Thanks to everyone who participated in the Advent Puzzle Hunt!
Special thanks to our 93 kickstarter backers who helped bring the physical advent calendar to life.
The idea for a puzzle hunt based on an advent calendar came about during the 2023 holiday season, when I (Eshan) was browsing advent calendars for other things like jam and hot cocoa and thought I could make my own. The target date of December 2024 gave ample time to put together a hunt, but a firm deadline to stick to.
With the advent calendar format, we hoped to keep some traditional features of puzzle hunts (style of puzzles, meta / supermeta structure), while being unique and memorable in our own ways with the casual format spanning over an entire month, along with the option to solve the hunt as a physical box.
The original pitch of the Advent Hunt from January 2024:
This is an advent calendar, with doors numbered 1 through 24 – one puzzle per day. There are metapuzzles on days 7, 14, and 23, each using the feeders from the previous week.
The plot: At the start of the month we meet Red, one of Santa’s elves. Santa has gone missing! Red is planning to look for clues to find him, but the puzzles are kind of hard and he asks us for help. Each of the 3 metapuzzles gives Red one piece of evidence to find Santa. In the supermeta, Red finds Santa and we learn that Santa set up the whole hunt as a Christmas present for Red because Red likes puzzles.
The target difficulty is easy, like fish puzzle difficulty. A successful test solve should take between 15 and 45 minutes. If this goes well, I’d like to produce physical copies of this (while also having a website for people who don’t want a physical copy). Therefore, each puzzle must be printable (no digital assets needed) and at most 1 page long. Half page puzzles are also possible.
Eshan and Jay started the first steps of planning after the Galactic Puzzle Hunt in November 2023. (We had set aside the full weekend for puzzles, but finished the hunt Saturday night and had spare time). We came up with the plot premise of Red the elf searching for Santa, as well as a rough idea for the supermeta and the three meta answers.
The Herrings 🎏 writing team was a group that had solved hunts together regularly for some time, but only sporadically written puzzles (some authors were new writers). To help give the hunt momentum and show it was serious, Eshan finished the supermeta and most of the metas in advance, before rallying the whole team to help writing.
Puzzle editing was done mainly over Discord, with a dedicated discussion channel per puzzle containing the author, editor, and all test solvers after their solve attempts. We required each puzzle to have at least one clean solo solve, with a target solve time between 15 and 45 minutes. The actual number of tests per puzzle ranged between 2 (Elf on the Ice Shelf) and 6 (Winter Wonderland) depending on the number of revisions needed. Fortunately, we never ran out of clean test solvers on the team, and we did not have to scrap any puzzles after they were drafted (although some underwent considerable revisions). We had one in-person meetup, towards the end of the writing process, in which we finished off the last straggler feeder puzzles, wrote the Kickstarter preview puzzle, and started setting up the website and Kickstarter.
When developing the puzzles, we went back and forth about difficulty. One tricky thing about this hunt was anticipating our audience’s experience level, as we figured we might attract both newcomers interested in casually solving a puzzle a day as well as dedicated puzzle hunters very familiar with solving techniques who might appreciate more of a challenge. It turns out that creating puzzles that take 15-45 minutes for solvers with different experience levels is a difficult sweet-spot to hit!
In feedback conversations, we often considered whether simpler puzzles were interesting enough as-is (e.g., Yule Log Tavern) or if we needed to add extra layers to be more challenging (e.g., Gingerbread Bakery started with just morse code and we added an extra binary step). One of our testers had a great idea for a more complex multi-step Winter Wonderland puzzle, which we eventually scrapped as we worried it would be too difficult (though we did share it in the Discord for anyone interested).
A fun meta-game that emerged for more experienced solvers was trying to "all-brain" some of the puzzles – which means no google, no writing anything down, just looking at the puzzle page until you know the answer. Our hats off to any “all-brainers” reading this!
We ran the Kickstarter for the physical edition of the box strictly after all of the puzzles were written. This was partly to be able to tell potential Kickstarter backers that all puzzles were already written and avoid overpromising, and partly to keep our collective minds on one thing at a time. In retrospect we probably could have run the Kickstarter a little earlier in the writing process, and finished the tail end of feeder writing in parallel.
We had to wait for the Kickstarter to close to get a final count on the number of boxes, before starting assembly. This left a short window of 6 weeks to complete all of the assembly. The total count of physical boxes was 131, including those ordered through Kickstarter plus those ordered by hunt authors for themselves and friends.
We ended up working with 7 different vendors (plus basic supplies from Amazon and Target) to fulfill the components of the box. The most dramatic delivery was of shipping materials, including the envelopes used for each puzzle, which filled Eshan's entire front hallway.
We had other vendors for the puzzle printing, custom printed boxes, stickers, ornaments, pins, and shipping.
Assembly was done by hand over the course of about 5 weeks, including stamping envelopes with numbers 1-24, stuffing envelopes, and folding and filling each box. The full process turned out to be more labor intensive than planned, and the order size of 131 felt close to the maximum we could handle without automating more steps or hiring helpers outside the team.
Some additional pictures from assembly:
The hunt website was a Django web application developed from scratch by our tech lead, Jay. A quick proof-of-concept (puzzle page with a PDF and an answer checker) was thrown together in a couple of hours on a weekend in April, and most of the development happened between mid-July and when we launched our Kickstarter with the teaser puzzle at the beginning of September. We have made it open source on GitHub here. You can find additional in-depth discussion about the hunt website in Jay's blog post.
Perfectionists (Solved all puzzles with the fewest wrong guesses)
Most determined teams (Solved all puzzles with the most wrong guesses)
Most common wrong guesses
Interesting wrong guesses
Crowd favorites (Most picks for favorite puzzle in the post-hunt survey)
For the puzzle The Gingerbread Bakery, we modified a real gingerbread recipe. To suit the puzzle constraints, we made some edits, but believed that the recipe should still work (some solvers noted that we used margarine in the ingredients but butter in the instructions, so chef's choice?). Karis, one of the authors of the puzzle, tried it out by actually baking the recipe:
The small surrounding cookies are a halo of herrings. Her family reported that they were definitely cookies, but tasted more like ginger scones than gingerbread cookies.
When we asked for fun stories to share at wrapup, we were expecting things like this:
We were stuck on the printer's devilry puzzle for a long time and so eventually we started trying to use ChatGPT to help.
It did not work well:
“Considering the phrase "The basil is old. It spews at a huge markup!", I think the word TAME might fit nicely, transforming "basil" into "bTAMEil" or something similar.”
What we were not expecting is the number of people who told us about doing these puzzles with their friends, family, and even classmates. It was genuinely heartwarming to see how many people made participation in our advent calendar part of their holiday rituals, and we’re delighted and honored that so many people used our puzzles as a way to spend quality time with their loved ones during a hectic time of year.
It was great to see the positive reception and interest in future hunts. We're enthusiastic about continuing to write puzzles in the future, but have no concrete plans yet. We may consider a future Advent Hunt, or something else. There is a possibility that we get roped into writing a certain much larger hunt in 2025, which would put Advent Hunt on hold.
Writing Team (Peppermint Herrings 🎏): Eshan Mitra, Zach Barnett, Karis Jones, Nathan Jones, Jay Qi, Alex Walker, Sara Walker, Zach Zagorski
Hunt Lead and Editor: Eshan Mitra
Tech Lead: Jay Qi
Business Lead: Sara Walker
Art Lead: Eshan Mitra
Test Solvers: All of Peppermint Herrings 🎏, Thomas Gordon, Kylee Hench, Chris Yu
Physical Production: Eshan Mitra, Emily Mao