Every January I go to MIT to join my team, Palindrome, for the annual Mystery Hunt, a weekend-long immersion in dangerously difficult puzzles. I solve other puzzle hunts throughout the year, and am helping to create one that’ll run just about a month from now, but these tend to be smaller events, lasting a day or a few hours. The Mystery Hunt is non-stop puzzling — some people on my team don’t even stop to sleep.
On the other coast are other, similarly intense annual events, including the Microsoft Puzzle Hunt. Since the Microsoft Hunt is on the other coast, however, I have never participated in it.
But aha! This past weekend, the Microsoft Hunt had an online “simulcast” — any team anywhere could join in the fun. And so a subset of Palindrome warmed up their internet connections, logged into a chat room, and pretended like they were in Seattle. Over a 24-hour span of time between Saturday and Sunday, I think I solved puzzles for, oh, 19 hours.
This was a particular pleasure, because my 9-year-old daughter, Lea, wanted to join in, too… and she was able to! Together, she and I did the legwork on “Twister,” a puzzle whose first step involved adding letters to words so that the new words belonged to the same category. (So GAPE, EACH, LIE, and PLOT became GRAPE, PEACH, LIME, and PLUOT.) Another puzzle, “Road Trip,” turned into a chance to test Lea on her state capitals — we then connected those capitals on a map to form letters.
Lea didn’t stick with it for 19 hours, of course. But I sat in that kitchen chair for hours on end, working on the various Google Docs that served as my team’s online headquarters. My goal, as always, was to walk away having made at least some contribution to my team’s progress: I still remember the complete cluelessness I experienced at my first MIT Hunt, and I am always out to prove, at least to myself, that I’m a bit more puzzle-savvy than I was back then. Happily, I made breakthroughs on a number of puzzles, and in other cases said the magic words that caused other people to make breakthroughs, which is nearly as good. I stuck to the word puzzles, of course, and left the more complicated stuff to my brilliant teammates — one puzzle involved “folding cubes over the periodic table.” I still have no idea what this means.
Amazingly, despite not being in Seattle for the actual Hunt, my team found itself leading the pack for the longest time. We couldn’t maintain this, however: When night rolled around, Palindrome was reduced to a skeleton crew of its most fanatical solvers. The four or five of us did our best, but we began to sink in the rankings. Then there was the problem of two largish physical puzzles, one of which all but demanded that a team assemble it together as a group. Scattered across the country, we couldn’t do that. Last I checked, one of my teammates was trying to do it solo, but I’m not sure he succeeded.
So ultimately we came in sixth — a pretty good showing all in all. Not that winning is the point of these things. Having as many aha moments as you can in the space of a weekend: That’s the point. On that, the Microsoft Hunt succeeded brilliantly. I really hope the online simulcast becomes a regular feature. I will definitely be back.
Now. If I may put on my critic’s hat for a moment, I found myself flustered by one major element of the Microsoft Hunt — or rather, an element that was missing.
I am used to puzzle hunts that work as follows: Teams solve a round of puzzles, and then use the answers to those puzzles in a final, extra puzzle — the “metapuzzle.” In small hunts, there might be only one round of puzzles: The solution to the metapuzzle is the finish line. In larger hunts, there might be multiple rounds, each with its own meta. (And sometimes the solutions to all the metapuzzles are assembled into their own metapuzzle — a meta meta!) As puzzle hunts have grown in sophistication, puzzle-making teams have sought to dazzle with ever more elaborate and elegant metapuzzles. Metapuzzles are like the climax of a great story. They transform a bunch of scattered puzzles into a unified, satisfying piece of art.
The Microsoft Hunt didn’t use a metapuzzle structure at all, and I’ve since learned that other hunts, notably out of Australia, also forgo this common form of presentation. That would be fine if metapuzzles were replaced with something as good or better, but that is not the case.
The individual puzzles in the Microsoft Hunt are as good as any I’ve solved, but the structure of it was bewildering to me: Every puzzle had an answer, of course, but those answers were completely incidental. They may as well have been chosen at random, and might well have been. We dutifully recorded these answers, because that’s what we’ve grown used to doing, but we needn’t have bothered: These words, which we struggled to discern by solving complicated, dastardly puzzles, were never used again. What makes this extra weird, from where I stand, is that after finally solving a puzzle and submitting its answer, you were given “evidence” to help apprehend various criminals… and in at least one round, this evidence took the form of words. You essentially traded your hard-fought answer words for completely different words. The disconnect here is baffling: Why not just form puzzles around the answer words that teams will ultimately need? Why instead choose a bunch of other, random answers that play no part in the larger structure?
Since my team did not reach the finale (and I was long gone by that point anyway), perhaps those puzzle answers WERE used in the Hunt’s final moments. But I don’t get the sense this was the case.
There may well have been a good reason to ditch the metapuzzle structure. Indeed, this may be, to some, an equally satisfying and artistic way of presenting a puzzle event. If so, I hope someone can explain it to me, because I do not see it.
None of this is intended to take away from the puzzlemakers’ achievement. The production values of the Microsoft Hunt were beautiful, the use of technology was marvelous (I loved the “Whistle Stop” puzzles and am sorry I missed the last of these), and the puzzles themselves were largely excellent and a lot of fun. But a metapuzzle framework allows a group of puzzlemakers to work together to build something larger, like musicians taking their individual instruments and playing a symphony. The Microsoft Hunt was not quite this. There was a uniting theme, but the structure was incidental. Instead of a symphony, it was a bunch of very talented musicians, all playing their own separate songs.
i was stunned at how well co-solving over the internet worked. i know the google docs are great, but the chatroom and g+ hangouts really added a lot to the experience, too. doing the whistestop puzzles in hangouts was terrifically fun, but there were a couple of other puzzles that we broke into hangouts for and that really helped.
FYI, the periodic table one involved doing a color PBN to get a map of the periodic table with a bunch of color hexominoes, then cutting them out and folding them up into cubes, then arranging the cubes in a certain way and reading off the chemical element symbols. it was pretty freaking cool.
great time all around. sorry i missed almost all of sunday.
Okay, so PB is obviously for “peanut butter” but what does the N stand for?
You’ve echoed pretty much my usual comments about the Australian-style hunt structure, but I will say this: this hunt did it that kind of structure right. Usually, you were satisfied that what you were given was the correct answer at the end. No puzzles with random answers that are only needed because of the meta puzzle.
That being said. I prefer MIT-style metas. (Although MUMS metas are actually quite close some years, but they’re all shell metas. You can’t do anything with them until Day 5. That said, they’re also really cool usually. I really like 2010 and 2011′s metas for MUMS)
PBN = paint by numbers.
Well, in an ideally constructed puzzle hunt, none of the answers would seem “random,” because the constructors would have figured out a way — SOME way, ANY way — to tie it thematically to everything else that’s going on. Puzzability was particularly good at this in the mini-hunts they created for the New York Times. If the answer word wasn’t directly tied to the theme (as in YANKEE or PITCH in a baseball-themed set), then the puzzle or, at a minimum, the puzzle title served as a bridge between the theme and the answer. For their puzzle set “Space Case,” celebrating NASA, one of the answers was PRIEST. What does that have to do with space travel? Nothing… unless you cleverly create a puzzle called “Mission Control.” Now it fits.
Patrick Berry is also very good at this trick, as a walk through his Adventures in Puzzling will prove beyond a doubt.
When a given puzzle, its answer, and the overall theme are all in perfect sync with each other, that’s the height of puzzle-hunt elegance. That’s a tough target to hit sometimes, but one always worth shooting for.
I agree about the answer words. I complained about it to Wombat at the wrap-up (we both solved on-site) and he said that after eight years of MS Puzzlehunts, he’s gotten used to it.
By the way, I think it’s really awesome how much you include Lea in your puzzling activities.
OK, I admit it – our team is searching the web trying to find out what people thought of our hunt. Thank you for the kind words.
I agree that hunts are better with the intricate structure you speak of. Given the experience level of our team (puzzle solving vets, but over half having never written a puzzle), I think our choice was to put on an event in this style or to not put one on at all.
Maybe next time we get suckered into hosting…
Great hunt, thoroughly enjoyed. The whistlestops were a great innovation in particular, and I hope we see more like that in future hunts.
I had a different problem with the meta structure though. Normally every puzzle you solve gets you one piece of the picture for the meta. Can’t solve the meta yet? Go solve more puzzles in that round, dummy. Solved them all? Then you have what you need for the meta. But in this hunt, that was a complete black box. Do we already have all the evidence for this suspect? Nobody knows. Will solving the last Chicago puzzle unlock something? Nobody knows. How many metas are there? Nobody knows.
On the other hand, the evidence-style metas are probably a lot easier to rejigger for an online simulcast. If a puzzle can’t be converted to an online format, just drop it – its answer didn’t matter anyway. Maybe a good compromise solution is to just give more insight into the actual mechanics. Like, maybe the game tracks how many officers are out hunting for evidence, and how many are required to find the next bit, and each time you solve a puzzle (a ‘case’) that frees up an officer to search for evidence?
The whistle stops are great. I personally wish for more since I’m a fan of twitch puzzle games, though I understand, having a team stop everything for one puzzle for a few minutes too often could get annoying. This is the second time that they’ve shown up at MS, and I’m keeping my fingers crossed that they migrate to MIT.
I’m really curious to see how “Color Me Crazy”, a puzzle which had a section in the middle where you went to a MS conference room and waved your arms for a specially-programmed Kinect, was simulcast.
I was part of the Seattle hosting team.
The location puzzles (‘Color Me Crazy’, ‘Nickel Backtrack’, ‘Magical History Tour’, and I think also ‘Gathers No Moss’) were not part of the simulcast. For the purpose of the final meta, simulcast teams were assumed to have solved those puzzles. However, simulcast teams got one hour less for the overall event, which I think still favored remote teams over the local teams. Not a big deal though.
In response to Eric’s comments about the meta structure, our team wanted the design flexibility of being able to cut a puzzle from the event, or move a puzzle to a different round without needing to re-plan other puzzles around that removal. We had quite a lot of design changes over the past year :-)
The style of MIT’s themed metas is admirable, but not always the holy grail. This year it led to a fair amount of backsolving for our team: “We need a mushroom with a goofy name that doesn’t sound like a mushroom. Jack-o-lantern? Parasol? Send em in and see what happens!”
Asher’s criticism was something brought up to us during the hunt. We realized that yes, those connections should have been more explicitly visible, so in response, we sent out a mass email on Sunday describing which puzzles fed into which metas. In hindsight, we might have tried a themed way of displaying this information through the hunt software.