Deanna ([personal profile] dr4b) wrote2017-03-27 12:07 am

Galactic Puzzle Hunt

Maybe my LJ/DW/whatever goal for the year should be to write at least something about all of the puzzle hunts I do this year.

So last week was the Galactic Puzzle Hunt, an all-remote week-long thing, "in the style of SUMS or mezzacotta", which are, of course, hunts I had heard of, but had never done. The idea is that there's some number of puzzles -- in GPH, there were 30 -- released over the course of some number of days -- in GPH, it was over 6 days, 5 puzzles a day -- and there are complicated scoring systems about when you solve the puzzles to determine a winner. GPH's rules were that you got full credit for puzzles no matter when you solved them, but if you took longer than 24 hours on them, you would get an "average adjusted solve time", which was the average time over 24 hours that you took divided by the number of puzzles you solved.

Only four teams solved the entire hunt in the appropriate 24-hour period chunks. We were 11th place since we didn't solve the very last two metas until Wednesday night.

I didn't really know what to expect for this hunt -- when I first heard about it, my inclination was to just plan to print out all 25-30 puzzles over the weekend and get together a bunch of people then to solve them, but I figured that maybe we could at least try the first day or two puzzles and see if it was going to be worth trying to solve them and go for a good score/time. As a result, my team was all people who work at Google MTV (me, Chris, Richard, Ken, Sean) or nearby (Glenn) or work at Google but decided not to actually meet up with us (Channing). I was originally going to recruit 2-3 more people but ended up not doing it. Team limit was 10 and we had 7, so. I figured that with a pool of 7 people we'd end up with an average of 4 who could make it on any given day, which is about what happened.

I reserved a conference room at Google that was close to a cafe and a printer and in a quiet side of the building where we wouldn't disturb anyone, and on Tuesday the 14th we got together to solve.

Well, I should preface that by saying that Channing jumped in and solved the first puzzle without asking anyone during the afternoon (puzzles were released at 2pm, but we agreed to all meet up at 6pm, I didn't want anyone to feel like they should put off doing their actual job to do puzzles). It was a little awkward but in the end it all worked out.

I had set up a hunt spreadsheet -- just one Google Sheet for the entire hunt, with an overall thing showing puzzle status (solved, not solved, and who was working on it, additional notes/links, etc?) and then each individual puzzle ended up being another subsheet of the main sheet. This was mostly so that we didn't have to worry about accidentally not sharing something to the right people -- everyone had access to the main sheet, so if you had that open, you were in with our hunt.

So that evening the rest of us (well, Sean left for a while) all got together and solved almost all the puzzles from that evening. Identities involved figuring out numbers on a logarithmic scale and assigning people to them, which I think Chris figured out early on; Angry Portals was basically playing a version of Portal but using Angry Birds to flick them around, and Ken spent a while just playing through all of the levels, and in the meantime a bunch of us slogged through the pairs of clues and cryptics in Zero Space. We finished Portal around 9:30 and then continued to stare at Learn to Play, which we had no idea what to do with -- the words all hopped around in weird ways when you resized the screen, and the HTML was full of numbers and weirdness. We would eventually solve it the next afternoon after getting a hint saying the HTML was irrelevant, basically -- we finished it shortly after that. Channing had the insight to use the game rules, IIRC.

The good thing, or bad thing, depending on how you look at it, was that by finishing Learn to Play an hour late, was that we were already off the Zero Second AAST, so there wasn't really a *ton* of pressure to stay on schedule.

But, when adjourning for the evening, around 10pm, we agreed to all meet the following night again and work on the puzzles on release schedule anyway!

(We also saw a team in the conference room next door to us -- we figured out they were also doing the puzzles when we heard the quacky Still Alive playing from their room -- funny thing is we'd kept our room door open for most of the evening, I guess they didn't overhear us much, because when we left they had solved 3 puzzles and we had solved 4 and all. They were the Puzzle Pirates, and from what I can tell they got pretty far but didn't finish the final metas.)

So, the next day puzzles came out. Ken and Sean couldn't make it to puzzling that evening -- well, they both had other commitments and said they'd be a little late. Day 2 also turned out to be the easiest day, though. We had already worked on the first puzzle a little bit during the day so that fell pretty quickly once we got the group together. Glenn just took How To Best Write an Essay and solved it on his own. Richard solved Very Fun Logic Puzzle on his own. I distilled out the information for The Superbowl mostly on my own and got help with it from everyone once they had finished their puzzles -- and Chris, being a basketball fan, did most of the work on A Glistening Occasion. I filled in a few of the people too, and then we split up watching One Shining Moment videos and looking for the people in them. We didn't quite find everyone, but we certainly found enough to get the solution.

Anyway, we submitted our answer for that final puzzle at 8:53pm, and seriously a minute or two later Ken showed up, and then about 5 minutes after that Sean showed up, and we were all like "Haha, we're finished, you guys missed out," and told them a bit about the puzzles, and then all agreed to meet the next night again, because clearly we had a good balance of people and were going to do just fine in this puzzle hunt. Well, Richard couldn't make it on Thursday, but everyone else could.

Thursday turned out to be super hard though. Ken and I had listed all of the punctuation and stuff for Puncturing Sensation early in the day, and everyone else refined it a little (like "swung dash"), but we stupidly couldn't figure out how to extract, which was frustrating because many teams had solved that puzzle. I did a reverse image search on one of the people in the Television puzzle to figure out it was Eurovision, and so I watched the Eurovision 2016 result video and ID'ed a bunch of them while Chris and Glenn got dinner from the nearby cafe to bring back. Ken showed up around then and found a music tool to type in the music phrases, so between the two of us we managed to solve that puzzle.

In the meantime, Sean and Chris had figured out what was going on with Scramble For The Stars but it was already a slog to fill in the crossword clues and start trying to figure out how the constellations were going to fit together. People also identified all of the Pokemon but we didn't know where to go next with that either (although it was Richard or Channing or someone who pointed out that all of the letters were amino acids, and I had noticed the lengths of the strings and that there were these huge similar subsequences between a lot of them).

We also all got completely and utterly sick of hearing Rock Lobster after listening to it a bunch of times for X-Ray Fish but not getting anywhere with it.

Really, it was pretty embarrassing after our strong finish on Wednesday, but when we all left to go home on Thursday, we seriously had one puzzle solved. Richard, however, got home from whatever he was doing, and figured out the extraction for Punctuation pretty quickly, so we had TWO solved when we went to sleep!

One thing about the format was that we tried to solve new puzzles as they came out, which made sense up to a point. Of course it would be better to solve the new puzzles in the first 24 hours, but once we figured out that a newer puzzle was also super hard, it almost made more sense to work on the older puzzles that we could get a hint or two on, and if we were truly stuck on the newer puzzles, go ahead and wait on that until the hinting time as well.

So Friday we had everyone again (except Channing). Though since it was Friday, we had to actually venture out to get food, since our building's cafe wasn't open (most cafes at Google that serve dinner don't serve dinner on Fridays). After solving the Famous by Association puzzle (an effort by everyone, kinda) Richard and Chris and Glenn and I got sandwiches/burgers from a food truck and Ken and Sean went all the way to Building 43 to get a real dinner.

Upon coming back, Richard started hacking on some code to solve A Basic Puzzle, and I dove into The Treasure of Apollo, and Ken roped Sean into listening to the Rock Lobster clips over and over and over again to solve X-Ray Fish, dammit (which he eventually did!), and everyone looked at Stephen's Speed Run and basically was like "lol no". Well, Glenn spent a while playing it anyway. And Ken spent pretty much the entire evening until I have no idea how late working on Scramble for the Stars (I joined him for a little while after getting stuck on extraction for Treasure of Apollo, to help work out a few of the early logic conflicts, but really, it was almost all him). We did a combination of drawing a graph on the whiteboard and of screencasting a constellation map and putting post-it-notes on the screen.

Also, we played a lot of music in the room that evening, because people kept humming the bass line to Rock Lobster, and that was unacceptable. For a while I was playing the broadway songs in Treasure of Apollo just to see if they'd inspire anything, and for a while we also put on various songs that were vaguely themed like our puzzles, and then Glenn just played Weird Al stuff for a while. (Also funny: Glenn has never seen or heard Book of Mormon, so when we were playing Hasa Diga Eebowai -- one of the songs in the Apollo puzzle -- he looked up from what he was doing at the time suddenly like "OMG you guys, I thought this was the Little Mermaid, and suddenly they are all swearing and talking about AIDS, what the fuck?")

I helped out with Watchers and Fliers (checking/confirming/drawing the cryptos and all; we sort of solved it as a team) and came back to Treasure of Apollo many times. Sean had the insight about the Pathfinder classes though (I had done the musicals and characters and stuff but we couldn't figure out how to index into it all). But we still didn't have an answer. Richard left fairly early because he's a morning person, and Chris and Sean left to bike home around 11pm or so. Glenn and I messed around with the substrings on the Pokemon trying to figure out ways to align them, and Ken continued slogging on the Stars, and somehow a little after midnight we came back to the Apollo spreadsheet and corrected a few of the classes and characters, and it was like hey, wait, we have the word STORE showing up now, that's promising... and a few rearrangements later and it was like "FF MP RESTORE? Are you kidding me?" but sure enough, the answer to the puzzle was ETHER. I was absolutely floored that it worked but glad to have that over with, because it was super frustrating to have done all of the leaps in that puzzle (we got the lyric strings out, identified musicals, characters, pathfinder things, etc, etc, etc, so not having an answer after feeling like we had made all of the leaps was so frustrating!)

After that Glenn and I both left. Ken apparently stayed for another hour or so and got enough of the letters that Channing was able to swoop in and extract an answer.

We still didn't have the Pokemon solved, and nobody wanted to optimize Stephen's Speed Run.

So, going into the weekend. I was beginning to think we might have been better off recruiting another person or two, but at this point it seemed kind of silly to try. We also switched venues to my favorite puzzle hunting room on campus -- a nicer bigger room next to a cafe and next to bathrooms and next to a printer and everything. (I didn't use it during the week though because I didn't want to disturb anyone, plus that cafe isn't actually open for dinner.)

Officially we began at 2pm both days, but Richard showed up earlier. Chris and I grabbed lunch from a nearby taqueria and got to campus around 1pm. Glenn showed up a bit later. Sean didn't show and Ken called in over video chat since he was kinda burned out (understandably) after working on the star charts until the middle of the night.

We were delighted to find that the cooler in the building that usually stores ice cream bars still had them over the weekend, of course :)

Richard had identified a bunch of the bases already, so we all jumped in on it and found more of them. (Believe it or not, I got to identify the bolded numbers as phonespelling in decimal -- they had been thinking that one of the bases would be hiding in plain sight.) So they kept working on that and it fell shortly after 2pm.

In the meantime, we had new puzzles that came out at 2pm. I picked up dʒʌmbəl and started working on it, because, you know, I always end up doing IPA puzzles. Richard started working on Thunk after the bases puzzle fell (I was like "it's an escape room text adventure! Someone else should do this!"). Chris and Glenn both worked on Unaligned and on YOLO Queue, and oddly, Ken worked on dʒʌmbəl with me over video chat and spreadsheets. Some of the words we got quickly but some took quite a while. (Oddly, though, the final phrase wasn't so hard for us, though.)

I worked on trying different indexing schemes for Drive for a while and then came back to Retinal Variants, the Pokemon puzzle. We finally asked for a hint about the cluephrase "INDEX NAMES BY LAST DI" and found we'd have to continue it somehow, though not in the amino sequences. Somehow I started googling things about amino acid protein sequences and how to translate them and I found out about BLAST (and was like "omg dr. blastoise that's what they meant") and about Pikachurin and was basically like "If I found out there was a protein called Pikachurin I too would totally want to write a Pokemon puzzle about it but holy fucking shit I'm really not sure how I was supposed to get here."

So then I spent -- seriously -- the next half hour translating the protein sequences. BLAST on the web is cool and all but it took around 2-3 minutes on average to translate each one. Fortunately there was a daily dungeon on Puzzles and Dragons going on around there anyway so I would play a board or two while waiting for the next sequence to go through.

Chris went and grabbed Subway sandwiches for us for dinner (Richard went home around then, and I forget what Glenn did) and then we went through the rest of the Pokemon puzzle and solved it and then we went home.

Sunday was more of the same, of course. Same room, though this time Glenn couldn't make it and Sean was pretty late, but Ken was on site this time.

Of course, Sunday was metas. So. Our accomplishment for Sunday was mostly solving the rest of the open puzzles, basically, with a few hints. (Infact, since we arrived at 1pm or so, Richard and Chris and I wrote out a couple of hints on the whiteboard specifically for Drive and Unaligned, to be asked at 2pm, and we also figured out one for Stephen's... and realized we could ask it whenever, so we did. Heh.)

So after we confirmed what we had to do for Stephen's, Ken took Glenn's solve paths and optimized them a bit and then I managed to figure out the clue phrase from the letters we did have. Richard and Sean worked out Unaligned after we got confirmation about only needing the shaded areas, and then some combination of everyone except me worked on the Duck Quonundrum while I was meanwhile still goddamn staring at Drive. First we had confirmed that the movie names were irrelevant, then we confirmed that the driver names were irrelevant, then we confirmed that we needed to fill something in on the grids. (I had already tried indexing into just about everything using either the movies, the drivers, the cars, the starting locations, the ending locations, summing the two of the locations, etc, etc). Someone came up with the idea of putting in the license plates, realizing that ECTO-1 was the same length as the Ghostbusters car trip, and then we all went on a Google Image Search scavenger hunt to find the license plates in the movies. And of course much like everything else we still ended up almost there but not quite (with CROOKEDST, CROOKESST, etc, etc) until Richard, who had gone home by then, pointed out that you had to use all 18 number-letter pairs. Oops.

So, all the puzzles were solved then, but we just struggled like hell with the metas. I had of course noticed the ingredients pretty early on -- and could not unsee TOMATO in SEMIAUTOMATON -- but weren't sure how to make them fit into the conference 2 lines. The Conference 1 "word find" was pretty much completely incomprehensible to us.

So we gave up around 8pm on Sunday and went home to get dinner and all, again with only one puzzle solved.

This is also where I can't really comment on how the rest of the hunt went, because I was driving down to LA at the end of the work day on Monday. I told people they should be free to get together and solve things if they want, but as far as I can tell, the rest of the hunt was basically solved over our spreadsheet and over email, with people conferring on which hints to ask and sharing ideas. We did finally solve Conference 1 and 3 on Wednesday evening, which completed the entire hunt, basically, and I do understand how all the puzzles worked, but I personally didn't really do any solving at all past Sunday evening.

But I am really proud of everyone for solving everything, at least.

Since this was my first time doing one of these hunts, I'm not sure what to think. I mean, I enjoyed having puzzles to do every day, and getting together with my friends to hang out and do puzzles is always fun. But I think we got pretty frustrated later on in the hunt when the slogs and leaps got harder, and I also think people got pretty burnt out after a while. (But only kind of. I mean, you could see us in the conference room later in the evening just like "aaaaaaaaaa what IS this stuff", but at the same time, people were always peeking in at the hunt spreadsheet during work hours and exchanging emails around the clock about solving ideas. So.)

Overall I'm glad we did it -- I'm always complaining there aren't enough puzzle hunts anymore and so I'd be a huge hypocrite for not taking advantage of the ones there are. I'd heard about SUMS/MUMS before but wasn't sure this format would really excite/motivate me (no physical component and not really interacting with other teams), though I think the leaderboard aspect here helped a lot because we kept looking at other teams like "argh, everyone else is solving this, it's clearly solveable, what's wrong with us?"

Ok, well, I've been writing this entry over a couple of days and I'm not sure it's entirely coherent or will even be interesting to anyone besides me. So.