For Mystery Hunt 2020, I hunted with teammate, the same team I hunted with last year. We ended up smashing all my expectations. By puzzle count, we were 2nd. By metas, we were 2nd to solve all metas. By Hunt finish time, we were 3rd, due to getting stuck on the last puzzle (more on that later). This was the first year where teammate got a call from HQ warning us that we were in contention for winning, and for much of Saturday, we were the lead team, which is a rather weird prospect to consider. I didn’t believe it until someone came back from trivia for Weakest Carouselink and said we played against Palindrome, indicating we unlocked the puzzle at the same time they did.

Based on questions Left Out asked us during Hunt interactions, I don’t think many Mystery Hunt veterans know where teammate came from. Very briefly, teammate is a mix of people from Puzzlehunt CMU, some Bay Area puzzlers from Berkeley, and friends branching out from there. Before teammate, these people hunted with ✈✈✈ Galactic Trendsetters ✈✈✈, but one year they decided to split into their own team, and teammate and Galactic have been sister teams ever since.

The two teams have similar team culture and age demographics. For the former, both are meme-heavy and very willing to backsolve. For the latter, the majority of both teams are younger than 30. I’m actually not sure if anyone on teammate is over 30, now that I think about it. This showed itself most strongly when Left Out came to deliver the Baby Shower Balloon. They started a clue with “BTS…” and three of us immediately guessed K-Pop. Then we got a question about a VH1 television show that showed music videos (Pop-Up Video), and our reaction was “What’s VH1?”

Before Hunt, members of Galactic and teammate leadership were seriously considering merging the two teams. The downside would be a less fun Hunt, since we’d be way over the recommended limit of 70-80 people. The upside was that we’d have much better odds of winning. We didn’t merge, but based on the solve graphs, we clearly didn’t need to merge to be a contender for the coin this year. We knew Galactic wanted to be more competitive this year, and weren’t surprised they got to the coin first.

* * *

I felt this year’s Hunt was really good, in inventiveness of its structure and polish in its puzzles. However, for me it still feels a bit sour due to how the Hunt ended for us.

On Friday night, we felt we were doing well. I stayed up until 6 AM, then slept for 4 hours and came back. On Saturday night, we knew we were doing well. We rightly guessed we had unlocked every round, we were making good progress on all the metas, and it started to look like if I went to bed, I would wake up after we finished Hunt. So I pulled an all-nighter, and I know other people pulled one as well. At 9 AM, we solved our last meta, and frantically sent someone to get the last penny.

Every time you solved a round in this hunt, you got to pick up a pressed penny. Each penny had 3 arrows, 3 images, and 3 sentences of text. In short, it was definitely puzzle data, but I was expecting a shell metameta, and didn’t bother looking at them too closely. Additionally, our home base was a fairly long walk from Hunt HQ, and Left Out told us it would be fine if we picked the pennies up in batches, to avoid having to send someone there and back for each meta solve.

We picked up the final penny, didn’t unlock a shell, and realized it was a pure metapuzzle, at which point we started looking at the pennies, and got horribly stuck. The Workshop isn’t on the hunt website yet, but the solution is described at wrap-up. We got the penny layout within the 1st hour, and were stuck on extraction for the rest.

About 3 hours in, we had concluded two things. First, if we had been in the lead, there was no way we had a three hour lead over 2nd place. Second, the coin hadn’t been found yet. Assuming the reasonable-length runaround, that implied that we were one of several teams stuck on the same pennies puzzle, and Hunt would come down to whichever team got the break-in first. And, every other contender for the coin would have made the same inference, and were likely staring at the pennies just as intensely. That was great for motivation, but bad for anybody looking to sleep, because any moment could be the moment we got the a-ha.

At 3 PM, Galactic found the coin. Our entire team had been looking at the Workshop pennies for 6 hours. We knew it was the only puzzle that mattered for Hunt completion, and we still couldn’t get it. Feels bad. As soon as the coin was found, Left Out called us, gave us some hints, and fastforwarded us through the endgame, which was both very fun and quite impressive.

So, what happened? Word on the grapevine was that Workshop testsolved perfectly fine within Left Out, and they were surprised it took teams as long as it did to solve. In my opinion, part of what played into this was that we knew the pennies would be part of a puzzle, but we assumed we could look at it later, when we had complete information. By the time we actually looked at the pennies, we were very sleep-deprived due to rushing for the Cactus Canyon meta, so we weren’t as sharp. As time passed, we got increasingly burnt out on trying to figure out the same freaking pennies for hours.

In many ways, this was similar to what happened to my team for Galactic Puzzle Hunt 2019. We put off learning Puflantu until right before looking at metas, under the logic that it’d be easier to learn it all at once after unlocking more artifacts. In practice, it turned the language learning into a big slog. What do you do, when there’s nothing to do but learn Puflantu, and Puflantu is hurting your brain?

I’m trying to figure out why the clocks for April Fool’s Day Town from Mystery Hunt 2019 didn’t have this issue for us. It has a similar structure, where a small bit of information from every round gets pulled into one final puzzle. However, we solved that meta without too much pain in the final hours of Hunt, once we had enough prank answers. I think the crucial difference was that during the Hunt, we started looking at clocks early because there was no indication the clocks didn’t matter for the current town. We tracked them early, giving us all the data we needed to extract once Sunday came around. In contrast, by getting pennies at the end of a round, we got a signal that we didn’t need to look at them until the end. (And if they were part of a shell meta, looking at them early may have been wasted time, compared to solving other puzzles.)

* * *

I’ve been thinking about puzzlehunt design more. You may also have noticed that this post is labeled “Part 1”. These two facts are linked! I’m one of the organizers for My Little Pony: Puzzles Are Magic, a puzzlehunt that starts in about 10 days.

Currently, I’m 50-50 between trying to get more sign-ups, and freaking out that people are going to hate it for not living up to expectations. I’ve been told this is a natural part of the creative process. I’m also still freaking out.

Puzzles are Magic should be fine, but in the next few days, I’ll be spending more puzzle-time on making sure the hunt website can handle the load, and less time blogging about Mystery Hunt. When things are less busy, I’ll share specific puzzle stories.

Last note: automated callbacks! I’m in favor of it. Somewhat selfishly, as teammate benefits from it a lot. I understand that some people like to hear a human voice now and then, and hunt organizers like how it lets them listen to teams celebrate a puzzle solve, but there’s no reason you can’t occasionally call teams for a check-in, or visit them in-person, while letting an automated system handle the majority of calls that are simply replies to puzzle guesses. It’s not a binary choice. Just saying.

Oh, and congratulations to the new couple. It was very, very sweet.