can taste it. Grrr.)

    Question: solving puzzles after the coin is found is one thing, but do you go back to solve any of the puzzles you backsolved for an answer?


    • I did some of that this weekend, mostly on my own at home!

      Re backsolvable metas: In addition to Cleric, we found Chemist to be very backsolvable too, though we had more puzzles solved when we solved that one. Our spreadsheet suggests we backsolved half of Economist and Thespians too, though I wasn’t around for those.

      We probably should have listened to SETEC’s repeated pleas that we stop backsolving, as we did miss a bunch of fun puzzles that way. That said, there was a bit of incentive to backsolve character puzzles: since submitting solutions to character puzzles was the only bite-sized way (ie, other than solving metas or waiting for events) to open more quest puzzles, and since solving quest metas was fun, and since solving quest metas mostly required a lot of quest answers, backsolving character puzzles was more helpful for solving quest metas than slowly and enjoyably solving the character puzzles.

      I don’t think it’s a particularly big problem, and probably didn’t affect most teams (since many teams either found the character metas harder, or were trying to win and would be happy with an answer gotten any ethical way). But it’s definitely the case that the experiment of interweaving unlock order between the classes of rounds turned the normal situation of “this solved round has some unsolved puzzles, look at them if they look fun, maybe backsolve them if after hours of trying everybody agrees they are no fun but it doesn’t really matter hugely either way” into “this solved round has a bunch of locked puzzles that we need to ‘solve’ in order to help us with this unsolved round and we already know the answers to them”.


      • For what it’s worth, I don’t think Setec (or any team I’ve run the Hunt with) objects to backsolving versus forward-solving in principle… If you were chided, it’s likely because you were jamming the call queue. Backsolving’s fine, but calling in every word of the form abcbb for every puzzle you left in a round (one team, which may not have been yours, was doing essentially that) is more likely to annoy us.

        Liked by 1 person

  2. I have some more things to say, but I think one really important thing to point out is that although this early coin-finding was in a sense unprecedented, there were other hunts that would have had very early coin finds if they were as clean as this hunt. In particular, I think it makes sense to think of this hunt as Normalville minus the orange star meta (when Random would have finished around lunch on Saturday). Similarly, without the event bottlenecks and some connectivity bottlenecks in the Paris round, Evil Midnight Bombers would have found the coin around lunch on Saturday for SPIES.

    Liked by 1 person

    • No way would we have finished Normalville lunchtime on Saturday without the orange star meta. We ended up finishing it at 6:30 on Sunday. We were hung up on the orange star for a long time, but not for 30 hours.


      • Turns out all the solve stats are posted for that hunt, except that they don’t have the first-tier meta solves. I was remembering the same thing as Dan, that we were stuck on that puzzle for more than 24 hours, but looking back at the data we didn’t get our 5th super-meta (Purple) until Sat 21:34:27. So Aaron’s definitely right and I’m wrong about when we would have finished. I feel like there must have been some reasonable measurement that gave 24+ hours since both Dan and I remember that, but I’m not sure what it would be.


  3. As someone who now plays in one of those middle-ranking teams, I would like to second the argument that if you want to have a “proper” Hunt experience, then don’t be with the big players. If you want to “win”, however, then the big players are the only option. (And that’s perhaps separate from the third option: I want to write Hunt puzzles but I know I’ll probably never get the chance because I’m on a middle-ranking team. But I imagine I’m in a moderately small minority here.)

    The notion that the Hunt ends at 6pm on Sunday rather than whenever the coin is found has been a liberating one. Yes, I realise that it creates a non-trivial amount of extra work for the runners, and I don’t envy them that. From a purely selfish perspective, I appreciated the opportunity to come ever closer to reaching the final runaround, and I bet I’m not the only one. That seventeen teams made it all the way this year is a success, nothing else.

    I too would like to note that I really enjoyed the Hunt this year. The puzzle difficulty scaled well and it was nice that there was some low-hanging fruit in the later rounds to keep us engaged. (On the other hand, I will note that I too solved the Cleric meta with only five answers, which surprised me a little…)


    • Why would it be inappropriate for you to join the writing team one year, even if you didn’t play with them to win? Especially if all you want to do is write puzzles, not take over a leadership position or control the theme or something major.


      • Mystery Hunt culture has always been that in order to write, you have to “earn” it by winning. Not saying it should be that way, but randomly jumping onto a team for writing would be historically unusual. (Mark Halpin did it for Luck in 2016, but he was rejoining a team he had previously solved with.)


  4. Only two things I disagree with. 97 teams registered, 83 teams submitted answers, and 82 teams submitted at least one correct answer. So, only about a dozen teams that registered didn’t actually participate. Also, Dan, you joined Setec Astronomy for the Carmen Sandiego Hunt which was before the early 2000s. You might have even been on my team for Enigmatology…, because wasn’t that the year we played Pyramid together,…. but I don’t have all my 1990s emails anymore 🙂


    • I meant that the early 2000s was when I started to get a sense of the Hunt landscape. I said at the end that next year will be my 20th anniversary, so I’m well aware that I joined Setec in 1998 (Enigmatology).

      Thanks for the correction… I agree that if 15 is a couple dozen, that’s a very small couple.


  5. I’m aware of the perennial discussions about capping team size and the reasons, which you mentioned, on why that will never happen. It’s interesting to me that the Hunt embraces a handful of benign obstacles that are directed toward team size, such as scavenger hunt requirements or the Gluttony event in the Hell Hunt. Do you think those types of size-based specifications are relevant? I imagine size-based specifications could be expanded to more significant elements of Hunt structure, but I expect that would lead to the same negatives that prohibit capping team size.

    I was not part of the contingency plan discussion on teams finishing before events. Was the option raised of having early finishing teams skip the events? I’m guessing that we presented personalized versions of the events as “fun hurdles” that allowed teams to experience the material, and we did not think that participation in the events would somehow affect the ultimate placement of the teams.


    • tmcay, to your last graf, I can say that we did indeed present personalized versions of the events as “fun hurdles,” and that the timing of them was such that final meta solve time was preserved, i.e., DaM finished metas at time X and Palindrome finished at time X + Y, and Palindrome found the coin approximately Y minutes after DaM did. There was a small opportunity for catch-up at the final endgame confrontation with Mysterio and the runaround itself, but those ended up running about the same length for both teams.


  6. I’m conflicted about this Hunt. I say this as one of the weaker (or intermediate?) solvers on a mid-tier team. We finished 15th, and finished the Hunt for the first time ever, which wouldn’t have happened if the Hunt were harder. Whatever other criticism I have, I did like the individual puzzles; one of the ones I worked on was Advisor’s Nightmare, so thanks for writing that. I also appreciated the theme very much, but of course I would, I’m that kind of nerd.

    I did not like the bottlenecks. Several times during Friday night and early morning on Saturday, we had a tiny number of open puzzles to work on. At one point, before the first event, it was down to the Despondent Dynast meta and one character puzzle. This creates 2 separate problems:

    1. If the team is stuck on that puzzle, there’s nothing else to work on. I forget where I read this, but it’s best if, on average, solving 1 puzzle unlocks a bit more than 1 puzzle, because if the unlock rate is 1:1, then any nonzero rate of puzzles the team can’t get will lead the infinite series to converge, and then the team will be completely stuck.

    2. Even a medium-size team becomes ineffective when there are fewer than, say, 4 puzzles available. This is especially bad if the puzzles are not parallelizable. Big crosswords are parallelizable, but the early bottleneck puzzles were not like that at all. The metas aren’t (and shouldn’t be) like that.

    I ended up writing quite a bit that night about public transit, because I had hours of downtime, in which I was ineffective at puzzle solving. There were too many solvers on too few puzzles, and I was not necessarily the strongest. (I’m good at two things: skipping between puzzles doing intermediate steps, and occasionally coming up with an insight of how to solve the problem; I think that this year, the only proper insight I had was that the Woeful Warlord meta had to involve tic-tac-toe.)

    As for the metas, I just think they were too easy. We technically solved the Despondent Dynast meta without a single object puzzle. We just read the flavortext very carefully, realized it was about mahjong and the dynast was down to three players, saw that there were 11 object puzzles that round, and guessed FIND A FOURTH. We finished the Hunt with no idea how to properly solve this, and having seen the solution, I don’t think we would have gotten it with infinite time. The only meta that I think was as hard as I expected a meta to be was the Dismal Dungeon.

    Another, possibly idiosyncratic issue: timing. I’m remote. I keep weird hours, so even though I live in Europe I ended up sleeping and waking up later than the American team members. The difference in hours is what made the fish round from 2015 so nice: there was a large supply of puzzles that the 2-3 awake members of the team can tackle. This time, there wasn’t. I was basically a zombie until I fell down at 11 in the morning on Saturday… and 10 minutes later, another European woke up and asked me whether I wanted to work with them on Intercity Rail. I of course was sound asleep, so they had to do it themselves.

    Fundamentally, easy puzzles like the fish round from 2015 are identical to hard but parallelizable puzzles. The idea is that a small team, or a team that is effectively small because it’s 3 am Donald Trump time and almost everyone’s asleep, can get something productive, it’ll just take longer than a large team. This was done in 2016, but poorly – it was released too late as I recall, and there weren’t enough puzzles.

    The enjoyment order of puzzle hunts I’ve done is still 2015 > 2014 > 2017 > 2012 > 2016 > 2013. 2014 is still the yardstick by which I measure other hunts, with 2015 excelling and the rest falling short in some ways (except 2012, which is hazy by now). No bottlenecks, no special frustration, just slow progress; as far as I remember we got 4 metas by the time HQ closed and finished 13th. Also as far as I remember, that hunt took around 36 hours, same as 2012, even though there were mega-teams, respectively Random Fish and the Sages.

    Finally… what was the purpose of the final items? I liked the item meta, but somehow I expected the metas to interact like in 2015, giving a meta-meta out of the character, item, and quest metas. Or was this part of the final runaround? (I don’t actually know what runarounds are like, since I was remoting this entire time, and our live video link didn’t work.)


    • I’m confused about your characterization of the metas. It sounds like you’re complaining that they were all too easy, but then you said you never would have solved Dynast the way it was intended given infinite time (while multiple teams solved it frontwards with four or five of the eleven answers). The two metas you cited as being most difficult weren’t the last solves for any team, as far as I can tell (Bridge and Minstrels were the ones that caused teams the most trouble), so I think your experience was atypical. As for the bottleneck, part of the structural design was that if you wanted to get ahead of the default release curve, you had to do so by solving metas, and I stand by that as a reasonable requirement.

      I totally disagree that the puzzle unlock rate should always be greater than 1:1; as you yourself noted, if that’s the case, the “series” diverges, and we didn’t write infinitely many puzzles for you! The greater the expansion rate is early on, the more of a crash there is once you’re out of new stuff to unlock, and for me the point in Hunt where you’re not going to get anything new is usually the part where I start getting less motivated. (Also, a big expansion rate creates a puzzle flood that benefits large teams… In this case, the biggest expansion rate came from meta solving, so the flood happened anyway as those teams cracked the metas.)

      The final items (specifically their modifiers) were used in the final endgame (right before the runaround). Eventually we’ll add documentation of that to the website, but for now it’s described in the wrap-up, which is viewable online:

      You will not do yourself any favors on this blog referring to US time as “Donald Trump time.”

      Liked by 1 person

      • The quests metas definitely had a lot more flavortext than I’m used to seeing in metas. In some cases (definitely Dismal Dungeon, maybe Despondent Dynast and Broken Bridge), the flavortext was what got us to jump to the final answer with less data than usual. Maybe the hunt would have been closer to people’s length expectations if it had been scaled back, but I don’t know if Modest Minstrels would’ve been solvable without some cluing about which music theory concept was relevant.

        I will say that as one of my team’s meta people, I enjoyed this hunt’s metas overall, even though the last one we got was Minstrels and nobody in our room had the background knowledge to contribute to it.

        Also, since Maniacal Merchants was the mechanism to make sure teams didn’t skip a bunch of puzzles, but it was structured like a meta, teams that were good at metas could short-circuit it too. (We had ~27 items and had only identified five or six of the eight groupings when we solved it- had almost extrapolated the last couple words of the answer from “TB BA N F…”)

        Liked by 1 person

      • I agree with Greg below about the metas being flavor-heavy. We guessed the Linguist meta entirely out of thin air (well, Ian did really)- we had a bunch of answers but weren’t inspired on the mechanism. Of course, that meant that we couldn’t backsolve anything else in that round.


      • I’m from Alon’s team (hi Alon!).

        I think his experience with the metas may have been colored somewhat by him being remote — Bridge and Minstrels actually were the last two metas we solved, Minstrels being bottlenecked by too few people having the appropriate skills to work on it, and Bridge to us staring at it for something like 6-8 hours and barking up all sorts of wrong trees (we concluded that a disproportionate number of answers had an ABA pattern in them, possibly exactly 8/16 of them, which suggested building some sort of bridge between ones with and without such a pattern…)

        Dungeon did in fact gave us a lot of trouble, but that was more because there just happened to be some puzzles we roadblocked hard on, and the nature of the meta didn’t give us a whole lot to play wheel of fortune with.

        I didn’t touch Dynast so I can’t speak to that one.

        Regarding unlock rate, I agree that >1:1 unlock rate causes problems for top teams (too many puzzles open), but I think it helps a lot with small/medium teams and especially worse-than-average solvers on those teams. Often I find myself looking through all the open puzzles and not seeing anything I feel like I can contribute meaningfully to, even in the early/easy rounds. This is only my 3rd hunt so I’m imagining this problem will get better as the years go by, but having a larger pool of available puzzles does a lot to help weak solvers like me find something to contribute on.


      • Okay, let me clarify.

        1. I think Dismal Dungeon was the meta we spent the most time on, but wasn’t the last to fall. This in turn is because of object puzzle difficulty – the object puzzles in that round, while on the harder side, were still less hard than in the rounds we finished last, i.e. Modest Minstrels and Broken Bridge. Curious Cube took a while, too, for the same reason (P.S. I love that meta – as I keep saying, I like the individual puzzles almost as much as I did in 2015, I’m just objecting to the unlock structure).

        2. What I’m saying re Despondent Dynast is that a) solving it properly looks like it would’ve been really hard for us, and yet b) we solved it pretty early because of flavortext. Sometimes solutions to puzzles are underclued, but in this case it was probably overclued.

        3. It’s reasonable to require teams to solve metas… but the problem is that the level requirements were such that we had to solve all metas at some points. It becomes a form of bottleneck, where everything hangs on one puzzle. It’s different if the meta is highly parallelizable (i.e. it has tons of puzzles in the round, and all or almost all are required, as in Dismal Dungeon, even if you know what you need to do), but I don’t think the Despondent Dynast was.

        4, I’m not actually complaining about the difficulty level of the Dismal Dungeon. I think it was hard, not too hard.

        5. Speaking of parallelization and metas: is it common for hunt writers to test-solve metas with different numbers of correct answers? In other words, to give the test solvers half the answers and see if they get it, and if they don’t then give them a couple more, etc.?

        6. While I think the expansion rate should be more than 1:1, I don’t think it should be massively bigger, precisely to prevent a flood. Probably the number of solvable but unsolved puzzles should remain constant, so for a given stuck puzzle rate a, the expansion rate should be 1:(1-a). Not sure what a reasonable a is, but for us, something like 0.2-0.25 wouldn’t be too far off, for an unlock rate of around 1.3:1. For stronger teams it creates a flood, but ultimately, the stronger teams are going to prevail anyway; none of us on the Transplantations has any illusions of ever winning. The issue of flood vs. bottleneck is more about teams of a zillion grinders versus smaller powersolver teams like Setec. (Put another way, the Sages wrote a hunt that only the Sages could have solved, and Setec wrote a hunt that’s optimized for Setec’s power-to-size ratio.)


      • On a tangent, I feel like this is a good place to mention that I do know of one zero-answer meta solve: in 2006, Death from Above correctly guessed CHOCOLATE MOOSE for the round 7 meta just on the basis of the fact that the spy was the Swedish Chef and the round had 14 puzzles.

        Liked by 1 person

      • Manic Sages got a sub-meta in 2011 with no answers. 1-1 was a seven-letter mushroom starting with P, 1-2 was a seven-letter flower starting with P, so 1-3 must be a seven-letter star starting with P and the first one we thought of was Polaris.

        Liked by 1 person

      • This is ancient history by now, but you ought to know that Alon Levy is a remote solver on my team, and he’s just 100% dead wrong about how our team got Despondent Dynast. We solved it completely forwards, the intended way. I know this because I was on-site and was the one who personally figured out how it worked and wrote down the answer letters (shortly after a teammate suggested Mahjong).

        Just want you to know that particular piece of negative feedback on your last hunt was completely unwarranted and based on a total misunderstanding of what actually happened. That meta was great, make more like it next year. -Kylee


    • The final items all had modifiers which you used to shift the letters in HIVE MIND to form GROUP HUG, the final step to defeating Mystereo in the endgame. The choice of modifier to use on each letter was given by numbers each character was standing on. To summarize, the endgame went like this:

      * 6 players stood on 6 hexes, surrounding Mystereo, who stood on another hex. (Everyone else on the team stood around the room to watch and help.)

      * In whatever order, the players attempted to use their powers on Mystereo, and in response, he asked them a question like “Who is my favorite magician?” The players answered using the appropriate answer pattern from the earlier Battle. So for this question, the answer was “Houdini.”

      * If correct, the player flipped over his or her hex.

      * When all six hexes were flipped, the players realized they were labeled with the numbers 1-8. (Obviously two of the hexes had two numbers.) These corresponded to the letters in HIVE MIND each player was responsible for. The players all shifted their letter(s) by the appropriate amount (i.e. according to the item modifiers), producing GROUP HUG. Then they gave Mystereo a group hug.

      * Mystereo’s hex was flipped to reveal the runaround directions. The runaround was pretty easy and led straight to the coin.

      I hope that makes sense. Actually, I hope they post detailed info on the non-puzzle bits (events, Battle, endgame) so that they can be well-understood and enjoyed (via pics or video) by everyone. Constructing teams often forget these things in their post-Hunt exhaustion, and they never get documented.


      • Documentation of those is being actively discussed. (I can’t promise for sure that discussion will lead to it getting done, but I agree that the longer it takes, the less likely it will be done at all, and I will prod accordingly. But keep in mind that everybody’s Real Tired right now.)


    • @alon: re the dynast meta: we certainly knew how to solve that. I can’t really speak as to what we had going on in the 2 hours or so before solved it (when we may not have had a ton of puzzles), because I was eating and dealing with a baby. I got back to puzzles basically right as we solved it (with 7/11 puzzles), and saw it get solved.


  7. I’ve been telling everyone who will listen that I thought this hunt had an amount of quality content nobody could reasonably call disappointing. And I think it’s really good — important, even — for 15+ teams to find the coin, at least once every few years. Wanting to experience the whole hunt, and not wanting to be on a super-huge team (or not wanting / being qualified to be on a super-strong team) should not be contradictory. So I hope this happens again, if not perhaps every year. If anything I doubt most teams could make a hunt with this much material that finishes so cleanly for so many teams. It’s definitely harder than it looks.

    …I won’t lie, though; there was some happiness in Codex HQ that we no longer hold this particular (modern) record. 😉 To answer your question, Sages solved the last meta at 5:19 PM and then got bogged down in our (long and imperfectly tested) endgame. So there’s precedent for “before dinner”, and anyway, as Noah says, without hard stop metas certain other hunts would have wrapped up their main phases much earlier.


  8. From what I experienced his year as part of the Luck contingent that solved with Desth and Mayhem this year, they are enthusiastic to make their mark, they have been receptive to war stories from those of us who have done this before (which, unsurprisingly, aligns with much of your thoughts in this blog), and they have remarkably strong leadership. I know it makes one a bit nervous when an unblooded team is up for Hunt, but I am confident this team is up to the challenge.


  9. So I think this Hunt was absolutely brilliant in theme, execution, puzzle construction, and all the other things Setec is so good at. It definitely lived up to expectations. OK, I miss getting a T-shirt (which seemed to be becoming a staple), and I wish I had a souvenir of finishing the Hunt, like the mini lodestone from 2015 that I have on my fridge. But the only real problem was the finish time, I think. But I’m not sure why I think that.

    Part of it is because I go way back to 1999, when the Hunt ended when the coin was found. And so there is still part of me that feels a bit deflated when that happens. Don’t get me wrong—I love the 6:00 Sunday endings, and after failing to take advantage of this extra time in 2007, we’ve done so every time it was possible since. And we finished for the first time in 2015, then again this year. So I’m not arguing to take that away. It’s just that there is something very epic about the crazy MIT Mystery Hunt that goes for an entire weekend, sometimes record length like in 2013. It brings a certain status to the event. And there’s something kind of awesome about running around campus on Friday and Saturday or maybe even Sunday night, knowing that no one has found the coin, that a huge race is on. (I still remember turning in a scavenger hunt late Sunday in 1999 and hearing from Acme about how some team—Setec!—was almost done but that Acme “still had a few hoops for them to jump through.” To my newbie ears, that sounded so epic after 50+ hours of Hunting!) I know we’re never close to winning, no matter how much I want to be, but I can pretend for a little while. When the coin is found, it stops being a competition and loses a little mystique. And this one was so insanely early that it even discouraged my team for a little bit, despite us being used to the new status quo. You’d think we’d be happy—it meant we had good odds to finish. But really it meant that all of our seemingly great progress was nothing compared to the leaders’. Getting to the Battle right then was probably the best thing to happen to us, because it lifted our spirits.

    That said, I don’t think anything can be done about this. The puzzles were maybe a little easy, but man, if they’re all hard, it gets old fast. (And…they’re still kind of hard.) The metas were easy to get with few answers and led to many backsolves, which is kind of fun but also kind of annoying sometimes because our best solvers wipe out potentially fun puzzles when they get the meta/backsolve. But sometimes the coolest meta just turns out like that, right? And the teams are huge, which I can’t complain about because my team is getting pretty big, even bigger than I thought once I looked at our login stats. (But we have a larger percentage of casual solvers, I’d guess, so we’re not Death and Mayhem or Palindrome.) In the end, I guess we all just have to acknowledge that Hunt is not really a fair competition, whatever that means, and so if someone wants to win badly enough to make an uberteam, they will simply destroy Hunts that respect Setec’s (laudable) design philosophy.


  10. One of the many problems with 2013 is that we on Sages were generally thinking of ourselves as a team of average-level-and-below solvers that won by being able to throw more manpower at everything. Therefore, if a puzzle was slightly too hard for our testsolvers, it’s just the right level for Mystery Hunt experts. In other words, we didn’t realize how good we were at solving, which made us that much worse at writing.

    One of my current teammates jokingly speculated that Setec had made the same error in the opposite direction.


  11. Hmm. Yes, the top two finishers were giant (and stacked) teams. And team size is important to discuss. But it seems to be a distraction from discussion of why this Hunt ran short. (I know you prefer to say “why the Coin was found early”, but for me and hundreds of other Hunt fans, the Hunt really did end early on Saturday.)

    You really should have mentioned that Left Out was a regular-sized team of great solvers, much like Setec Astronomy or Luck in previous years, and they still finished at 9:00 am Saturday. That means their Hunt lasted 21 hours, more than a third shorter than the previous record of 34-35 hours. Furthermore, at least 6 other teams finished the Hunt before that 34-hour mark. If, at noon Friday, D&M and Palindrome had collapsed into a black hole of writhing wriggling solver mass and fallen through to the center of the Earth, we would STILL be having this “why was the Hunt so historically short?” discussion. (Well, maybe not me, from beyond the event horizon.)

    I’d say the biggest factor was really just Setec’s preference for puzzles that are clever and elegant but not too hard or grindy. As someone pointed out on Facebook, the Mystery Hunt typically has plenty of puzzles scaled to stump (or at least delay) both the large teams and the experienced solvers. Really hard ahas with no help from the flavortext, steps requiring a large amount of esoteric research, data that looks like nonsense unless you happen to have read every issue of Shonen Jump, etc. From your post it doesn’t sound like that’s the audience you wanted to write for. You wanted the teams in the middle of the pack – people who don’t live and breathe puzzle hunts – to be able to solve the Hunt, and that’s what you got. But that’s incompatible with keeping the top teams busy all weekend.

    On top of that, Setec purposefully wrote almost half the Hunt specifically as “fish” puzzles. It’s a clever idea to have a part of the Hunt designed for more casual teams, and I think you pioneered it with your own Hell Hunt’s opening round. In recent years, 2013 had an easy opening round (7 puzzles), 2015 had the actual Fish round (57 puzzles), 2016 had an easy opening round (25 puzzles) and the Dreamtime round (25 puzzles). But these were thought of as bonus content for a wider audience; we knew the top teams would power through them quickly. Treating your 67 Character puzzles across 6 rounds as a regular portion of the Hunt really threw off your time/difficulty estimates. What’s worse, they were actually the puzzles that gated all the unlocks! I know you didn’t intend to write an easy Hunt, but you clearly did intend to write an easy HALF-Hunt. Of course there were consequences.

    Oh, and also you failed to write a frustratingly broken metapuzzle that was so hard nobody could solve it for half a day. That’s a classic Hunt tradition, and it would have fixed the length issue. 😉

    I don’t mean any of my criticism to be particularly negative. I know all too well how easy it is for the Internet to gang up and tear apart the thing you’ve poured your heart and soul into for a year. It was a fantastically well-written and well-executed Hunt, and for many teams I imagine it was one of the best Hunts they’ve ever done. But personally, I hope future Hunts return to being long and brutal, the kind that really do require 50-100 people to work hard together for 50+ hours, stealing sleep where they can. For those who prefer more scaled-back puzzles, or the “fish” variety, there are LOTS of other puzzle events through the year to satisfy that urge. There has only ever been one event at the level of the Mystery Hunt, and that’s the Mystery Hunt. 🙂


    • I don’t remember whether this was something Chris said at wrapup, or at some point when he stopped by our HQ after we finished, but it sticks with me:

      Chris was staffing a booth for the Mystery Hunt at some activity fair at MIT, and he said that a lot of students came up to him and said, “Oh, 2015 was my favorite. There were puzzles I could do, and I didn’t feel stupid.”

      Now, I’ve been solving puzzles for a long time, and I have enough confidence in my abilities to not feel stupid when I can’t solve a puzzle. But I’m not the core audience of the Hunt. Those students saying they liked 2015 because they didn’t feel stupid? They are the core audience.

      So I hope that this year, like 2014 and 2015 before it, was not an anomaly.

      Liked by 6 people

    • “I’d say the biggest factor was really just Setec’s preference for puzzles that are clever and elegant but not too hard or grindy.”

      Given the choice of a shortish hunt that is full of clever and elegant puzzles, or a full weekend spent with puzzles that involve tracking down the minute details of strings of seven coinflips in a list of a million flips, or having to solve a fractal word search down 87 levels, I’ll go with the short, fun hunt every time.

      Liked by 1 person

      • No need to be mean.
        There is a space in between a “fish” puzzle and one that is basically impossible, and I too look forward to trips to the MIT Mystery Hunt as one of the few habitats of the long-form puzzle.


      • I agree with Melinda that Lance’s comment was a bit harsh. There is middle ground.

        Having said that, I think there are two types of “big grind” puzzles. There are puzzles that are long due to their revolving around a big epic idea (like many of Derek’s puzzles), and there are ones that are long due to size (identify 200 sound clips!). Both types have their fans, as is evident in the comments here and on some online posts about previous Hunts. For whatever reason, I don’t think you’ll find a lot of fans of either on Setec, which is why none of us were particularly excited about writing them, although I’d argue that Hexed Adventure, Marvels of the Ancient World, and Mirrorball all had a certain amount of “bigness” to them. For my part, I’d take the epic idea puzzle over the slog-for-the-sake-of-slog puzzles any day.

        Also, I don’t think it’s fair to conflate Fifty-Fifty and In The Details (the fractal word search). I’m generally not a fan of Derek’s epic puzzles (sorry, Derek), but I actually really liked In The Details and thought the process was reasonably clued. Whereas Fifty-Fifty has become a go-to punchline in Hunt discussions (see Advisor’s Nightmare for one more jab). The amount of guidance in these puzzles versus expecting the solver to guess what the author is thinking can make a huge difference… It’s becoming evident that some solvers actually enjoy the experience of being adrift for hours until they finally figure out what to do, but I don’t.

        As for team size being a distraction from discussing why the coin was found so early, given that I proposed three non-team-size-related factors in my original post, I’d like to think I’m encouraging people to think about other reasons as well. (Though I admit that a lack of big giant puzzles was a factor I considered listing but didn’t… If anything, it would have made #2 even more pronounced.)


      • I do apologize for being mean; and I mostly regret it as well. My ad hominem tone was sparked, really, by what I felt was an unnecessarily adversarial tone from Derek, such as insisting on referring to the Hunt as “ending early”. I also think that he not only missed an important point that Setec made, but in fact willfully missed it: that the Hunt’s audience is MIT students, and that therefore “keeping the top teams busy all weekend”–especially when that involves brutal difficulty–should be the goal, with easy puzzles scattered here and there.

        Underscoring Derek’s deliberate ignoring of that point is his final statement, that “For those who prefer more scaled-back puzzles, or the ‘fish’ variety, there are LOTS of other puzzle events through the year to satisfy that urge.” Setec’s point, and I think it’s a really important one, is that it’s a mistake to think of the audience of the MIT Mystery Hunt as “puzzle solvers”, who can go to any number of other puzzle-solving events. The audience is “MIT students”, and Derek’s final statement would only really apply if there were lots of other *MIT events* that featured easier and more tractable puzzles. There aren’t. This may be one of any number of puzzle events, but it’s the only MIT puzzle event, so for the actual audience of the MIT Mystery Hunt, there aren’t lots of other similar events through the year. (Though I’d love to see interested students come to DASH and BAPHL!)

        It was the deliberate dismissal of that point that sparked my ad hominem attack. So I regret and withdraw the ad hominem content, but stand by a rephrased point: “Given the choice of a shortish hunt that is full of clever and elegant puzzles, or a full weekend spent with puzzles of brutal difficulty, I’ll go with the short, fun hunt every time.”

        Liked by 2 people

    • “Research something really obscure” is a time-honored Mystery Hunt tradition, dating back to Brad Schaefer’s puzzle in Linear B. I feel like there was relatively little of that this year, but I suppose obscurity is in the eye of the beholder.

      (I wrote the anime puzzle in 2013 and the comments on it I saw were split about evenly between “how the hell were we supposed to know anything about that?” and “yay anime! some of my favorite shows were in it!” My view is if there’s an MIT student group devoted to the subject it can’t be THAT obscure.)


  12. “One of my personal pet peeves is the statement, ‘The Hunt ended early.’ The coin was found early, but the Hunt ended Sunday at 6pm.”

    Yes, this, so very much this. I’ve had to correct a couple of people on this point, in fact. All the more so because this was an important design decision, and one I really appreciate y’all deciding.


  13. Now for my way too long answer.

    Estimating the length/difficulty of a mystery hunt is extremely difficult, and I think it’s impossible to hope for better accuracy for a given team’s finish than a margin of 6 hours either way. It’s still pretty hard even if you give yourself 12 hour error bars. So even though I think this hunt was around 12 hours too easy (as I’ll explain below), I don’t blame the writing team for this. I don’t think you could have known in advance it would play out this way.

    If I were to pick a goal for how long/difficult a hunt would be I think I’d pick “IIF finishes the hunt early Sunday afternoon.” IIF is a good strong team that has been hunting for a long time and doesn’t have wild swings in terms of its composition, so that’s a reasonably stable goal and it also just seems pretty reasonable that a team that good should be able to finish. That’d give 14 or 15 teams finishing this year (depending on sleep), and the extra 12 hours only buys you two or three more (which is great for them to happen some years, but doesn’t need to happen every year). I say early Sunday afternoon instead of Sunday dinner, so that if you miss by 6 hours they’ll still be very close, but I think IIF at 5pm Sunday also has a strong argument. By this measurement the hunt was around 12 hours too short. Looking at some stronger teams that still seems reasonable, we had a great time at Plant but finishing the last puzzle around 6pm on Saturday feels a little fast.

    That said, a goal like “IIF finishes Sunday afternoon” means there is a genuine problem for the 2nd and 3rd place teams who don’t get enough puzzle time (and also don’t get to win). I don’t think the solution is ever longer hunts, but I do think it’s a real problem that we should think about solving. Keeping HQ open solved the “coin is found early” problem for most teams, but for the top teams what we need to do is to keep their teams solving puzzles after they found the coin. No one person or small group sees anywhere close to all the puzzles in a weekend, much less in 18 hours, so this should be a solvable problem. One specific idea is to have a website available where they can see all the puzzles but none of their answers and let them continue to call in answers. A friend on one of the top teams found a small newbie team of friends to join on Saturday morning and then got a whole weekend of still solving and had a great time. There should be a way to get more teams this experience. Another idea would be to let teams that have finished the hunt see how often a puzzle was solved, so that they could concentrate on the fun small-group puzzles and not the big stumpers. A third idea would be to have set up some kind of achievement-based rounds that small subgroups of the team could tackle, like “solve all the crossword puzzles” or “solve all the puzzles whose answer starts with an A.”

    In terms of why it ran short I think you’ve hit on several of the big reasons in your answer. Here are a few more. I think the answers and cluephrases were a bit too cute and nice and made it too easy to solve metapuzzles from very few answers and backsolve puzzles. Only for “The Fighter” and “The Dismal Dungeon” did we figure out how a metapuzzle worked and still need more puzzle answers to extract (and for DD that’s only because you released the grid piecemeal). A really extreme example of this phenomenon is that we solved The Maniacal Merchants with 18 letters out of 40 (SELLT…..E..SSJ.NK..DGOO……….SPR..) It would have been better at that point for us to be forced to solve a few more of your wonderful clean puzzles to move on instead of getting by with only solving the easy ones and backsolving. It’s fun when an answer gives you that spark of themey-ness, but I’m not sure it’s worth the cost in length. I think the quest metas weren’t so much too easy, as they were overclued. They were great beautiful clean puzzles, and mystery hunt teams don’t need big hints to solve them. “The Economist” was way way less of a hint towards state coins than “mastermind” is to Mastermind. I loved the metas in this hunt, and don’t think that replacing them with harder metas is an improvement, but having less strong cluing I think would be an improvement.

    Finally, I wanted to say one thing about why I thought this hunt was too easy in a way that’s unrelated to length. One thing I like to get out of a mystery hunt, in addition to a couple days of solving fun puzzles with friends, is the feeling that at some point our team overcame a great challenge. I don’t want that feeling constantly, but I do want it once or twice. This time everything fell so quickly I just didn’t feel the same sense of accomplishment. For comparison, in the 20K Leagues hunt (which also had a low average difficulty) I felt like our team was really challenged by the castle metapuzzle. I think this hunt would have been improved by having one more difficult obstacle, for example a meta-meta for the last few quest puzzles, or an inventory puzzle that felt more like the Hell meta-meta and that was challenging and forced us to solve more quest puzzles. That probably would only have added an hour or two, but it would have made me feel better in terms of having met a big challenge.

    Liked by 2 people

    • I’m worried from some of the later comments that I maybe didn’t strike the right tone here. I really really liked this hunt and had a great time! I thought it was a great hunt. My comments are all really small tweaking suggestions that come from liking to argue about hunt writing minutia, not serious criticisms. The most important paragraph here is the one about trying to find the “early coin solve” problem for the teams near the lead, and that’s not a criticism since no one else has solved that problem either it’s just a “important problem to solve in the future.”


  14. @ Alon Levy (from January 22): <>

    I can’t speak for Setec, but I’ll assume this was done. We (Team Luck) definitely were using Puzzletron to do just that for the metas on Huntception. A solver could choose to open up any/all of the round’s answers before attempting to do the puzzle. I set up a randomizer so as to pick answers ‘out of a hat’, starting with about half of the solutions and then revealing another one every time I felt I still wasn’t making progress.

    From that, we knew that hunt had metas that generally required most of the answers in order to solve the puzzle. (I felt they were fair, if generally on the difficult side.)

    One thing I think is harder for play-testing teams to do is to adequately replicate how puzzles will be solved; for metas, we didn’t really know what meta answers we’d be more likely to see than others (though afterwords we were definitely able to provide feedback on which answers were seen as ‘critical’; this got logged back to make sure the puzzles they were in were appropriate). Also more often playtesting is done by individuals while rested at a computer, rather than groups of people talking with eachother on sometimes little sleep.

    Not a perfect process, but we all do the best we can.


    • I can confirm that Setec tested metapuzzles using an interface that started with just a few (pseudo-randomly chosen) answers revealed, allowing the solvers to gradually (and pseudo-randomly) reveal more answers over time, as they felt they needed them. is the code we used. Solvers reported the progress they made, along with the set of answers they had as they made that progress, in their Puzzletron testing feedback.


  15. Apologies – it cut it off, but I was responding to Alon’s question about parallelization and metas/different number of correct answers above.


  16. Hi, I’m with Setec, and I wrote a few puzzles. I was also far, far removed from action this time around, so please don’t imagine that I represent Setec philosophy here.

    There’s one important subject that I’m seeing above in a few comments (Derek’s, Noah’s, others’) that’s new this year:

    Some people are sad at the lack of grindier, sloggier puzzles!

    This may be the first year where people could lodge a complaint like that. It’s worthwhile for future teams to note.

    It’s also totally surprising to me.

    The only comments I remember about sloggy grinds were complaints. “That was impossible!” “I spent four hours on it, then got pulled off because it was already backsolved.” “I spent five hours on it, then got told it was backsolved an hour earlier.” “I spent six hours on it, and it was okay I guess, but I missed so many other puzzles and events and meals and fun.”

    I do not remember anyone responding with “No you guys, that grindy slog was awesome!”

    I do remember people feeling so negatively and strongly about these grinds that they bring them up years or omg-decades later. (And the ones I remember are not just the ones that people personally tease me about.)

    Dan says he remembers some positive comments, but, eh, not me.

    So let me ask a question. In the comments above, the people who miss the grindy slogs seem to say that they (the puzzles) have the following benefits: they bring cachet to the Hunt, they lengthen the time it takes to find the coin, they ensure “the right” teams find the coin in “the right” order, and they help you feel like your team did something big. I hadn’t thought of all these reasons before, so thanks. But I did not see anyone claim that the slogs are fun to solve! My question is: Do the proponents of sloggy grinds actually miss working on them personally, or just miss knowing that they’re out there?

    Liked by 3 people

    • I don’t know if I would necessarily call them slogs, but I have certainly sunk lots of time into a single puzzle and had lots of fun doing it. World’s Tallest Cryptic (2008) took me at least 6 hours, and I was elated when I pulled an answer out of it.

      My friend Kyle says he likes when a puzzle has a lot of research as the first step, because it’s something he can do when he’s not at his peak and there’s nothing available in his wheelhouse.

      However, I am entirely with you on the sentiment. I think it’s possible for a 4-hour puzzle to be lots of fun, but if it is, then it’s not a slog. Grindy and fun may be possible, but it’s hard to get right.


    • Maybe we should have a virtual puzzle called “The Grindy Slog” that simply gives you an answer after 12 hours and doesn’t include any actual content. 🙂


      • I definitely don’t want slogs. What I wanted was some variation in the difficulty of the metapuzzles so that the team would get the buzz you feel when you just did something you didn’t know you could do. I really really liked the metas this year, but I also felt like we just cut through them like butter. Now there’s a very fine balance here, I’d rather have the experience of quickly solving them than what happened in Normalville (I felt really happy when the final meta at wrapup said it was written by you and we hadn’t gotten stuck on it at all). The examples I had in mind (the 20K meta-meta and Hell meta-meta) weren’t slogs. Other examples would be the 2nd layer metas in Normalville, the world meta-metas for video games, resolving the first round for History (though not the related scav hunt which was grindy). Something that was a bit of a “showpiece.”


      • They did that joke last year: Time Suck was just a 10-hour video of an annoying song. (although you could use tools to find the important points in the video without actually watching the whole thing)

        I wouldn’t call myself a proponent of the slogs, but I guess I’d be in the “cachet” category: most of what I get out of the huge puzzles is stories to tell my friends how ridiculous Mystery Hunt is. (I’m in Seattle, so I like to joke that the Microsoft Puzzle Hunt is scaled for teams with 12 people, and the MIT Mystery Hunt is scaled for teams with infinity people.) But there are other ways to be ridiculous without being huge: it can be things like using a car for the entire scavenger hunt or there being five different board game puzzles, one of whose answers was [the answer to Marvels of the Ancient World].


    • As a data point both for and against your comment: I really really liked spending many hours working on The 10000 Puzzle Pyramid in 2015. And it also got backsolved when I was getting close to the last step and I stopped.


  17. I have a young son, and so I don’t start Hunting (remotely from Seattle) until after he goes to bed on Friday night (though I then stay late into the night and return on Saturday morning). In most years, this results in lots of entertaining puzzling. This year, I missed over half the Hunt. So from that perspective, this year’s Hunt disappointed me. However, all the puzzles I *did* work on Friday night were fun and clean, and I enjoyed them a lot. So the time I *did* spend was quality time.

    Ideally, organizers would know how long their Hunt is and inform the teams in advance, so expectations are managed. In practice, I realize this is virtually impossible. But I think that’s the fundamental issue: from year to year, nobody knows what you’re going to get, and so you can’t set your expectations properly.


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