First, congrats to Speedrun or bust, We Regret To Inform You That It Is Projectyl, and Dragon Phoenix Dog 1 for being the first three teams to finish (the first in under an hour, the second and third both being solo teams, and the third for 100%ing the hunt with no incorrect guesses!) Also congrats to mattapuzzle, our fourth place team, for being the first team to 100% the hunt. Congrats to Ski Nautique, Lil Stoop, and Kwargers for winning the high school division (Ski Nautique also being first to reach 100% in the division), and solosis, those without sanity, and Peehs for winning the middle school division!

Other records: Speedrun or bust also got the first solve of the hunt, solving Bottom Lines in 31 seconds from the start of the hunt, and mattapuzzle got the first solve of the first meta. And returning to the final meta, Me-Stik Spiral was the team that solved it the soonest after unlocking it. A quick shoutout to first solves on all the other puzzles: Congrats to

Finally, thank you to everybody who donated to the low-tech charity drive: we received receipts for over $1,200 of donations!


We were all blown away by how many participants we got. We each tried to predict how many teams would solve at least one puzzle in the entire hunt; the highest prediction any of us gave was 300, and that was passed in less than 15 hours.

Here's a chart of the top 30 teams' progress! (Times are in UTC.)

We also have charts of the progress of every finishing team, the bigboard we used to monitor hunt progress, and a separate stats page for the overall hunt. And on the puzzles and teams pages, every puzzle and team also links to statistics about that specific puzzle and team.

You can also download the full guesslog for the puzzlehunt; if you find other interesting statistics or visualizations, we'd be interested to know!

Story and Theme

Solvers were introduced to Dippy the dragon-phoenix, who was celebrating their rebirthday on Dilemma Planet, but was unable to find their friends to celebrate with. Dippy asked solvers to solve puzzles on Dilemma Planet to find them. It turns out that those friends were different birds who had hidden themselves among the first round answers, and solving the first round meta helped Dippy message all of Dippy’s friends by TWEETING.

In the second round of puzzles, solvers had to solve puzzles to fuel Dippy’s rebirth pyre. Each second round puzzle was related to fire in some way, and putting all of the answers together revealed Dippy’s new form, a PYROTECHNICAL FOWL. After solvers solve the main round meta, Dippy completes their rebirth, rises from their pyre, and gives the solvers one of their featherscales as a parting gift.

The idea and name for the DP Puzzle Hunt basically came from the existence of the BN puzzle hunt and Co-Puzzle Hunt earlier this year and the observation made by several people that Caesar shifting “BN” by one produced “CO”, so “DP” would be the logical next hunt.

On August 15, betaveros came up with “Dippy the Dragon-Phoenix” as a half-joke in 15 minutes. As the hunt developed, Level 51 and Joman came up with various ideas for worldbuilding and the story, and somehow our hunt developed to the point where it had a plot and kind of coherent metas, all in the space of two months.

Design Goals and Philosophy

Our general goal with DP Puzzle Hunt was to write an easy puzzlehunt. Broadly, we were inspired by the many people who have been running and announcing puzzlehunts on Puzzle Hunt Calendar lately (perhaps in part due to COVID-19 lockdowns making other forms of entertainment harder to participate in?), and wanted to join the fun.

At the same time, we realized that a lot of our favorite puzzlehunts that we’ve written and participated in are fairly difficult, perhaps because they’re targeted at people similar to the writers and because often, the most interesting puzzles require a certain floor of difficulty to retain their interestingness. However, this also means that completely beginner-friendly puzzlehunts are a lot rarer, so we set out to write one, while still holding our puzzles and puzzle mechanics to the same standards.

In our initial planning doc, we wrote that “the median puzzle should be a 6x6 crossword with some squares circled”, and in the rules, we compared our hunt to Colby’s Curious Cookoff (which is a recent polished, beginner-friendly hunt we highly recommend). In hindsight, we realized we likely overshot the advertised difficulty with some of our Round 2 puzzles. We expected The Magic Match to be difficult due to its length, but were somewhat okay with it because it was at least still approachable to solvers who hadn’t done puzzles before. Probably bigger offenders were Deity Duos, which was nerfed dramatically from an even harder version (!) but had burned through most of our testsolvers at that point, so we didn’t have a solid grasp of its difficulty; and Geocaching, which was written quite late, so we weren’t able to spend as much time as we might have liked to calibrate the difficulty. On the bright side, based on the feedback form we got, the vast majority of responses rated the overall difficulty level of the puzzles either a 2 or 3 (on a 5-point scale), so we think we probably didn’t overshoot the optimal difficulty.

A specific consequence was that we wanted to include an easy metapuzzle. Metapuzzles are one of our favorite parts of puzzlehunts, but we saw that they were missing in some of the recent puzzlehunts on Puzzle Hunt Calendar. Also, metapuzzles are a popular target of aggressive innovation and boundary-pushing in puzzlehunts (the 2019 Galactic Puzzle Hunt is a striking example), which makes sense because they’re many people’s favorite part of puzzlehunts, but also can make them even more inaccessible to new puzzlers than the regular puzzles. By contrast, the substring metapuzzles we included in our puzzlehunt are far from groundbreaking; we know of many other substring metapuzzles and wouldn’t be surprised if either exact category of substrings had appeared in other puzzlehunts. Instead, we simply hoped to demonstrate that metapuzzles could be simple to write while still being thematic and fun to solve. (Empirically, it looks like we succeeded in that our final metapuzzle received the highest fun rating among all puzzles in our hunt.)

A side note on metapuzzles for anybody out there who might be writing their own puzzlehunts: Obviously, most metapuzzles place constraints on the answers and necessitate them being written before any puzzles can really be started. This could limit people who want to write puzzles with extremely constrained or thematic answers, or who already have puzzles written before the meta was written. For the specific metas we chose for our puzzlehunt, we would probably not be able to accommodate such puzzles. However, we believe that it’s still possible to write interesting metapuzzles that are flexible enough to accommodate many answers. For example, the MUMS 2016 meta was an extremely flexible and (in our opinion) entertaining metapuzzle that was able to accommodate 22 of 23 random puzzle answers. As the meta answer page says:

One of the first decisions made in this Hunt was to use completely freeform answers. This was partly because we had plenty of first-time contributors to the MUMS Puzzle Hunt, but also because many of them had written their own puzzles before in preparation for their own puzzle hunt. Instead of modifying lots of puzzles to use restricted answers, we decided we would accept all puzzles (and their answers) this year, and can think about a more restrictive meta next year. This turned out not to be completely true: after 23 puzzles were finalised we finally decided on the current meta string. It turned out that we could fit 22 of them, and one of them (Pedigrees, fortunately the most flexible one) had to be changed to begin with a W. The remaining 2 puzzles came after, with one-letter restrictions.


Although, as solvers, we liked free-form hint requests a lot, multiple members on our writing team had participated in running hunts using free-form hint requests before and had learned that it was extremely time-consuming. For our hunt and our team size, we were fairly concerned that we wouldn’t have enough manpower or timezone coverage to properly respond to many hint requests from teams. So we settled on pre-written hints as the primary mechanism of delivering hints.

One big drawback of pre-written hints is that there’s always the risk that the team is stuck somewhere that the pre-written hint doesn’t help with. (We know this concern was already discussed in e.g. 2017 GPH.) So we tried pretty hard to make both of our hints useful to solvers no matter which stage of a puzzle they were stuck on. Even so, there were definitely places we didn’t expect solvers to get stuck and didn’t succeed in cluing in our hints. For example, we failed to anticipate the number of solvers who got to the last step of Nought But a Crossover, but didn’t call it in because they didn’t recognize the letters they read from the O as an actual German term instead of just random letters.

We’ve seen other hint systems that tried different ways to resolve this issue, which we might consider incorporating elements of in future hunts. For example, REDDOThunt has an Early Hint and a Late Hint for every puzzle; the two hints can be “purchased” separately, and they provide you a hint with an early or late step of the puzzle. This is also an improvement, but sometimes it’s hard for solvers to know when they’ve passed the stage when the Early Hint will be helpful or entered the stage when the Late Hint will be helpful.

Separately, we’re not big fans of voluntarily purchasing hints with actual points, which can lead to highly regretful and unfun situations. For example, you might purchase a hint and then realize that you were really close to solving the puzzle, so you just have to live with the fact that you irreversibly decreased your maximum score by making that choice. Even worse, you might purchase a hint that turned out to be useless based on where you were stuck. Now you effectively have to solve an equally difficult puzzle for fewer points.

It might be counterintuitive because this gives teams less choice, but we prefer mechanisms where hints are automatically released and, optionally, puzzle scores automatically decrease at scheduled times; this guarantees that the optimal strategy is always just “try to solve puzzles using the information you have available”, and you don’t have to worry about the meta-strategy of whether a hint is worth spending in a particular case. (Similarly, we don’t like to assign many different point values to puzzles, even if they’re different difficulties, partly to keep the strategy simple and partly because puzzle difficulty is very subjective anyway.) Another way to avoid the situation is to use a separate, unscored currency for hints. We used both mechanisms in our puzzlehunt.

As the hunt progressed, we eventually did release hint requests, and at least for this hunt, our concerns about the hint volume being too large to manage proved to be unfounded. After we hesitantly released one hint request per team, we did not get hit with a deluge of hint requests; we only got 27 hints over the whole hunt (2 of which were obsolete, and 1 of which was a test hint by our friend), and we were generous about refunding them when we were giving really small hints.

We’re not completely sure what caused the stark difference between our prior experiences and our experience with hint requests during this hunt. Perhaps the pre-written hints were a very good first line of defense. Perhaps our hunt was just at a difficulty that people were bottlenecked at finding enough time to work on it rather than getting stuck. In any case, we may be more generous with hint requests if we run another hunt of a similar structure and difficulty in the future.


(this section is by betaveros)

A secondary goal I had with running this hunt was to use it as an end-to-end test of running a hunt with the GPH website code, to see if I can recommend other people who want to run puzzlehunts to use it, and if not, what’s missing. I also tried to host the hunt on a pretty tight budget. Throughout this process, I found Alex Irpan’s blog post a valuable reference; in particular, it convinced me to at least casually load-test our website.


Here's a Puzzlord-generated image of our progress:

Funny answers

(These are just the ones we thought fit to print...)

Fun stories people sent us

Hindsight is 5780:

We watched “supa hot fire vs. b bone” as a joke, because I asked if anyone had famously declared themselves not a rapper. As soon as it finished, the meta clicked for all of us and we started yelling

We received various comments about “Hades”, with some teams complaining that it was too obscure of a game, but some other teams relaying stories about how they were playing it right before the hunt or started playing it after. A highlight were StriketeamA’s comments; after politely suggesting the game might be too niche, they followed up with:

Ignore the comment on Deity Duo from my last survey. I apparently had it in my steam wishlist and totally forgot about it. Good memory, huh?

A long story from Denver Nuggets:

Unbelievably, our Puzzle Hunt team—the Denver Nuggets—seemed to dictate how the actual Denver Nuggets performed on the basketball court. Our puzzle solving ability on the night of Friday, September 21 correlated perfectly with the Denver Nuggets’ point scoring against the Los Angeles Lakers in Game 1 of the Western Conference Finals. (Note: The Nuggets are substantially weaker than the Lakers.) We started off strong, solving many puzzles quickly. This was paralleled in the Nuggets’ first quarter, leading 38-36 at its end. Leading the Lakers is no easy task. Then, we couldn’t go forward because of the locked puzzles and point requirements. During this time, the Lakers made a 17-1 run to take a huge lead. That entire night of puzzle solving was filled with frustration and anger. Meanwhile, in the NBA game, the Nuggets were down 27 against the Lakers. The Nuggets lost the game, and the game was far more lopsided than the final score of 114-126. Similarly, we were wrecked in this puzzle hunt, and barely proceeded after Saturday night. The hunt carried its bad karma in the recent Game 2, when the Nuggets led 103-102 after trailing 92-100 and by as much as 16 during the game. Then the Lakers’ Anthony Davis made a game-winning/saving buzzer-beating 3-pointer, and currently, the Nuggets trail the Lakers, 0 games to 2. Hopefully typing this story for you to see will save the Nuggets, who made two 3-1 comebacks in the two rounds before the Western Conference Finals, becoming the only NBA team to do so. GO NUGGETS! ELEVATE!



About us

Q: Names please, I’m really curious. Maybe lydian and greenturtle were a part of this. But maybe I’ll never know.
Q: are you ever going to reveal your identities? the public demands answers.
Q: We are interested to know about the team behind the puzzles, and their puzzlehunting and writing experience/stories to share.

Q: Who testsolved these?

Writing the hunt

Q: Please describe some ways puzzles changed in testsolving!

Q: Which puzzles were the hardest to write/playtest?

Q: How long did it take to write the clues for Downright Cryptic/A Fork in the Road?

Q: What on Earth prompted you to include a duck konundrum?

Q: How long did it take you to make this?

Future hunt plans

Q: Will you do an EQ puzzlehunt?
Q: Are there going to be more hunts like this? This was awesome!
Q: This hunt was great! And I can’t believe how quickly you put it together. Do you have plans to do another one in the future?

Q: If you make any future hunts, will Dippy be returning (or will you have a new cast)?


Q: Do you guys have answer times saved?

Q: How many people submitted DUAL MASTER for 21st Century Celluloid Man?

Q: showing fun & difficulty ratings publicly was… unconventional. not sure i have an opinion one way or the other re: good/bad. i suppose difficulty ratings could be used as a way to determine which puzzles to solve or skip, but that information is also conveyed by # of solves/guesses. mostly, just. Why?


Q: Did seriously nobody register the domain puzzlehunt.net before this hunt?

Q: How about open-sourcing the website? :)

Q: Why does Kipper sometimes speak with emojis? (Geocaching)

Q: Did the tour guide from Geocaching ever put 2 and 2 together.

Q: Why was “spark” bolded in the section intros…?

Q: Do you have any ideas for how to reach out to more young puzzlers?

Q: What DP stands for and how did you name your crew?
Q: How did you choose the name “DP” for the hunt?

Future Puzzling

Puzzle Hunt Calendar is a great place to find other upcoming puzzlehunts. One of the hunts we’re looking forward to is Matt and Emma’s Birthday Bash, which is coming up in about a month and at least partially run by our first-place team.

Thanks for playing!