UPDATE: The good people of Escape From St. Louis are running DASH 9! Yay!
Hello, friends of DASH St. Louis –
With all of your help, the DASH puzzle hunt has had some very successful years here in St. Louis. We’d love for that to continue, but Michael and I cannot run it this year due to scheduling conflicts. I’m reaching out to you – former GC, players, or anyone who might be interested in helping. For DASH 9 to happen in St. Louis, someone needs to step up to run it.
Finding locations and arranging permission.
Finding and scheduling volunteers
Setting up a ticket selling mechanism
Communicating with local teams
Regular internet meetings with national organizers
Budget management (generally this is budget-neutral, usually you get lunch and a shirt out of it)
Puzzle materials handling
Route and puzzle testing
In particular, the puzzles themselves are already written – you don’t need to do that.
Michael and I will help out as much as possible with the setup and planning, especially with regards to advertising and team/volunteer contacts.
The original date announcement was May 6, but there is currently a poll with options to change the date to any of April 22, 29, May 6, 13, 20, so the sooner you respond the more chance you have of influencing the date.
If you’d be willing to champion the DASH St. Louis puzzle hunt this year, many will be grateful!
If you’re looking for more puzzle fun in St. Louis:
Nir Chezrony runs escape rooms in the U City loop, under the banner of Enigma Productions. His next room is coming up in March 2015.
Michael Goldwasser and I will be running the St. Louis site of the international day-long puzzle hunt DASH 7 on May 30, 2015. Contact me if you want more information, or have a local team that would like to sign up.
I’ve authored a little two-puzzle series that’s location specific to St. Louis. It’s buried deep inside Larry Hosken’s online puzzle trail Octothorpean, but that just gives you an excuse to play his excellent games. Or maybe, to visit us here in the Lou.
Finally, show some love for St. Louis’ best known professional puzzler, Patrick Blindauer. He brought us DASH 4 and 5, and invited me to Nir’s escape room last summer, so Sensation really started with him.
Sensation ran for two months, with 36 groups and 259 people playing. 23 groups solved it, for a 64% escape rate. The quickest exit was Greg M’s group, getting out with all puzzles solved, no hints, and 35 minutes remaining. The next fastest groups escaped with 28:15, 26:00, 22:40, and 20:00 left on the clock. 3 groups escaped with 15-20 minutes left, five with 10-13 minutes left, three in the 5-10 range, and seven groups had the excitement of an escape with just a minute or two left. The day after Thanksgiving, my family cut it the closest, brute forcing the fifth symbol on the lock with just 2 seconds remaining. Thirteen groups out of 36 didn’t make it out in the one hour limit.
Every group got the five lock box open. Based on when the mouse first moved, the median time to open the box and take out the Ouija board was 33 minutes. The quickest time to spirit contact was Team Chai’s 13 minutes, and they escaped with 18 minutes remaining – a time which held the speed record for half of the run. Six groups never solved the flashlight sight key puzzle, five never followed their noses to the air freshener, three didn’t find the touch key in the ice cube, two couldn’t get the chocolate lollypop, and only one group missed the creepy little girl in the chair, which was my fault because her audio didn’t play for them.
Spoiler alert: Don’t read further if you want to solve the sight puzzle.
Of the spirits, the easiest was Galileo, contacted by 28 groups. Mozart was #2 with 23 contacts, then Julia Child and Chanel with 19 each and finally King Midas with only 16 groups asking for his assistance. Also telling, Chanel and King Midas were never the first spirit called.
Very few groups got to hear Houdini congratulate them, since that required contacting all five spirits before solving the puzzle. If you solved sensation and want to hear your kudos from Houdini, read the theme and metas post.
Sensation Thanks & Credits
Thanks to everyone who played – I hope you had as much fun solving it as I did building and running it. I never got tired of listening to groups laugh, scream, struggle, and cheer.
Thanks especially to my colleague Michael Goldwasser, who helped design and test puzzles, did a bunch of setups and breakdowns, ran the show for groups when I was busy, and volunteered his daughter for a long morning of alpha testing. Dave Letscher had the crucial idea to make the five lock box open with four keys. My wife and daughter were subjected to much unpleasant experimentation involving sight, taste, touch, and especially smell. Laura, Bobbie, and Tammy tolerated my 36 room reservation requests and the mess in Ritter 111.
My spirited voice actors: Mark Batshaw, Marshall Cohen, John Cross, Krista Denton, Stuart Oberheu, Stanley Pitchford, Carla Scissors-Cohen, and Joan Tracy. You guys were great, and I’m sorry none of you could play the room!
Finally, thanks to 10 years of students in SLU’s math/CS club. This one was for you.
The cooler for the taste puzzle had a taste sticker on the outside. On the inside, players saw this:
Not all groups noticed the key in the ice cube when they took the soda out of the cooler. Some came back to the cooler later and saw it, some got a hint to check again, and some just never found this key.
Once the players found the key, they had to get it out of the ice. Waiting for it to melt on its own took too long. The intended method was for someone to hold it in their hands (or mouth) until it melted, which took only a minute or two. Crushing the cube was also possible, although I believe this led to bad karma and difficulty reaching spirits. Either way, you had to touch it.
The “touch spirit” puzzle was a wooden box with a leather covered hole, sitting on a table by itself. The first step to solving this puzzle was to reach into the hole. Inside the box were six squishy puffer balls, intended to feel pretty gross and give a jolt of icky surprise. Players had to take the six balls out of the box, and notice they were numbered 1-6.
Each ball was slit open, and contained some plastic tiles with letters. Arranging the letter tiles into words, and then putting the words in order gave the message “touch back of my golden sides”:
Most groups got to this point without much trouble, although GOLDEN also anagrams to LONGED and DONGLE. Groups then had to notice that the two long sides of the box were painted gold.
The next step was for someone to stick their hand back into the box and notice that the insides (“backs”) of the two golden sides were covered in bumps and divots. There was Braille writing on the inside of the box, and the plastic tiles were Braille training tiles I’d purchased from an online store for the blind (free shipping via USPS!). At this point, players had to read the Braille inside the box with their sense of touch, using the tiles to translate the letters and spell the answer to the puzzle: KING MIDAS.
Of all the spirits, King Midas was called on the least, so by that measure this puzzle was the most difficult in the room. Some groups got stuck after reading the tile message, some just couldn’t handle the Braille. A few thought the answer was King Sadim or Bing Fidas. Too bad, because I thought John Cross was excellent as King Midas, telling players their lock combination began with O.
The Golden Touch
I’ve had to solve quite a few puzzles where the answer needs to be decoded from Braille. These puzzles all asked the solver to look at a pattern of dots and convert using a key. When I had the theme for Sensation, I realized I could make a puzzle that required solvers to read Braille the way it was intended: by touching it.
The “feel box” puzzle was the idea that convinced me I needed to actually build Sensation, and of all the puzzles I think it’s the most elegant: the decoding key is built in to the tiles that give the instructions, and the answer, King Midas, is hinted at by both the golden sides of the box and the “golden” and “touch” words of the instructions.
Here’s a view of it during construction, showing the bumps (thumbtacks) and depressions of the word “king”, with “Midas” not visible on the facing side. Real Braille just has the bumps, but that would have been beyond most players’ ability to read.
The hole was covered with faux-leather on the outside and the sleeve of a flannel shirt on the inside to make it nearly impossible to see inside the box. I painted the long sides of the box with gold acrylic paint, but the first few groups had a hard time recognizing it as gold, so I had to repaint with some metallic gold spray paint.
My original idea for hiding a key with smell was to put the key in a plug-in air freshener. Groups would smell it, look for it, and find it. Simple. But air fresheners can really stink up a room, and the math conference room is fully enclosed, no windows. Michael and I were worried that it would make the experience unpleasant for groups. We were also worried it would make the other smell puzzle impossible to solve. And we worried that our colleagues with actual job related meetings in the room would be upset.
If I’d been on the ball, I’d have gotten ahold of an Oscar Meyer bacon smell generator, but I was too late. Instead, Michael and I furiously kicked around puzzle ideas in the final week of planning, imagining some kind of scent trail to follow, or even some sort of physical device. There must be a way to make a really good smell-based puzzle that hides a key, but we didn’t hit on it.
Instead, I went back to the original plan and hid the key in an air freshener, plugged into an electrical socket, modestly concealed behind a table:
While this technically was a smell based puzzle, I think only one group actually said “hey, what’s that smell?” and looked for it. Most groups just happened on it. Some groups never found it, and I have a hard time believing they never noticed the sickly sweet smell of $6 AirWick Apple Cinnamon Scent. Luckily, the ventilation in the room was good enough that the other smell puzzle wasn’t ruined by this one.
It’s another sense memory that will always bring Sensation back to my mind. I had to keep the thing on the shelf in my office so it wouldn’t leak into the suitcase, and all semester my office smelled vaguely of something not quite entirely unlike apple pie.
From the beginning, the plan for the “smell spirit” was always to use scratch’n’sniff stickers for some sort of matching problem. By the end, I spent way more money on this puzzle that any other, and quite a bit of time as well.
The puzzle consisted of twelve plastic cards with segments of 7-segment LED printed on them. Each card also had an orange sticker in one corner. The cards came stacked, shuffled, in the middle ‘nose’ slot of a tray (stolen from my Monopoly game). The other six slots were decorated with pictures of smelly things: Mint, banana, chocolate, cinnamon, ginger, rose.
Players had to smell the orange stickers and sort the cards into the appropriate slots of the tray. There were two cards with each scent, and each pair of cards combined to form one letter of the answer, ‘CHANEL’.
I mean, people don’t smell very well. Yeah. People aren’t good at telling the difference between different smells, and this made the Chanel puzzle hard to design and hard to solve.
My original plan was to make my own scratch-n-sniff stickers, using ordinary white stickers and essential oils. I bought a couple of essential oils, some soap fragrance oil, and some air freshener oil. I put drops of these oils on post it notes and started testing. The scents faded quickly (a couple of days) and were hard to distinguish. Like, it was hard to tell cedar from flowery soap from sugar cookie. On top of that, the scented oils were expensive, running upwards of $10 for a small bottle.
Supposedly, scratch-n-sniff stickers last so long because the scented oil is encapsulated in tiny wax globules that burst and release their scents when scratched. Since I was having no luck making my own, I decided to find some that were commercially made. Unfortunately, scratch-n-sniff stickers always have a picture of what they smell like, printed right on the sticker. It seems impossible to get blank (or at least identically marked) smell stickers without having them custom made. The first couple places I contacted told me that a custom run of blank stickers would cost me thousands of dollars, so somehow I was happy when I found a place in New York that could hook me up with six sheets of differently scented blank stickers for only $125, with a huge lists of scents for me to choose from.
Would I have liked bigger stickers? Yes. Would I have liked it if the different scents arrived at my house looking identical instead of being different shades of tan? Oh, yes. Would I have liked to pay less than $125 to get this puzzle working? Hell, yes.
I had to color the stickers with orange magic marker to make them indistinguishable. I had to replace the stickers a few times over the run of the room. And the puzzle was still hard because the small dots didn’t have strong smells. But groups could solve it, especially once they knew they were looking for a spirit name.
In the end, I was happy with the puzzle. And I’m in possession of six mostly complete sheets of 1″ round blank scented stickers, if you’d like to buy some!
One prominent object in the room was a drink cooler, sitting on the main conference table. This cooler has two compartments – the main drink area and a smaller compartment inside the lid. The main compartment was open, the small top compartment had a metal latch and a 40-digit dial padlock holding it closed.
Inside the cooler, groups found six clear glass bottles containing brown soda. Each bottle had a number on its side. There were logos for six types of brown soda taped to the sides of the cooler, in pairs, with each pair separated by a dash. I also provided a stack of dixie cups and a roll of paper towels.
Players needed to taste the sodas, identify them, arrange the bottles in front of the logo stickers, and then read off the combination for the padlock from the numbers on the bottles.
When you buy a cheap padlock, you don’t get to choose your combination. I wasn’t thrilled with what I got, 20-22-36, because it meant that three sodas would all get numbered with 2. I numbered root beer with 2, Vanilla Coke with 0, Coke and Dr. Pepper with 2’s, Pepsi with 3, and Diet Coke with 6. As expected, Dr. Pepper and root beer were easy to identify. People had a tough time with Vanilla Coke. The most difficult task, to my amusement, was to distinguish Coke from Pepsi.
Once the players identified the sodas, and figured out that they gave a lock combination and not some cruel subtraction problem, they could open the upper compartment. Inside, they found this chocolate lollypop:
Usually, this got laughs. It’s a great moment in a puzzle when you think you’ve solved it, and then the big reveal seems completely unhelpful. After the laughs, it didn’t take groups very long to decide to open the lollypop and eat it. Or break it. Not eating the lollypop was one of my private little indications that your group wasn’t fully engaged. If you’re doing a puzzle about taste, and your are given a chocolate lollypop, you should actually eat the thing.
Embedded inside the chocolate was the “taste” key, at that point usually covered in a disgusting mess of licked chocolate residue.
The “taste spirit” puzzle was one of the hardest in the room, and as my friend Patrick Blindauer described it, “scary”. This puzzle was a tray of red jellybeans, sitting alone on a table:
Under each jellybean was a letter. The solution was simple, but you only got one try: Eat the jellybeans. Notice that some are cherry flavored and a select few are cinnamon. Mark the letters which had cinnamon jellybeans. Observe that the marked letters reading left to right spell “JULIA CHILD”.
Very few groups took this approach. The majority thought that each vertical column would have one letter to fit in the square underneath the column. Why did they think that? There are nine columns of beans, and ten squares. The columns don’t line up with the squares. Some columns have no cinnamon beans, while others have two. Yet still, the vast majority of groups ended up with something like CJHUI LILAD written in the boxes, and then they had to anagram it out.
The taste puzzles were by far the biggest pain in the ass for Michael and I to deal with. They are consumable (duh), and need to be replaced. They make a mess, and we had to be careful about sanitation so people wouldn’t get sick.
For the soda quiz, I used three sets of six glass soda water bottles with plastic caps. They needed to be filled before each group played. I had cans of soda, and each can was enough to fill two of the soda bottles, so I would make two sets at a time. Some groups got flat soda.
Then there was the cleanup. Luckily, nobody spilled soda. But there was no way to be know if a group poured into the cups or drank from the bottles, so I had to wash all six bottles after every group played. And there’s no dishwasher in the office where I work.
All in all, between filling and washing bottles, the soda puzzle took about 10 minutes of work per group that played the room. This leads to the most important rule of puzzle construction:
Don’t design a puzzle that requires you to wash the dishes afterwards.
Making the chocolate lollypops was fun and easy. I had eight copies of the key, and a mold that made seven pops. It took about 10 minutes to make a batch. I used cocoa flavor candy melts instead of real chocolate, because they are much easier to work with. Put them in a ziploc bag, give it a couple of minutes in the microwave, cut off the corner, and squeeze out the chocolate. Unfortunately, the flavor of candy melts is a sad mockery of real chocolate. If you didn’t like my choco pops, you can have your money back.
My jellybean tray was made of a clear acrylic 8.5×11″ sign holder and the clear plastic lid from a set of watercolors, joined with superglue. Each jellybean had its own little holding dish, while the sheet with letters slipped into the sign holder and was visible through the two layers of plastic.
It took quite a bit of shopping to come up with this design, but it paid off in the end, because I could remove the paper sheet. Groups were nervous (or maybe not hungry enough) to eat all the beans, so they often nibbled on them and put them back on the tray, leaving little red pools of spit behind. It was good that I could take out the paper and then easily and thoroughly wash this puzzle.
The taste puzzles were responsible for most of the per-group cost of running the room. Each group needed 6 half-cans of soda, 36 jellybeans, and a chocolate lollypop, which probably came to about $3 per group. In the end, with requests still coming in, I “sold out” sensation at 36 teams because I was out of lollypop sticks and bags. Though really that was just an excuse – mostly I just couldn’t give up any more of my time.
The “sight key” puzzle for Sensation centered around a crypto-flashlight: A black light flashlight that only turns on when the players set its knobs to the correct combination.
The flashlight started out as a cheap lantern flashlight. I took the business end of a black light LED flashlight and superglued it into the cowling of the lantern light:
This also required drilling some holes in the cowling to pass power wires through from the flashlight body.
Next up, I wired six 7-position rotary switches in series:
I printed colored indicators and attached them to the flashlight sides with clear tape, then drilled holes in the sidewalls:
Installing the switches into the lantern was the hardest part. Tight workspace, and they had to be secure enough that they wouldn’t turn with the knobs. I had to heat the metal plate on each switch with my soldering iron so it would sink into the flashlight plastic and stay put.
Finally, I wired the black light flashlight’s 3xAAA battery pack to the switches, the main flashlight switch, and the LEDs. I tied the battery pack to the flashlight bottom with a loop of wire.
The knobs I used came in red, yellow, green, and blue. For purple, I covered a red knob with purple nail polish. For orange, I used an orange sharpie on a yellow knob. Here’s the finished flashlight , from the green/blue/purple side:
All in all, the flashlight cost around $50 for materials and took me about 4 hours to build.
A few words on building stuff that hundreds of people will play with
When I set out to build Sensation, I wasn’t expecting it to be so popular. I built stuff to be strong, but some things weren’t strong enough.
The flashlight broke at one point, because a couple of groups opened it up, the battery pack came loose, and its connecting wire was severed. In fact, one group actually solved the flashlight by looking inside and seeing how to set the switches to make the correct electrical connections. After the wire broke, I had to repair it, find a tougher way to attach the batteries (velcro!), and then I duct taped the thing closed to prevent future groups from opening it again.
The worst casualty of the room was the pull down projection screen. Sort of a rollershade contraption, it began to lose its spring after a dozen groups, and after 20 groups it just broke. Every group that pulled it down thought they’d broken it, and it wouldn’t roll back up again. Sorry, groups! I really liked the hiding place, and was just a bit too lazy at that point to find another solution. I had to do a quick repair to the screen (with pliers) after every single group.
The puzzle leading to the “sight” key was set out in plain view on a table by itself. It consisted of a colorful word search and a crypto-flashlight with six knobs, colored red, orange, yellow, green, blue, and purple. Each knob had an indicator allowing it be set to one of the same six colors. I’ll say a bit more about the crypto-flashlight in my next post.
A quick look at the word search shows that there are many hidden color names. A closer look reveals that some of the color names are not in their own color. The idea was that this word search was a “Stroop test“, a classic psychology experiment. I don’t think I really captured the challenge of the Stroop test with this word search – it wasn’t particularly hard for players to find the word names written in the wrong color.
When the group found the six incorrectly colored names, they set the named knob on the flashlight to the word’s color. The switch on the flashlight turned on a black light.
Invisible ink clues were written on most paper surfaces in the room. On the word search sheet itself were the instructions “turn off the lights, look around the room”. Invisible writing in a few other places sent the group to look above the whiteboard, where an innocuous looking sign revealed, under black light, to “pull down the screen”. The whiteboard in this conference room has a very well concealed projection screen – subtle enough that only one group pulled it down without solving the flashlight. When groups pulled down the screen, usually in the dark, the “sight” key fell out with a pleasant chime as it banged off the whiteboard marker tray.
Since Sensation’s puzzles were based on the five senses, you can’t try most of them at home. This one, you can, if you like. One wall of the room featured seven “artworks” and an answer key below them. Here’s a PDF of the sight puzzle, and it’s installation on the wall:
That’s Dr. Case smiling down on the puzzle. Dr. Case was a longtime member of SLU’s math department, and one of three portraits that hang in our conference room. They had nothing to do with the puzzle room, but at least Dr. Case seemed to approve.
SPOILER ALERT: Don’t keep reading if you actually want to solve the sight puzzle!
Each artwork is an optical illusion, concealing some text. They read “giraffe”, “liar”, “love”, “little”, “hello”, “of course”, “read your book case”. When you fill these in the answer grid, you get the sight spirit: “galileo”. All of these pictures were found on the web except the horse, which is my adaptation of a rather unpleasant puzzle you’d probably rather not know about. The actual Chinese writing above the horse is supposed to say “turn your head”, as a hint, but my Chinese players tell me that Google translate completely failed on that one.
Most groups solved this puzzle before they got the flashlight working, but each artwork did have an invisible ink hint to help see the illusion. Illusion 5 is by far the hardest, if you haven’t seen it before. The invisible hint was to hold it at an angle and look from the bottom.
This puzzle took me many iterations to get right, going back and forth with different illusions and different answers. Every time I tested it, I had to make it easier. I’m happy with the final result, but it did turn out to be the easiest puzzle in the room. On the other hand, it was really good to have an easy “spirit” name, so that groups would have something ready when the Ouija board came out. The other spirit names were much harder to solve. In fact, half of the groups contacted Galileo as their first spirit.
Here’s Stuart Oberheu as Galileo, telling players their combination has a U in it. There was only one U anywhere on the main combination lock.
The key for the “hearing” puzzles was hidden in the case of an iPod touch. The iPod touch was, in turn, extremely well hidden in the hollow of the leg of an old metal desk chair that generally lives in the conference room I used for sensation. Unless you flipped the chair over (which I think one group did), you were not going to find the iPod just by looking for it.
However, the iPod had a recording playing on it, which I would start as the last step of room setup, after making sure the group was present and ready to go. To make the recording, I had my 8 year old daughter record a bunch of attention-getting things. Most of them came naturally to her, others I suggested.
Here are her lines, some of which had a few versions:
Hey, look at me! Look at me!
You guys still haven’t figured this out?
Aaah, I’m so frustrated
Would you listen to me?
Can you please play with me some more?
Are you guys having fun?
Let’s have some more fun!
Hey, aren’t you going to play with me?
Don’t forget about me over here.
Hey, I’m over here
Look at the gun show over here! (this is a Night at the Museum quote)
and of course, the crowd favorite, “Bubble bubble snap snap cherry pie!” which only an 8 year old could have come up with. I remixed the approximately 90 seconds of voice samples into a 70 minute long track. After a ten minute silence at the beginning, the volume starts low and short phrases start playing around once every couple of minutes. As the track continues, the volume ramps up and the frequency of sounds goes up, until it’s almost impossible to ignore by around minute 30, which would typically be 20 minutes after the group started the room. Every group eventually heard the iPod. Some took quite a while to decide to look for it, and usually got irritated with her attention getting behavior. Just like being a parent.
My daughter was pleased to hear that nearly every group referred to her as “creepy” at some point. The disembodied little girl voice in the chair was definitely a highlight of the room – most groups were startled or “freaked out” by it. It was also the one puzzle that was harder for larger groups, since they generally made more noise and drowned out the iPod sounds for longer.
Though the iPod was hidden, there was one obvious hearing puzzle in the room. Michael provided an old CD player boom box, on which we slapped a green “nose” sticker. The boom box was plugged in, turned on, and had it’s CD compartment open with an unmarked CD inside.
This CD had four tracks:
The Beatles, “Eight Days A Week”
The Clash, “Magnificent Seven”
Bob Marley, “Three Little Birds”
Groups had to listen to the CD, realize the songs were all numbers, and then identify the numbers. Not so hard for the first two tracks, although the majority of groups thought U2’s song was called “One Love”. The Clash proved to be the most challenging. The first words of the song are the shouted title, “magnificent seven”, but Mick Jones is pretty hard to understand.
Often, playing the CD made it harder to hear the iPod, which was kind of the idea, and so groups tended to have these song titles before they knew what to do with them.
Once the iPod comes out of the chair, it’s pretty clear that you need a four digit code to unlock it, and these songs gave it: 1873. They also gave a jolly soundtrack to the room. I’ll always reminisce fondly of “Sensation” any time I hear “Three Little Birds”…. every little thing’s gonna be alright. However, if I never hear “Magnificent Seven” again, I’ll be fine with that.
I removed everything I could from the iPod without jailbreaking it, which seemed like too much trouble. The rest (Apple’s unremovable apps) I stuck into three folders labeled “ignore” “these” “folders”, and groups did ignore them. That left the music player as the only thing of interest, and it only had two songs: the attention grab, and a second track, “Paragon of sound” by Samuel Morse. The Morse track was a morse code recording, and the cover art was a morse code key. Groups had to play the track, listen to the morse code, and decode it. This was actually challenging, especially when under heavy time pressure. When they did decode it, they got the sound spirit, “MOZART”.
Running “MOZART” through the Ouija board gave ‘D’, the middle letter of the lock. Here’s pre-eminent physician Mark Batshaw as Mozart.
These puzzles held to a couple of standards I wanted to achieve: One, groups had to actually use the appropriate sense to solve them. Two, there were no instructions telling groups what to do, and three, I avoided using a code sheet, which I was at first tempted to provide. I was pretty happy that the album cover art could display the Morse code key, and keep the puzzle self-contained.
The “Sensation” puzzle room was called “Sensation” because the theme was the five senses. For each sense, there were two main goals: To find a key, and to learn the name of a spirit guide associated to that sense. The sense puzzles were clearly coded with sticker icons:
The box with locks
The first metapuzzle was a large pentagonal box with five locks on it. It was the largest and most prominent thing in the room, and many groups assumed that once they got it open they would be free. In fact, it was just step one of two towards unlocking the room. Every group did get it open, usually within the first half-hour.
The box is a pentagon, about 22″ across at its widest point, and 4″ deep. I made it with a plywood base, 1×4 pine sides, and a hardboard top. There are five 1″ dowels embedded in the base of the box, and they stick up through holes in the lid. Once the lid is on, 10 wires run across the box in a pentagram pattern. Each wire has a small loop soldered at each end. The looped ends of four wires fit through horizontal holes in each post, and are then secured with a lock.
Besides looking really cool, the point of this complicated design is that unlocking any four of the five locks will open the box. I knew from the start that I wanted this feature, since it gave teams a chance to move along in the room even if they hadn’t found all five of the keys. The internet didn’t turn up much except for some cool park and ranch gates which are built for access by unlocking any one of multiple locks. I came up with some ideas for 2-out-of-3 lock boxes, but I’ve got to give credit for the pentagram wiring idea to my colleague Dave Letscher. Mathematically, this is the complete graph on five vertices, or K5, and the idea would generalize to any n-1 out of n locks, although the number of wires quickly becomes prohibitive. Dave didn’t know what I was locking up, but as soon as he drew K5 on my board, I was sold.
The Sixth Sense
Unlocking the box, players find themselves in possession of a Ouija board, moving from the five senses of the keys into the metaphysical ‘sixth sense’ to finish the room. Reactions ranged from disappointed (didn’t win yet) to freaked out – many of the best screams came at this point. One thing I learned about Ouija: my people call it “Whee-cha”, with a strong J sound, but apparently the norm in Missouri is “Whee-gee”. I did not know until the room was fully built that the mediocre horror film Ouija was coming out this Fall. I didn’t see the movie, but thanks, guys, for making my room seem hip and current.
I used a run-of-the-mill Ouija game from Hasbro. Inside the box, players found the board itself, a laminated sheet of instructions, and a wooden planchette (the pointer device):
I custom built the planchette around and old optical bluetooth mouse I owned. The wood came from an old cigar box I had left over from Benjamin’s 7th birthday party (teaser: more on this in a later post!). The mouse was connected to the laptop which was supposedly only running the countdown timer, but of course it was handling all ‘spirit communication’ as well. When players discovered spirit names, they would spell them on the Ouija board using the planchette/mouse, and the computer would then play a pre-recorded audio track of the ‘spirit’ responding out of a hidden speaker in the ceiling. This was also good for some scares.
Getting spirits to respond was one of the more fragile puzzles in the room. Groups had trouble when they moved the board to the far side of the room, or twisted the planchette when sliding from letter to letter, or spelled too fast or too slow. Nevertheless, all but one group did get ‘spirit communication’ working adequately.
Each spirit (one per sense) would tell the players one letter of the final combination to exit the room. The exit lock was a Master Password Plus Combo Lock. This lock has five wheels with letters and numbers on them, and is customizable for whatever combination you want. Or so they say – in fact, your choices are very limited. The lock comes with eight wheels, and you use any five of them for the combination. But there are only four wheels with letters – the rest are numbers and symbols. In the end, I scratched the letters ‘H’ and ‘1’ into the lock, and used the combination OUD1N to spell “HOUD1N1”, the solution to the room.
I wanted groups to be able to brute-force the lock with, say, four of the five spirits contacted. Again, not having to solve every single puzzle seems like a nice way to keep groups from getting frustrated. But I didn’t want it to be too easy to guess, so the combo was shifted up one slot from the obvious row. In the end, though, the H and 1 gave people enough of a hint that quite a few groups were able to guess Houdini with help from only three or even two of the spirits. Groups that did contact all five spirits received a little bonus, though – Houdini himself would speak, tell the players to shift the combination up one line, and congratulate them on their work. Here’s Marshall Cohen as Harry Houdini.
If you’re going to contact spirits with a Ouija board, there’s really nobody better to involve than Harry Houdini, famed escape artist, spiritualist debunker, and target of annual Halloween séances. Because I have kids who trick-or-treat, I was unable to give a group the opportunity to play Sensation on Halloween. That would have been really cool.
Most groups, I think, understood that it was some sort of computer program tracking the mouse, but there were plenty that assumed I was triggering the audio tracks based on what I heard over the baby monitor. That would have been a total failure – many groups chanted the letters, but others spelled in silence.
I wrote the program in Python, using the PyGame library, which had the graphics, mouse, and sound utilities baked in. The one-hour countdown timer was straightforward, although it took some care to get it to run without cranking the fan on my ancient MacBook.
So, how do you determine when someone has traced out the name of a spirit using a mouse on a Ouija board? I had many ideas, involving complicated AI, but in the end I tried something simple and it worked well enough.
Mouse movements arrive in tiny little increments. My first step was to add these together to make movement vectors, separated by ‘pauses’ where the mouse stayed still. This gave a stream of vectors coming in to the spirit recognizer. I converted these vectors to a stream of distances separated by turning angles between vectors. Actually, I used the cosine of the turning angle, because that’s easy to compute using dot product (Calc III students, take note!). The nice thing about the distance/angle representation is that the mouse orientation doesn’t matter, as long as it doesn’t twist it too much when moving.
For each spirit, I recorded the sequence of distances/angles that result when spelling the name. As the distance/angle stream comes in, the most recent data is matched against each spirit name using least squares. I tweaked it a bit to get good response and cut down false positives, but it worked really well almost immediately. It was accurate enough that spelling a name with one letter off by one position would usually fail.
The code for the countdown timer/spirit handler is on my github and is GPL, if you’d like to peruse it or adapt it for your own uses.
The only hitch was a nasty little surprise – when I switched from my testing photocopy of a Ouija board to the actual glossy Ouija board that came in the box, the mouse failed to track accurately. I had to glue the photocopy to a piece of hardboard and use that for the room, instead of my authentic Hasbro toy.
At the end of last August, some friends and I did “Trapped: Volume 3”, a St. Louis room escape done by Nir Chezrony’s Enigma Productions. We failed to escape from his room, but by the end of our postmortem at the Moonrise rooftop terrace, I was thoroughly inspired to try to make one myself. Before the weekend was out, I’d come up with a theme and many of the puzzle concepts, and after that I couldn’t let it go.
I decided early on that the puzzle was a gift to SLU students, as a sort of farewell to the Math/CS club, which I’ve been sponsoring for a decade. When it ran, I opened it to anyone at SLU, and about half the groups that tried the room were staff or faculty. Apparently SLU (and probably most universities) have a lot of fun things for students, but the staff are neglected, so it was a pleasant surprise that so many non-student groups signed up. It certainly helped that I didn’t charge – the room cost me some money up front, but since it was on campus I didn’t need to rent the space or deal with any liability issues that would probably jack up the cost of a privately run enterprise.
I used the Math/CS department’s conference room for the puzzle, but it was still being used for other purposes during the semester. So I needed a room that I could set up and break down fairly quickly, and store in the relatively small amount of space in my office across the hall. Originally, I wanted to keep setup to 15 minutes and breakdown to 5, but in the end both took about a half hour, so each group really took around 2 hours of my time – although I could do some small amount of work while groups were solving and I was in my office.
The general format of the puzzle was copied directly from Nir’s “Trapped” – groups left their possessions in my office, then went into the room with no real instructions other than to open the lock on the door. Each group had 1 hour to escape, with a countdown timer running on a laptop in the room. I (or Michael) would listen to the group over a baby monitor, and occasionally write hints on index cards and shoot them under the door when teams needed help.
Design and build for the puzzles took all of September and most of October. Michael Goldwasser helped design and test some puzzles, and then helped run the show. We had a successful beta test on October 21, and then ran the room until December 16. I advertised it with posters on campus and in the dorms, and also put a notice in the “Newslink”, a daily email sent to everyone at SLU. Amusingly, nearly all the teams that signed up found out about it through the Newslink, and my postering efforts were largely wasted. By the time word-of-mouth started bringing in more groups, we were “sold out” – I capped it at 36 because some supplies were running out, not the least of which was my available time.
All told, 36 groups and 259 people did the room, 23 groups escaping and 13 still stuck after their time expired, a 64% escape rate. In the next few posts, I’ll run through the theme of the room and the various puzzles, and finish up with some kind of result summary.