So, before I put the Mystery Hunt out of my mind for a while, a few people asked me about the duck konundrum. This made me aware that though I'm generally associated with popularizing the genre, I've never really talked much about theory. That's probably because my method isn't particularly sophisticated. There are essentially three steps to writing a konundrum:
1. Figure out a structural gimmick. The first Duck Konundrum I wrote (which at that point, as far as I knew, was the only one) was basically a musical chairs game with one set of instructions. Since then I've done them with parallel sets of instructions, instructions you look up from a table based on which card you draw, instructions that are triggered by the actions of others, and so forth. Switching up the foundation keeps each iteration from seeming too much like the last one. The one real constraint is that, while you can solve one of these with one or two people, lots of people look forward to doing it in a group, so building it for that sort of experience is key.
2. Figure out what needs to happen. Usually you need everything to end up in a certain place or state at the end to yield an answer. For the most recent Konundrum, I also had necessary midstates that would generate which routine to do next. In some ways, this made things easier for me, since it meant I could write the whole thing in six independent parts. In other ways, it made it harder, since it turned out that moving stuff from state to state in the last two routines was a colossal pain. But these things happen. Anyway, some constraints are stronger than others (maybe you just need ten of something to be somewhere to yield a J) but whatever you decide on should become canon so you know where to aim.
3. Write directions. This, as you've probably guessed, is the time-consuming part, although it's not as hard as it seems. You basically just want to move stuff around in an interesting way that gravitates toward your desired endstate, possibly passing through events you want to happen along the way. (For example, in TDK2, I knew I wanted to switch pieces to different boards and spell out stuff on the Scrabble board.) You want to take very careful notes, because you may need to back up if something causes a problem, and you probably want at least one test where you hover over a group with the step-by-step solution in hand. (I did not do this for DK2. Chris probably would have murdered me during the test session if I weren't safely in Pennsylvania.)
Writing the directions involves a lot of creativity and instinct, the latter of which I've developed doing it a bunch of times. How to steer things in a fun and timely way is just something you eventually get the hang of. But there are a few philosophies that I think are particularly useful.
Don't trap the solver. The least fun part of solving konundra, at least based on feedback I've gotten, is getting far in, realizing that something's wrong, and not knowing where to start over from. As a result, you want to be very careful about anything you put in that's misleading. Let's say a solver makes the "obvious mistake" at a step. When do they realize it? If it's ten steps later, are they going to have any idea what they did? Are they going to be able to easily reset? Fun is a big goal here, and so you want to be careful about whether your actions are going to make the solvers miserable. That said, Mystery Rallye (in the 2008 Hunt) had tons of traps, which is apparently something rallyes have. Many people were cool with that, so your mileage may vary. (No pun intended.)
Don't coddle the solver. On the other hand, you don't want to make things too easy. DK3 (The Fellowship of the Duck) was probably my least popular Duck Konundrum, and when I talked to people about it, many of the "meh" people said they didn't feel challenged. We are at least pretending this is a puzzle, so it should make people think. Little embedded challenges are a good thing; you should just be careful that they're not designed to lead solvers waaaaaay down a false path.
Reassure the solver. Checksums are a big plus. (No pun intended again.) I don't endeavor to make sure solvers know every single thing is in the right place at all times, because that would drive me insane and it would get boring to read. On the other hand, making it clear the solvers are doing the right thing once in a while is very good for morale. Let's say the solver is instructed to pick up an item from the blue box, and the only item there is the Fiery Sword of the Serra Angel. You could say "Pick up the item in the blue box," and that's enough for the solver to know what they're doing. On the other hand, if you say, "Pick up the item in the blue box and wish you had your flame-retardant gloves," the solver can follow the instruction and knows things are going well. Of course, in this example, the solver might read ahead and know the Sword is going in the blue box. So you might want to be a tad more subtle.
Tell a story. Once upon a time, some people followed some directions and then the answer was BANANA. The end. Blah. I try to give the characters in my Duck Konundra personalities, which usually develop as I go. For example, in DK5, I noticed that Oscar randomly was grabbing a lot of objects early, rather than having them thrown to him. So I decided, okay, Oscar is a repentant kleptomaniac who will alternate between stealing things and feeling bad about it. This gives solvers a through-line, and it makes the Oscar player (if everyone is playing a character) feel like they have a niche. In both DK3 and DK5, one character naturally ended up a selfish ass, and another was a hapless victim. I'm not sure what this reveals about my psyche. Setting up expectations is also good; Cthulhu creates tons of dramatic tension in DK5, and in DK4 I love the player who draws a royal flush only to discover that under the rules for that round, his previous hand was better.
Find your voice. Writing a konundrum probably gives you more flexibility than any other puzzle, because while there are rules that apply to clues for, say, cryptic crosswords and abstract logic puzzles, here you can tell people to do things in any way you see fit. My writing strengths are wacky juxtapositions and dark comedy; those are the wells I went to when I wrote plays, and so I throw them in my Duck Konundrums. If you let your mind wander and write what pops into your mind, the writing's likely to flow better than if you try to force it. In a way, it's less like writing a puzzle and more like writing fiction, which probably makes it a much easier form for some people and a much harder form for others.
Hands off the duck. That's my calling card, thank you very much. (The funny thing is, I don't even have a thing for ducks. I was just trying to throw my initials in the name; if I'd been in a different mood in 1999, I might have written five Deer Konundrums by now.)
Feel free to share your opinions about konundra here. What aspects do you like? What don't you like? Are there cool examples you've seen outside the Mystery Hunt? What would you like to see or not see in future puzzles of this type?