MITMH 2020: 2019 Oct

Careful: I'm pretty blasé about spoilers for this year-old hunt. If you haven't already solved the puzzles you wanted to solve, maybe don't keep reading.

Every so often, fearless leader Corey sent out a "State of the Hunt" email. Here are some excerpts from October's emails to let you know what the writing team was doing. (And to remind me, as I write this, what we were doing.)

Puzzles: Castle, Enchanted Forest, and Spaceopolis are pretty complete but we’ve swapped at least one puzzle after BTS1. BTS2 will test Big Top, Harvard Ballooniversity, Movie Lot, and Safari, so we’re focusing all the writing and testing effort there. Most of those puzzles are in good shape but many still need testing, and some still need heavy writing. Yesterdayland, Wizardy World, Wild West, and Water Park still have several unassigned answers, but they’re lower priority compared to the BTS2 puzzle lands.

Server: Core player functions are complete. Admin pages are being added now. Fastpasses and Hints will be added in two weeks. BTS2 will have some non-standard puzzle flows which will be added in two weeks.

Fabrication: Plans are underway to purchase or create costumes for Penny, for the Loonies, for Park hospitality hosts, and for Events staffers; see #costumes. … The final runaround will involve creating a really big pressed penny (the coin!) but those details are still pretty sketchy. … We have many puzzles with physical artifacts and we’ve reached the point now where we need to figure out—where should everything be warehoused until January? How do we avoid losing something? Garage space in Boston would be very welcome; contact me if you have some. … We’ve not yet started planning the build for sets for our intro or the final runaround.

Events: We have 5 events planned: Midway, Character breakfast, Line [the event that eventually became Build-a-Wand workshop started out as one part of a waiting-in-line event], Parade, and Fireworks. Midway has solid puzzles and fun but we’re adjusting scope in order to pull in budget. Breakfast tested well at BTS1; the open question is getting MIT’s approval for cooking. For both Midway and Breakfast we’re planning to have additional food and have that food sponsored by corporate donors. Development on Line is on hold about two weeks pending gathering on-site information. Parade is still early development. Fireworks has a good concept but needs revision before next testing. We’ll test Midway and at least one other event at BTS2. …

Testing. We have 25+ puzzles ready for testing. If you’re on Left Out, you must like puzzling, so testing our puzzles should be right up your alley. 8) Please follow this guide to get access to Puzzletron and some tips to finding testsolving Left Out friends.

BTS2. Our next big test solve is November 2-3 and we’ll host two teams: Left Out (“Right In”) and a non-Left Out team. If possible, please take part and RSVP on the Google Calendar invite.

BTS3. Our third and final big test solve will be December 7 - 8. We’ll test the final set of puzzles (two inner lands and two outer lands) that weekend; please RSVP on Google Calendar.

January travel. Now’s the time to make travel plans for the Hunt weekend! We need as many people as possible on-site in January, but even if you can’t be there, please sign up for remote Hunt Ops -- we have need for online-only staffers. Either way, please add your plans to this spreadsheet. You should book your hotel through the discounted room blocks - the same as players will - and that information is in the spreadsheet, too.

Story: The intro is written and stable. New this week: we’re confirmed to have Kresge for kickoff! The mid-game interactions are written but need feedback from unspoiled testers (target: BTS2). The end-game is still just a sketch but Todd and Ben are working on it. We have a lead on a room for the end-game, something thematic and basement-y in the Stata Center, but we’re still pondering locations. …

As of this week—all answers in our entire Hunt have been assigned. No more blanks!

We have art for all our puzzle lands now; see #art for lots of great examples. Some of these are just placeholders, until after BTS2 when Natalie plans to revise and add more details.

Costumes. Rachel is heading up our costume design and acquisition. She posted a request for materials and costumes; please take a look and respond on Slack if you have any of these things.

Big Test Solve 2 (BTS2) would be our biggest BTS. (You remember that a Big Test Solve was a weekend spent testsolving a few rounds of puzzles. BTS1 had covered the first few rounds. BTS2 covered another few rounds—but they included long rounds like Safari.)

I once again put on my Production Czar hat and assigned puzzles to Production Team members for HTMLization. For BTS1, I'd assigned one puzzle to each team member on a Friday and had puzzles back on that Monday. For the deluge of BTS2 puzzles, I had two puzzles per team member. Extrapolating from BTS1, I thought this would require two weekends. That's not how it went, though. Things went slower this time.

I think production folks had "sprinted" for BTS1: they had one puzzle to get done, so they did it. For BTS2, they saw that one sprint wouldn't be enough; so they paced themselves.

Anyhow, I shouldn't have estimated the team's HTMLification rate so confidently. We got through it OK, but I was nervous for a while there.

The Czars (Team Left Out's team leads) had a weekly video meeting. As Production Czar (HTMLification team lead), I typically didn't contribute much to these meetings beyond "HTML stuff is on track." But there were a couple of weeks around here where I instead said "HTML stuff is likely to fall behind for BTS2, but I think we can catch up for BTS3."

You might recall that back in June, we'd been dismayed to find out that someone else had reserved Kresge Hall, MIT's biggest auditorium, on Hunt opening day. And we'd scrambled to find another big space (and also yet another "spillover" big space) to hold our opening skit.

We got Kresge Auditorium after all. I don't remember how. Maybe someone in Puzzle Club found out who'd reserved it and asked them to switch or something?

BTS2 had a couple of puzzle rounds with puzzles that didn't fit the usual "answer is a word or phrase" pattern.

Tech Czar Doug was pretty busy getting our hunt web server to work with these puzzles. Rich Bragg adapted the (open source) Twitter web emoji picker to help teams enter emoji answers. But the web server's answer checker had to deal with special cases.

E.g., the server could display a custom message for special wrong answers. E.g., maybe a puzzle yields the message DESSERT COOKBOOK, but that message isn't the answer—maybe entering BAKERY instead displays the answer "BRING US FOOD, YO: BRING COOKIES TO HQ". And if a team brings cookies to the running team, they get the puzzle's real answer, SUPPLEMENTAL HARDWARE GUIDE. Some special wrong answers display special messages: that's pretty normal.

But for multi-answer Safari puzzles, we wanted some special correct answers to display messages, too. E.g., in the Rhino two-answer puzzle, the team assembled a structure to get the first answer. To get the second answer, they would take a picture of the first-answer structure and mail that picture to Guest Services (HQ). So we wanted the web server, when it saw the first answer, to display a message asking the team for a picture. The web server already displayed a message for correct answers, a congratulatory one. How could we get teams to notice that this message was both a congratulation and a picture request?

Emoji answers had other problems. Consider the ♥. Teams that completed the scavenger hunt knew their answer was a ♥ shape, but there are a few ♥-shaped emojis. Which of 💕💓💖💗💘💝💞💟🧡💛💚💙💜🖤❤♥❤💓💖💗 were they supposed to choose?

For BTS2, we used the web server's special-message-for-special-wrong-answer function: if a team guessed 💗, the server showed a message like "💗: Close! You want 💓". This was kind of janky: if the right answer and the wrong guess looked nearly identical, it wasn't super-clear what was happening.

For the real hunt in January, Doug improved this. In January, the server understood that there might be alternative correct answers. So if a team guessed 💗, the server just said "Correct!" and displayed the correct heart. That worked pretty well, but it took us a while to figure out what would work well.

Some bad communication between collaborators led to a dicey situation: I was in a group that had agreed to provide a puzzle that solved to a particular answer phrase. But in October, darned late in the year, we handed this answer phrase back to Yar, saying we couldn't write a puzzle for it. (This surprised me: I was still designing a puzzle for this answer. Did I mention there was bad communication?)

Fortunately, Martin Reinfried picked up this "hot potato" and quickly came up with the Pirate Ship puzzle. This puzzle was surprisingly tricky to HTMLify, but I tried to do a good job nonetheless; this puzzle saved our bacon.

One of our puzzles, Operation, was not for the squeamish: part of the puzzle involved recognizing medical operations by viewing videos of these operations. A doctor could probably handle it; most Left Out folks weren't doctors.

Enough of our testsolvers could endure the videos. Here "enough" means a few: we got a few testsolves on early revisions; we got a couple of "clean" testsolves on the final version. But it was tough to get that last couple of testsolves. We were running out of not-easily-grossed-out people who hadn't already seen the puzzle.

(It turns out I'm easily grossed out, at least when it comes to medical operation videos. I remember doing some HTMLification work on this puzzle, doing some JS work so it would display the videos despite some URL-scrambling things our hunt web server did. This work was made trickier because I only ever looked at those videos by holding up my hand over the screen so I only had to watch the parts that showed between my fingers.)

We didn't consider at the time, but obvious in hindsight: Our team of ~60 people only found a handful of willing testsolvers. During the real hunt, small teams might not have anyone willing to look at operation videos.

Had this puzzle fallen late in the hunt, then probably only larger teams would have encountered it. But it was early in the hunt and many teams bumped into it.

During the second half of October, HTMLifying puzzles was an eight-hour-a-day job for me. It didn't consume my life, but it did consume a sizable fraction.

I didn't take detailed notes during this time. So now, as I try to write about interesting things that happen, I don't have much to say. But this was a busy time.

The busy-ness paid off, though: we caught up on HTMLification.

Art Czar Natalie got the team an account at the clip art site. This meant that if I wanted, say, a cute picture of a cow to decorate a Minotaur-themed puzzle, I could get one from the site. I wouldn't have to beg art from one of Left Out's few artistic folks.

Nov [>]

comment? | | home |