The night hunt ended, Anand started making slides for wrap-up. He was going over them with me and I noted a blank slide towards the beginning with just the title “Design Philosophy” and nothing else.
“What is that slide?” I asked him.
“That’s where you’re going to talk about your design philosophy in running this hunt,” he said, as if that explained everything.
“My what? I don’t have a design philosophy. I just ran the hunt by the seat of my pants.”
“Yes you do,” he insisted, “and it led you to run the hunt the way you ran it.” Which led to us having an extended conversation about what this supposed design philosophy might be and what I was supposed to talk about. And it turns out, he was right — I did have a design philosophy, even if I wasn’t cognizant of it at the time. It also isn’t solely mine — it’s mostly a shared vision that most members of Alice Shrugged believe in. I just was granted the ability to make the final decision on things that did or did not facilitate said philosophy. (And to incorporate more baseball than usual into the hunt, which was a vision mostly shared only by Jason and Harvey.) Given how well-received our hunt was, I’m writing this down with the hope that One Fish Two Fish Random Fish Blue Fish reads it and considers many of these issues when designing the 2015 hunt.
As a preamble, a lot of this stems from my personal belief that the ballooning team sizes of recent years is a problem. We did intentionally bottleneck the end of the hunt so that large teams wouldn’t have an advantage solely based on manpower. Our release rate was designed such that a team should have ~15-20 puzzles open at a given time and no more (and maybe less), but that they should simultaneously advance through the story of the hunt at a faster rate. (And so puzzle release was done on a round-by-round basis — solving puzzles in the Mock Turtle round would not help you open puzzles in the Tea Party and vice versa. But once you advanced past the MIT round, you probably had three Wonderland rounds open at any given time… or you were approaching the end of the hunt.) There were some large teams that complained about this. To this, Alice shrugs.
Make the hunt fun for small teams – Too often, and we did it in 2004, the team that wins the hunt writes the next hunt only thinking about how their team works and how they would approach things. The problem is, the team that wins is usually a huge mega-team of 100+ people or a “super” team composed of cracker jack solvers. A small team of >10 undergrads who have never hunted before is, realistically, never going to win the hunt. But that doesn’t mean that you should discount their experiences. We had a handful of people on Alice Shrugged that had previously hunted on small teams, and listening to their thoughts was invaluable.
Small teams of undergrads are, in my mind, the heart of the hunt. The big teams that have been around for years are going to continue to be around, almost no matter what you do. By contrast, small teams of inexperienced solvers — teams that are mostly undergrads, who by the ASA charter ought to be the target audience of hunt — are only going to come back year after year if they’re having fun. These teams are also full of new solvers that will actually remember most of what they saw, as opposed to new solvers on large teams who tend to not really contribute much their first year or two because more experienced solvers just get to things first and faster.
Nothing was more satisfying than running into Nu Tau Epsilon in Lobby 7 at 4 am on Saturday morning, who saw me camped out as the Cheshire Cat where I was waiting to meet another team. They had, at this point, apparently only solved three puzzles (and eventually solved 12, calling in their last answer Sunday at 2 pm). But they were having a great time, they were quick to tell me, and were still actively working and solving puzzles. I suspect that this is directly correlated to our opening “MIT round” where I held the 24 puzzles in that group to some pretty strict testsolving benchmarks (2 hours maximum from our testsolvers, preferably under an hour). While it’s true that the larger, more experienced teams blew through these by early Friday afternoon, it didn’t hinder their enjoyment to have this opening round of puzzles (and Bumblebee Tune-A is pretty hilarious, even if it only takes you a half hour to solve), but it made all the difference in the world to the Nu Tau Epsilons and Duck Soup (the last team to reach the Jabberwock, I think).
We had 54 teams call in a correct answer this year. I think this might be a record. I want it to be broken soon and often.
As many teams as possible should reach a satisfying completion point – I think the Hell hunt of 2007 was the first hunt to stay open after the coin was found. This has been a refreshing change, both for teams that ultimately finish the entire hunt, but also for teams that just enjoy solving as many puzzles as possible before the weekend ends. We had eight teams finish this year — possibly another record? (Data is spotty.) The mantra “The hunt doesn’t end when someone finds the coin; it ends when HQ closes,” was repeated a lot amongst Alice Shrugged. I hope it gets repeated amongst Random as well.
In addition, this year we had the “hunt the Jabberwock” interaction which came after the MIT runaround, in which teams were given a deck of cards as a sort of substitute coin. 31 teams (out of 54 that solved a puzzle) reached this point, including two that got there just after 6 pm Sunday when we were nominally closed. In addition, we grabbed one team off the MIT runaround and let them have a Jabberwock hunting experience along with the last team to officially reach that point, just because we wanted to make sure they had that chance before we collapsed in a pile of our own exhaustion. We received a lot of great feedback on this — teams that knew they were never going to win the hunt were actually able to reach a satisfying capstone to their experience and go on a runaround that tied a lot of things together (i.e. the locations of the puzzles on the map). This was an intentional design choice and it worked as planned. Plus everybody likes getting stuff.
Remember that it’s the MIT Mystery Hunt – Emphasis on MIT. There are a lot of people who take part in hunt who never went to MIT. Many of these folks complained about Cards Against Wonderland having too many inside jokes or getting totally stuck on the relay puzzle because they don’t know the MIT courses. Well… too bad. If I were to participate in the Game at Stanford or the Scavenger Hunt at U Chicago, I expect that Stanford/U Chicago students and alums would be at an advantage. And that’s how it should be. So write puzzles that utilize art around campus, The Tech, Technique, or knowledge of MIT dorms. Make people get up and runaround and explore the various courtyards and other strange areas of campus. Utilize the fact that you’re at MIT when designing your story. Plenty of other puzzlehunts (DASH, BAPHL, etc) are designed for a universal audience. This is the only one at MIT.
Being on campus should be an advantage – A more general point that encompasses the above point, but with a slightly different angle about it. Out in California, there are a lot of “West Coast style” puzzle hunts that involve getting in a car and going from location to location to find all the puzzles. People who do these hunts refer to the Mystery Hunt as a “conference room style hunt” — and I’ve seen it remarked that “conference room” hunts are less fun because you could just as easily be at home in your underwear solving the puzzles.
Make that last statement not true. Give teams objects that they actually have to use to solve parts of the hunt. Hide puzzles in the first aid kit. (Incidentally, the safety office did not require us to hand out first aid kits this year, but since we had already written that puzzle, we actually had to go to MIT Medical and request them anyway.) Write runarounds that involve using things on campus. Utilize interactions. This isn’t MUMS — give people a reason to not just stay at home solving puzzles in their underwear. (And while you’re at it, do so while using the fact that you have the bulk of MIT’s campus as your staging ground with all of its hidden nooks and crannies.)
Events should be fun and encourage inter-team collaboration – Back in the day, there weren’t “events” but a mid-hunt party, where HQ would shut down and stop taking calls and all hunters would get together and just take a break and have fun. Eventually, this morphed into the party containing a puzzle… which in some hunts became a puzzle that wasn’t solvable at the event and kept teams competing against each other at all times, which tended to cut down the fun of the event. We had a rule: if a team came to an event, they were getting an answer when they left. We had another rule: if a person came to an event, they were going to have to work with people from other teams to complete whatever task we gave them. (I don’t want to even call them puzzles, because Cards Against Wonderland was literally just “Come play this fun variant on a popular game for awhile and then we will give you an answer.”) Focus on making the events fun, and leave the clever, brain busting puzzles for the rest of the hunt. We had a small amount of feedback complaining about the lack of real puzzles, but the overwhelming majority of people praised us for planning fun activities that encouraged socialization with people not on their team.
Give yourselves ample opportunities to have fun while running the hunt – If the team running the hunt is having fun, this will trickle down to solvers. We received a lot of feedback about how great it was to interact with us during hunt and how positive and supportive we were. And that’s because we were having fun ourselves. We enjoyed dressing up in silly costumes and so we scheduled a lot of that. (Maybe a little too much too early as Friday evening started to get crazy, but overall, I think it was good.) We were having fun with it and so were solvers. If you’re not having fun, then you’re doing it wrong. [This is probably good advice throughout the writing year too.]
Focus on what’s fun for the solvers not what makes the constructors look smart – During testsolving, we added a feedback slot for “fun” on our puzzles. This proved to be an incredibly useful tool in determining how to tweak a puzzle, or maybe even outright cut it. Novice puzzle writers especially often try to prove how smart and clever they are by trying to write puzzles that are too difficult and too sloggy for the solvers. “Oh, testsolvers got that right away, so I should make that more obtuse,” is a common gut reaction from new writers, but it’s the wrong reaction. If you’re writing the Mystery Hunt, that means that your team was good enough to win the Mystery Hunt, which means that you’re the cream of the crop of solvers. So don’t be afraid to take the metaphorical red pen to a puzzle that your testsolvers aren’t enjoying or can’t solve. A puzzle author’s victory is not when a puzzle is so clever and tough to crack that it’s only solvable via brute force and unchecked a-has; it’s when it’s satisfying and fun to solve. As I said to Boston magazine: writing a hunt is like designing a game that you want to lose. Let the solvers be the winners and then you’ll win too.
It all comes down to the puzzles – In 2004, we designed a hunt structure that was a departure from the traditional hunt and included many novel elements that hadn’t been seen before. We had objects and maps and a kickass runaround (finally online after 10 years). But all of these things were overshadowed by the fact that the puzzles were often terrible and underclued and broken and had answer words that weren’t words. You can have the most fascinating theme, the zaniest final runaround, the most well-designed costumes and props, the highest production kickoff, the most professional looking website, but if your puzzles are terrible, very quickly no one is going to care about the rest of this stuff. So do whatever you can to write good puzzles. It’s a bit of a science; it’s a bit of an art. It’s definitely a learned skill. If you’re a new writer, your first puzzle is going to suck — mine did. Listen to feedback and figure out how to make it not suck (which might involve throwing it all out and starting from scratch). Seek out the advice of more experienced writers. But don’t throw out the ideas of the novice writers, because they are full of new and outside the box ideas that keep the hunt fresh.
Mostly, just have fun. Writing a hunt is hard. Being in charge of a hunt is especially hard and often stressful, but if you have the right mindset, it’s also incredibly satisfying when it all comes together.
Colossus of Rhode 
This is fantastic! I hope Random sees it and takes it to heart. I don’t agree with every detail, naturally– different teams will certainly have different values– but the core lesson of being clear about what you want and focusing on it is something every writing team should learn. And as a team with a lot of novice writers I particularly hope they understand that if your puzzle can be fun, clever, or hardcore, fun wins the day every time.
(The hunt was fantastic too, of course!)
“I think the Hell hunt of 2007 was the first hunt to stay open after the coin was found.”
Actually, if memory serves me, it was SPIES in 2006, but they announced it right as Hunt was starting and caught everybody off guard. I get the sense that more teams took advantage of it in 2007, and it might have helped that we advertised prizes for teams that filled their “sin meters” (I distinctly remember the Simmons team calling in an answer Sunday afternoon to complete their wrath collection). I definitely don’t think the idea was ours, and I’d feel bad taking credit because it’s such a good one.
Ever since then I’ve staunchly pushed the idea, both as a constructor and a solver, that people should keep solving after someone wins; even for our own team, this has been an uphill cultural change, as we lost all but a dozen hunters by the time we finished 2010, we had more than that for the final runaround in 2011, and then in 2012 a majority of the team stuck around to be in Ogre of La Mancha. I think/hope this was the year when the community finally said, “Hey, you can actually have a whole bunch of teams do the final runaround,” and that cutthroat teams finally stop equating the end of the Hunt with the first-place finish time.
And as I’ve said elsewhere, the “satisfying completion point for small teams” concept was something that I think Evil Midnight first tried to innovate with Zyzzlvaria, but we botched it in a number of ways. The strict standards Erin mentions are the reasons it worked so well; for that I thank her and also forgive her for burying my only two MIT puzzles in the back. (During the Hunt, I wasn’t even sure anyone solved Across The Hall forwards, though I think a few teams have mentioned doing so.)
Good post. (And good Hunt, but I’m biased.)
“And as I’ve said elsewhere, the “satisfying completion point for small teams” concept was something that I think Evil Midnight first tried to innovate with Zyzzlvaria, but we botched it in a number of ways. The strict standards Erin mentions are the reasons it worked so well[…]”
Just to clarify due to potentially confusing pronouns: Evil Midnight did the botching (not that Zyzzlvaria was a bad Hunt by any means in my opinion, but finishing Inner Zyzzlvaria was not as every-team-friendly as we intended). The completion point that worked really well was this year’s MIT round and subsequent Jabberwock runaround.
I don’t know how well Inner Zyzzlvaria worked for truly small/inexperienced teams, but there’s a middle tier between “tiny and doesn’t know what they’re getting themselves into” and “competitive for finishing hunt”, and my team that year (Optimus Funk or R2-DISCO or whatever we were called) really enjoyed getting the “finished a meta meta” feeling from that. (Especially nice given that we solved no Outer metas.)
Based on survey feedback, at least five teams forward solved Across the Hall. And it got all 4s and 5s for fun too! I didn’t notice that I had both of yours buried. That had more to do with the fact that they were both diamonds, nearly all of which were intentionally pushed towards the back.
SPIES stayed open after the coin was found, but I don’t *think* we announced it as Hunt was starting: I’m fairly sure it was a spur-of-the-moment decision we made because the coin was found really early.
In the Hell Hunt I think you kind of prefigured the completion-point-for-small-teams idea by putting “The Runaround” in the middle of the Hunt so that all teams would get to experience it, which I thought was a nifty idea. In the Mario Hunt we tried to echo that with the mini-runarounds after each of the first few gameworlds. We also had the aim of the Mario gameworld being a mini-Hunt that small teams could feel satisfied to finish, though it didn’t occur to us to impose standards like the two-hour-testsolve limit, which is an interesting idea that seems to have worked pretty well for you guys.
Plant forward-solved Across the Hall! It was one of my favorite puzzles, actually. Seth and I would be on one side of the room solving the clues, and then we’d should across the room, like, “is BEXLEY a dorm?”, and the people on the other side of the room would say “YES!” and find it on the map.
If I remember correctly, in 2006 there wasn’t a plan to stay open, but since the coin was found relatively early we decided to stay open basically on the spot. (At the very least I can’t find any emails discussing staying open prior to the email actually announcing to all teams that we’d stay open a few more hours.) We also only stayed open for around 3 hours after the coin was found, while 2007 stayed open quite a lot longer. So I think the Hell hunt deserves much of the credit for really taking the “stay open” idea seriously.
I don’t think that this year was all that different from 2011 and 2012 (the other two years with announced end times) in terms of teams trying hard to finish after the coin was found. In 2012 five teams finished, and if they’d stayed open until 6pm instead of 3pm there’s a very good chance that would have been seven teams, so I don’t think it was substantially different from this year. (In 2011 we only had 4 teams finish, but that’s because the puzzles weren’t as clean as in 2011 and 2014. Two teams finished at the last possible moment.)
My 15-person team in 2006 solved our only meta about half an hour after the coin was found, and we really appreciated that somebody still came down and acted out the James Bond skit for us. It was so exciting! And also nice to feel like HQ cared about our experience.
I don’t remember ever hearing about a plan to stay open late (until it actually happened, anyway). But then again, we weren’t really paying attention at kickoff so we kind of didn’t know there were antepuzzles in addition to metapuzzles, either. So maybe I’m not the most reliable source.
We didn’t announce at 2006 kickoff that there were going to be antepuzzles in addition to metapuzzles. They were supposed to be a surprise. In retrospect that may have been a bad move, since it seems to have led to a bunch of frustration (on the part of teams that didn’t find them) or confusion (on the part of teams that did find them and didn’t know what to do with them).
Huh. For some reason I thought somebody at wrap-up had told us that the puzzles were mentioned at kickoff. I guess I feel less dumb for not hearing about them?
We did solve the first ante pretty early on but didn’t call it in because we didn’t know we were supposed to — we just sort of assumed it would play into the meta eventually.
So I think the confusion about what was announced at kickoff was, at kickoff we announced that the name of the agent in Boston was Ethan Hunt, and when you called in a meta answer you needed the agent’s name as well. So people would forget that, and call in the first meta answer DISAVOWED, and we would say “for what agent?”, and they wouldn’t realize that was information they had. (In other rounds, the name of the agent was the solution to the previous round’s antepuzzle.)
Figuring the antepuzzle would play into the meta or something else eventually and not calling it in was a “confusion” response that we just hadn’t anticipated, but a lot of teams seemed to have had. There was one team—I don’t remember which—who solved everything in round 1, including the meta, and then were stuck with nothing to do, and someone from HQ went and visited them and they were like “we’re stuck, we don’t know how to unlock round 2, all we have is this JAMES BOND / CAMBRIDGE we found in the puzzle header data” and HQ was like, um, have you tried calling that in? Finding something that *looks* like a puzzle answer, even in an unexpected place, and *not* calling it in was something that it hadn’t occurred to us that teams would do—it had worked for TAKE THE RED PILL in 2003, after all. (In retrospect there are a variety of ways that wasn’t comparable, of course.)
Once it became clear the 2006 Hunt would end on the early side I knew we would try to stay open and keep people solving, though I’m not sure whether I communicated that to much of the team. I think it was the coin is found email that said we would still answer phones while people were solving, but it ended up that almost everyone packed it in really fast so we weren’t open much longer. Hell gets all of the credit for the “nothing has changed, you can still do every single thing in the Hunt” that was so massively successful.
Nice writeup. As a puzzle editor and meta author for [Alice Shrugged], that largely matches my sense of our team’s design goals. (I have personal opinions that differ from the team’s/Erin’s, but they’re mostly a matter of small details rather than overarching philosophy.)
In particular, I’d like to reemphasize three points, for any Random Fish who might be reading:
-Have a rigorous formal editing and testsolving process, led by experienced people, and stick to it. As Erin said, the main disaster of 2004 was that the puzzles were bad and/or broken; we did much better in 2014. Of all the statistics in wrapup, I’m most proud of the fact that every single puzzle was solved at least seven times. (Well, it’s a tie between that and the QUAGMIRE stats. 🙂 )
-Interacting with solvers–over the phone, and especially in person–adds a human element that drastically increases their fun. If someone drops by a small, noncompetitive team not for Hunt-plot reasons but rather just to make sure they’re having fun (and maybe drop a few hints to get them unstuck on a few puzzles), that really makes an impression.
-I think having a more-approachable MIT round was a great idea this year (I can say that since it wasn’t my idea originally). I hope that, no matter the overall structure, Random front-loads some easier puzzles. Sure, competitive teams will blast through them, but so what–small teams of first-time Hunters will have more fun, since that’s the part of the Hunt they’re likely to spend more time in.
I don’t regret including MIT-related content in the Hunt, and completely agree that this is to be expected in an MIT puzzle hunt. However, some of the criticism on this point is more that if there’s a puzzle that involves sending a few people to perform it (such as an event or the relay), it would be useful to know in advance if MIT knowledge is required/recommended. Someone largely unfamiliar with MIT culture could certainly get the answer to the Cards Against Wonderland event, but would have far less fun than an MIT undergrad. I think that point is essentially fair, since unlike “normal” puzzles, for these interactions it’s much harder or impossible to find the MIT people on your team to take a look at something.
In-person interactions are a big thing. I was very pleased that Alice Shrugged made at least 5 or 6 visits to Immoral, Illegal, and Fattening’s headquarters. Then I asked my friends on Left as an Exercise for the Reader how many times they got visited, and they said 2 or 3.
II&F and LAAEFTR are next door to each other. There is no reason there should have been such a disparity. Sure, some people came just to check out our sashes, but if you’re already on the hall, stop by and say hi to the other teams while you’re at it. I think last year or the year before I heard complaints from other friends on Grand Unified Theory of Love that they felt ignored by HQ most of the time. So Random, if you have the people to spare (and I suspect you will) make visits to all the teams, not just the ones you’ve heard of or have friends on.
I’ve been hunting for about a decade now, and I’ve never witnessed an instance when we were bothered by someone from HQ coming by. And I’m sure that if we ever were too busy for it, we’d politely ask them to come by another time. So by all means, if you’re on a team that’s running the hunt, please take the opportunity to go visit a team or two and watch them solve your puzzles. It’s a great experience for everybody!
Having been on a lot of running teams and having done a lot of visits, I can say that I have visited teams (this year and in other years) that either (a) gave the impression they were uncomfortable being visited or (b) were rude or obnoxious to the point where I had no interest in visiting again (for example, by immediately criticizing puzzle authors in a tactless manner).
I’m not saying that either of these was true of Grand Unified Theory or Left As An Exercise (I visited both and don’t remember any major issues with either), but I just wanted to clarify Mike’s and Nathan’s points in saying that there are some teams that at least seem to want fewer visits, and there are others whose behavior makes us less likely to stop by.
Dan’s point is sad but fair. I do wish that teams would keep in mind that, no matter how frustrated or stressed they may be by hard puzzles, the running team has endured a lot more stress over the past year, so be polite. Perhaps II&F is extra conscientious of this fact since we have a lot of folks who have helped run hunts in the past.
“If you’re writing the Mystery Hunt, that means that your team was good enough to win the Mystery Hunt, which means that you’re the cream of the crop of solvers.”
I just want to agree with this, loudly. Don’t fall prey to the Impostor Syndrome.
So, regardlng “solving puzzles in the Mock Turtle round would not help you open puzzles in the Tea Party and vice versa”—I’d like to explain why, at least in the past, I’ve criticized systems like this.
What I call Principle II of Hunt construction is “solving a puzzle should be helpful”—that is, every puzzle you solve should get you closer to the coin. Most of the time, the main reward you get for solving a puzzle is that the answer is an input to the meta. But if you solve a meta without solving all the puzzles that feed it, then all of a sudden you might have no Hunt incentive to go back and solve (or backsolve) any of the remaining puzzles in that round, which is sad.
In many Hunts, this problem is dealt with in endgame: some phase of the final runaround depends on having puzzle answers, so you have incentive to solve puzzles that you don’t need as inputs to metapuzzles because they might be required at a later point. One problem with this is that it only affects teams that actualy make it to endgame; it’s nice if other teams can get a reward for backsolving or late solving as well. But I also soured on that system even for winning teams because in both 2006 and 2010 some particular puzzle answer was demanded during runaround that a team just *hadn’t* solved or backsolved, and they basically had to pause the runaround for however the half hour or whatever it took them to backsolve the answers they needed, and that just wasn’t fun for *anyone*. So for 2011, I pushed for the reward for solving to be simply that solving puzzles in *any* round brings you closer to unlocking your next puzzles—so that puzzles from early rounds wouldn’t become obsolete just because you’ve solved their metas.
Solving more puzzles after you had finished the meta did help, to a point — you needed so many points for the train tickets to the rest of Wonderland, if I recall correctly. But yes, solving puzzles in one round had no effect on any of the other open rounds, and when the release schedule was sometimes 2 puzzles for every 2 answers rather than 1 for 1, it could be frustrating to have 4 puzzles in the Caucus Race still locked and not get to see any of them after solving another.
I totally agree about the danger of requiring puzzle answers in the runaround, though I have to say that watching Evil Midnight conjure the answer WYCLEF from thin air in 2006 was pretty amazing.
When We’re Up All Night to Get Lucky solved the Caucus Race meta with 5 of 14 answers, and 4 puzzles there still locked, it was this very solve that opened our last round — and we soon noticed that the ticket meter disappeared from our pages, telling us there were no more rounds. So we indeed ignored the other Caucus puzzles for the next 11 hours while we solved the other metas. They came up during the answer check puzzle during the runaround, but since we only needed to extract a letter from each one, we had enough answers to get by.
Great hunt! In particular, the puzzle editing was really great (probably the best ever).
I did some poking around for records. I’m quite confident that you’re right that 8 teams finishing is a record. I think two hunts had five teams finish (2012 and 2007) and one had four teams finish (2011). (In 2007 there were several more teams with all but one meta.) For the “teams finishing a major subhunt” I think you also got the record: in 2011 we had 27 teams finish the Mario world, in 2007 there were 31 teams that did the midhunt runaround, and in 2009 I think 21 teams got through inner Zyzzlvaria. I can’t think of other hunts with something comparable. As for the number of teams solving a puzzle, I couldn’t find anything over 45. So, I’m pretty confident you have the record there too.
Great job! Those numbers set a high bar, but as you say I hope we see some of those records fall in the future. In particular, I think someone should be able to get to double digits of teams finishing, which would be a real milestone.
10+ teams finishing would be pretty crazy, in terms of runaround logistics. This year, We Had A Plan (TM) for how to deal with up to three teams in a competitive runaround, and we managed to juggle three teams again on Sunday afternoon. But just from a personnel perspective, we were pretty strained.
For Random, that’s probably something worth thinking about more carefully when designing the endgame.
I’d happily settle for double digits finishing everything but the endgame. That said, I’m a fan of short endgames. The 2007 hunt certainly would have had no problem getting any number of teams through endgame!
“I’m a fan of short endgames.”
Me too. Although I can say that most teams that finished in 2007 found the lack of an endgame runaround anticlimactic; so there is such a thing as too short.
Yeah, even though I like short, 2007 was too short. I really like the idea of having the key part of “endgame” happen at team’s HQ like with the Hell meta-meta and the 2010 first round re-solve and scavenger hunt (and in a different way, the beginning of the 2011 endgame). But I think there still has to be at least a runaround after that or else people are going to be disappointed. In my opinion, Normalville had the perfect endgame: short, funny, fun, and very low likelihood of teams getting stuck. (Though that was my first endgame, so I might be biased.)
I think that 5-6 hours is too long for sure. Personally, I’d be perfectly satisfied with about 1-2 hours, but it’s sensible to plan for how to deal with logistics in case your endgame lasts 3 times as long as you think it will.
Noah, wasn’t Time Bandits your first endgame? Or were you not there for that?
I think one sensible plan for how to deal with endgames that last too long is just to decide that you’re not going to let each step take longer than x time. If teams get stuck for that long you just hint and help them after a certain amount of time. There’s a reason that sequential hunts handle getting stuck differently (hints, skipping people on puzzles, etc.).
(None of this is meant to criticize this year’s endgame in particular. I think this endgame was well above average, even though it had some problems. But I think in general endgame is an aspect of Mystery Hunt that’s mostly broken, like say events were prior to this year.)
Nope, I spent time bandits trying to stay awake on Jeff’s couch in Berkeley wondering what the hell was taking so long.
“I think one sensible plan for how to deal with endgames that last too long is just to decide that you’re not going to let each step take longer than x time.”
We actually did that this year, although it was still pretty long. But after 45 minutes, you were supposed to get a hint. After 1 hr 30, you were supposed to get a blatant hint. After 2 hr 15, we were just going to move you along (although no one took that long).
This was definitely all made up seat-of-our-pants when we had three teams on the final runaround and Random was taking far longer on the bed than any of our testsolvers. (We had two independent groups solve it in ~25 minutes. No actual team came close to this — I know Codex did ~45 with a hint. I was really surprised by the disparity, but maybe how tired you are really comes into play in a way we didn’t predict. Also, our team generally loves hands-on stuff and that might not be universal.)
It seems to always happen once or twice a hunt that the testsolver experience is just wildly unrepresentative. Although it’s tempting to come up with case-by-case explanations, I also think it may just be that the naive odds of two testsolves both being off by 1 standard deviation in the short direction is over 2%, so you should expect it to happen a couple times per hunt.
The bed in particular is probably way harder in an enormous group. I was asleep during Codex’s run of endgame (boo), but from the video it looks like we only really cracked it when one skilled person took charge of manipulating the bed.
So, the bed. I’m probably the person who knows second-most about the bed puzzle after Feldmeier (who singlehandedly built it), and I watched six of the eight teams solve it on the final runaround. In retrospect, I have a few ideas about why it took longer in the runaround than in testsolving.
It’s a physical puzzle that’s very easy to play around and fiddle with while getting nowhere on, and it’s less obviously a “standard puzzle type”. With a large group, if there’s an obvious puzzle type (e.g. nurikabes and chess in the Lion and the Unicorn stage), people will let the people who are best at those puzzles at them. But for the bed, anyone can fiddle with it, so lots of people did.
It also requires a few steps to be performed in sequence, while *not touching* the rest of the parameters–so a group of people would only solve it if everyone was following one lead. And if any single person takes charge (which could sometimes take a while, since there was often a lot of fiddling), it’s more likely to be someone who is enthusiastic about directing the solving effort rather than the person with the best chance of figuring out the puzzle. Random eventually solved it after Eric Price took command, but it still took them 95 minutes or so. Lucky took a similar amount of time or even longer–after their blatant hint at 90m, they eventually brought in their lockpicking person, who hadn’t even been involved until then.
So it sort of makes sense that a small group (when I testsolved it, there were three of us) would solve it faster–it’s much easier to coordinate in general, and much easier for the one person with a good idea to be heard, which was what was necessary to solve this particular puzzle.
(There were a few other factors–e.g. because of context the testsolvers had a stronger inclination that the puzzle was self-contained aside from the allen wrench, whereas teams knew to use the allen wrench but were also tempted to e.g. use answers from Hunt rounds somehow. Also, our testers weren’t sleep-deprived.)
In general groups a group of 35 is going to solve a lot of puzzles slower than a group of 5 would. For the bed I wonder in particular whether having fewer than 9 is a huge advantage? Once you have 9 people it really looks like the puzzle wants you to have one person actually holding each of the handles. But in reality you don’t want people touching the handles as they’re more likely to mess things up that way.
For Plant I think the two main things that slowed us down were:
1) Not knowing in what ways the bed was *not* like a lock (e.g. you don’t have to worry about losing progress, so the person pulling and pushing on the end didn’t have to be careful like they would on a real lock).
2) Difficulty communicating between the people who were lock pickers (of which there were several at the table) to the subset of people who weren’t lock pickers (e.g. me, though I got out when it was clear I wasn’t helpful). In particular, what it meant for a handle to be stuck or not took a while to communicate. This was largely because it would often be that you could move the handle, but not easily or say only between numbers 2 and 3. This lead to much confusion of people saying their handle was stuck when it wasn’t.
The following feels like it could have been written specifically for Manic Sages:
“””
Focus on what’s fun for the solvers not what makes the constructors look smart […] Novice puzzle writers especially often try to prove how smart and clever they are by trying to write puzzles that are too difficult and too sloggy for the solvers. ”Oh, testsolvers got that right away, so I should make that more obtuse,” is a common gut reaction from new writers, but it’s the wrong reaction. If you’re writing the Mystery Hunt, that means that your team was good enough to win the Mystery Hunt, which means that you’re the cream of the crop of solvers. So don’t be afraid to take the metaphorical red pen to a puzzle that your testsolvers aren’t enjoying or can’t solve. A puzzle author’s victory is not when a puzzle is so clever and tough to crack that it’s only solvable via brute force and unchecked a-has; it’s when it’s satisfying and fun to solve. As I said to Boston magazine: writing a hunt is like designing a game that you want to lose. Let the solvers be the winners and then you’ll win too.
“””
I think this was probably the number one mistake we made in writing the 2013 hunt, and possibly the number two through five mistakes as well.
Which is just to say: LISTEN TO THIS PERSON RIGHT HERE!
What are your thoughts on the somewhat-conflicting goals of “don’t swamp small teams by flooding them with too many puzzles they don’t have time to solve” and “give small teams a chance to at least see all the puzzles in the Hunt, so they can solve the ones that interest them and skip the others”?
I was theoretically in charge of deciding the time-unlock pace during the Hunt, based on how the overall Hunt was progressing. We had a lot of debate seemingly all year about whether and when to open floodgates and risk swamping smaller teams.
What ended up happening was that we dumped all the puzzles on Sunday morning, which felt awfully hasty. A lot of this was because I was hesitant to step on the story progression by unlocking Wonderholes for teams who might otherwise have “earned” them on their own (by solving an MIT meta, with or without hints).
Our graph structure and non-linearity made the time release hard to control, but overall I think we undershot the ideal release rate and we should have kept up a steadier trickle of new puzzles throughout the first 24 hours. We *had* talked about establishing a canonical puzzle order and just steadily releasing puzzles in that order over time, but we never seemed to come to enough of a consensus to implement it. We also had discussed some complicated schemes for automatically opening puzzles based on solve rate (but that still wouldn’t have fixed the question of how to open the Wonderhole rounds in the first place).
To get into specifics, we waited way too long on Friday and Saturday to trickle new puzzles out to people. It was 6pm before we did the first time unlock, and then trailing teams spent most of Saturday with just the 24 MIT puzzles open. We had sort of decided in advance that we’d try to hint teams into the Wonderholes instead of pushing them in with a time-release, but by late Saturday it seemed clear to me that any team who hadn’t made it there yet preferred more puzzles instead. And then on Sunday, we were forced to do a big unlock to try to “catch up”. (Admittedly, one of my motivations for getting the late rounds open early on Sunday was to make sure everybody could get to the Duck Konundrum, which was, by bad luck, located in a later round.)
I’m not sure how to more effectively measure when small teams are tired of their current puzzles and would prefer to see new ones, other than by just visiting often (which would be ideal, but perhaps not very scalable). Some teams were willing to let us know via the ‘Contact HQ’ link, but most teams are likely too shy to do so.
As far as more general thoughts about unlocks (not specific to what we did this year), I’m usually in the camp that it’s safer to unlock too many puzzles, and teams who don’t want to look at all of the puzzles can just … not look at them. But I’m sympathetic to the idea that it can be overwhelming — I wonder if there’s a balance to be struck by allowing teams to control their own release pace somewhat. (The implementation may be tricky, but perhaps they could tune a setting somewhere to move between “give me all the puzzles as fast as you want please” and “don’t show me everything, I’m happy with what we’ve got”.)
This was the source of much debate amongst those on our team who had been on very small teams. (Jason (aka aerionblue) was of the camp that said that he liked having access to as many of the puzzles as possible and could ignore the rest. Other people, including one who had been on the same small team with Jason, said exactly the opposite and that he hated being swamped and liked to continue progressing through the hunt over the whole year.
And so what happened, as Jason more or less laid out, is that we went in circles and were never exactly in agreement over what to do and at the last minute (literally Thursday night), I just punted all the responsibility and decision making to Jason. It was by far the least thought out thing that we did, and as he said, it definitely deserved more thought than we gave it.
I suppose my final opinion on that matter is that small teams differ on what they prefer (even within the small team), and that if there was a way to give each team the experience they think they’d have more fun with, then that’s what should be done. Which is a pretty non-opinion opinion. (But if I was on a small team, I *think* that I would ultimately want access to all puzzles, even if it meant being overwhelmed, so that I could find all the cryptics and Nikoli style logic puzzles that I love.)
I’m pretty sure I’m the “other voice” Erin was talking about. My feeling is that I get a really big solver high when the team I’m on solves a meta puzzle. Even better when doing that gives something tangible, an interaction, a new round, or whatever. I prioritize this even over getting puzzles that are in my proverbial wheelhouse, instead preferring to slog through puzzle types I dislike to advance the meta knowledge on a round that we are close to. I do think that I’m in the minority here, possibly the very small minority, so I apologize if my opinion had an outsized effect on the outcome. This wasn’t the only point of distinction between Jason and me back in our smaller team days. We also disagreed, and still do, on whether football is interesting enough to take precedence over puzzle solving.
If I had to give advice it would be to make sure teams know that they can liberally use the “contact HQ” message for anything. They’re stuck on a puzzle, they’re not sure about some aspect of the plot, whatever. On Friday evening we got a “contact HQ” from a team that said that they weren’t having fun. Harvey and I immediately went to visit them, and we steered them through one puzzle which they were doing correctly but somehow got some very bizarre transcription errors. The fact that they felt free to contact us and we would either send someone or talk to them on the phone, really helped their experience. I think if teams were willing to contact HQ with, “we’re tired of looking at the puzzles we have, can we have some more please?” it would solve a lot of the tricky unlock problems and allow you to tailor the experience to smaller teams. Because, honestly, when you’re at the point where smaller teams want hints and you are willing to give them, worrying about parity between teams is dumb. Better to make sure each team is enjoying things to the best of their ability, and really hammer in the point that they should talk to HQ often.
The other piece of advice is to make sure you at least have the ability to simply unlock any given puzzle or round to all teams or a specific team. We didn’t have that and it required a specific set of people to be able to unlock puzzles on the fly.
More baseball than usual? I must have missed the puzzles about baseball. Of course, the normal is 1 puzzle or less so it doesn’t take much to be more than usual.
http://devjoe.appspot.com/huntindex/keyword/baseball
There was 1 baseball puzzle: http://web.mit.edu/puzzle/www/2014/puzzle/round_tripper/
And of course, a meta that incorporated non-trivial amounts of baseball. (“Make a travesty of the game” is the best phrase.)
Oh, right, I remember Round Tripper now. I even helped solve it, just not the baseball part of it. When I heard my teammates from the next table over saying they had numbers 46 and 48 through 57 that they couldn’t figure out, I told them they were ASCII numbers. That was it. It turned out that they were baseball stats that were supposed to help them identify the players who they had already identified from team data alone.
And yes, the White Queen round.
“It turned out that they were baseball stats that were supposed to help them identify the players who they had already identified from team data alone.”
It seems kind of weird that you’re trying to explain how the puzzle works to the people who wrote it, but for what it’s worth, I think you’re getting it wrong.
Jason has now reminded me that I really wanted to talk about the Jean Segura incident when it happened and how it was related to the White Queen meta, but obviously spoilers prevented me from doing so at the time. Also, the tale of Germany Schaefer, which you can read about in that link.
Oh, and Local Shorts included some baseball teams.
And PWTA I WANT TO HOLD YOUR HAND included a mini-puzzle involving baseball teams.
You can see this was mostly Rhode’s fault (not that I didn’t try to help).
I’d love to hear Erin or whoever also talk about our policy of hint-giving and team visits. I think doing both really helped small teams have more fun.
During the Mario Hunt, we tried to be more organized about team visits by sending someone (in costume) to visit a team whenever they unlocked a new round; but from post-Hunt reports I gather a lot of these visits just ended up kind of awkward.
Not sure what there is to say. Our official policy on hint-giving was:
-For small noncompetitive teams, give out hints freely on an ad-hoc basis in a manner that increases their fun/gets them unstuck.
-For competitive teams, strictly no hints on regular or meta puzzles, in the interest of fairness. The Director (that is, Rhode) might revisit this policy during runtime, but only if things are seriously broken, and she’s the only one who can make that call.
-On the final runaround, in the interest of time, at each puzzle location each team would get a hint at 45m, a blatant hint at 90m, and get forcibly moved on at 135m. (The last part was never invoked.) We cut these times down a bit for the later teams who did the runaround Sunday afternoon.
Our official policy on team visits was “feel free to visit teams”. We had ideas about visiting small teams more frequently early in the Hunt and larger teams more frequently later in the Hunt; I’m not sure if that actually happened. There was a board where we tracked which team had been visited when, so we would try to visit teams that hadn’t visited for a while. But it wasn’t officially timed or organized or anything, it was just if there was a non-crisis moment in HQ, some people who liked doing team visits would look at the board and then decide to go visit a few teams for a few hours.
I visited a few teams, and I don’t know how other people chose where to go, but Robbie and I basically looked for 5-6 teams that hadn’t been visited recently, and went around to visit them and teams who were geographically close to them.
Team-visiting could have been organized better. That being said, it’s hard to say how long an interaction will take–sometimes teams will engage with the visitors more than others. Also, there were a number of periods when staffing was an issue–we had a bunch of interactions, so there were times when we were coordinating three different “character interactions” and several different meet-HQ-to-get-this-puzzle interactions while still needing people in HQ to answer phones, monitor servers, and generally fight fires. Visiting teams is important, but because it’s often less *urgent* than other things it’s easy to put it off until we don’t immediately need people for something else.
One thing which is really obvious in retrospect, but something I didn’t think of beforehand, is that it’s extremely useful that everyone on the team that wants to go on team visits, know pretty well how the first 10 or 20 puzzles work, or at least have access to the solutions. In hours 6-12 of the hunt, the small teams are going to be stuck on these puzzles. You are going to be strapped for resources because the large teams are blowing through the early rounds and needing HQ interactions, so you really have to make those visits to smaller teams count. You shouldn’t be afraid of hinting early, especially to the team of 6 undergrads who have never done mystery hunt before.
As I learned through handling a lot of hint dispensing to smaller teams, hinting is hard.
Pingback: Linkage | An Ergodic Walk
Pingback: Mystery Hunt participation trends over time | An Ergodic Walk
Pingback: 12 Rules for Escape Room Puzzle Design - the Codex
Pingback: 13 Rules for Escape Room Puzzle Design - the Codex
Pingback: Two hundred puzzles, fifty years later | MIT Admissions