Zendo and the puzzle maker mentality

Zendo, by Kory Heath, is a good game. I recommend it. Go play it.

One participant creates a puzzle. The other players compete with each other trying to solve it. The puzzle-creator isn't the adversary. (The deets about how the puzzles work and terms of the player's competition with each other are beyond the scope of this post.)

The object of the puzzle-creator is not to make a puzzle so difficult that it can't be beaten. Nor a puzzle that's too easy. The puzzle-creator isn't competing with the players. She is challenging them.

Someone recently used the metaphor of puzzle-maker and puzzle-solver to explain to me something that I had thought was incoherent in trad roleplaying. The puzzle-solver can have the mindset of competition. Trying to beat the puzzle. Struggling.

The puzzle-maker has a completely different mindset. They can make an unsolvable puzzle easily. Once I played a very cruel prank on someone. We were sending puzzles back and forth. And I sent a square of random letters, numbers, and diamonds. A fake "puzzle". (The intended outtome I had in mind -- which in hindsight was unlikely -- was that the solver would realize that the only winning move was not to play, would see the Gordian cut. Of course than did not happen. Instead, it eroded trust in the implied social contract between solver and maker.) That was a mistake. I was very young, and I learned from it. The puzzle-maker's mindset is to find the sweetspot of difficulty. Goldilocks. The puzzle-maker "wins" if the puzzle is a challenge, is pleasant to try to solve, is instructive, is doable, is difficult, is possible, is fair

So that is how it is. The players "cheating" is different from the DMs "cheating". I finally heard a good argument for that. I enjoy solving puzzles. I enjoy creating puzzles. I had never connected that mentality -- those two orthogonal but sharply intersecting mentalities -- to DM and player in a trad or old school roleplaying game.

I had a mental model of a cheating DM to be similar to someone cheating in chess, someone who pretends to play chess with you but who actually just wants you to have a compelling experience. But. That's not what good chess teachers do. They neither play to beat you nor pretend-to-play. What they do is set up puzzles. Can you see the back rank mate in this situation? Can you see the distraction, the prophylactic move, the fork? This has truly been a missing puzzle piece for me to understand the trad trpg mentality.

However. However, the game of Zendo has rules for how the puzzle-maker interacts with the players and the puzzle. The puzzle consists of a secret "rule", and of "structures". The players use the structures to try to find the secret "rule". The puzzle-maker is allowed, once the game has started, to create structures that make the rule easier or harder to find. Again, with the goal of finding the sweet spot of difficulty, and trying to be fair to all players. They are, after all, competing with each other, even though they're not competing with you.

The puzzle-maker is not allowed to change the secret rule once play has started. That would break the game. It is not allowed to have no rule in mind, building new structures to disprove the players guess while continuously allowing many different variations of secret rule to be possible. It is not impossible to play the game this way. Eventually, and with great difficulty, one secret rule might become nailed down. Conversely, having no rule in mind but agreeing with the first suggestion by any player, would also break the game.

Both of these are powerful techniques, in Zendo, to achieve the puzzle-maker's goal of having the puzzle challenge be just the right difficulty and length. When I am the puzzle-maker in Zendo, I am generally happy if a particular rule goes around two or three turns around the table. If every participant first has some wrong idea of what the rule might be, but then when one player is getting close to finding what the rule might really be, so are the others. Having no rule in mind allows me to prolong the game as I wish, by continuously making the rule more difficult, and to end the game whenever I want, by declaring any valid guess -- any guess that isn't disproven by preexisting structures -- to be correct.

That, however, is against the rules of Zendo. That breaks the covenant between puzzle-maker and puzzle-solver. That… is cheating. The rules of Zendo are that the puzzle-maker does have one secret rule in mind, and sticks to it.

In D&D… I wear two separate hats. The prepper, and the runner. When I am running the game, I may not alter stats on the fly on the sly. HP, saves, number-appearing… the prepper should set those when she makes the world. The prepper makes sure the runner has a location, a living, populated location and the life, the population, has stats. Mechanics.

Similarly, in Zendo, I create a rule, trying to make that rule a good, fair challenge. I have some liberty once the game starts to build structures that are easy or hard. But I can not touch the secret rule.


  • This is 100% correct and well explained IMO.
  • I also think this is an excellent analogy. I agree with this mindset. I think it's one of the reasons modules are popular for this kind of play (and not so much for other styles of gaming): it makes that separation between the "prepper" and the "runner" easier and clearer.

    Having said that, I do have two minor nitpicks:

    1. It's interesting to hear you use the term "trad play" to refer to this. I understand what you mean, because I know your gaming history and interests. But some people use "trad play" to refer, rather, to 90s-style heavily GM-led (or even railroaded) adventures, a la Whitewolf or Dragonlance or similar fare.

    2. I don't think the "prepper" is always in that "looking for the Goldilocks challenge level" mode. It's also possible to prep in entirely unbiased or distant modes, more as a "simulationist", and then figure out what challenges the players will face later, in play. Sometimes that can be an interesting variable to play with!

    For a simple example, imagine drafting up dungeon which holds an ancient Egyptian tomb. The prepper interested in challenge might devise a puzzle or trap which attempts to suit the skills of the players - so, maybe if the trap seems impossible to avoid, this prepper will add something to the environment to give the players a chance (perhaps the corpse of an earlier victim!).

    The heartless, simulationist prepper might not care about that, though, and just do historical research, making the trap/puzzle/tomb as authentic as possible, not caring whether that makes it impossible to solve, too easy, or anything else.

    Of course, in practice, not many of us are purely one or the other. However, it's worth considering those two possible angles and being able to choose the right one for the right situation.
  • edited July 2018
    I definitely agree with nitpick 2 and with the point that modules make it easier. I run homebrew too but yeah, modules more often. When I made this point IRL, I made sure to emphasize both those things. You're creating an experience, not necessarily a fair, solvable, winnable challenge. Situations and challenges can and will arise in play and players tend to create their own. In the article that landed on the cutting room floor but I'm glad it got brought up in the thread.

    Point isn't that the prepper has Goldilocks mindset. Point is that prepper and runner obey different rules. (Clarification for new readers: "runner" means DM, who runs the dungeon, not the PC, who (in another sense of the word) is making a dungeon run.)

    Now let me explain nitpick 1, the trad vs old school thing. I have started following your usage, Paul, and using trad for 90s style game and old school for old school play. I know "trad" has another seme in "games with a GM/PC split" etc etc. Games that would encompass AW, V:tM, CoC and old school. I've written posts up to quite recently with that terminology, trad as a superset of neo-trad, 90s game and old school. But now I'm onboard with using trad more narrowly, to mean incoherent 90s game.

    So how come I used it here? Because as you know, I've seen trad as so incoherent that it makes my eyes bleed. And using the metaphor of the completely different mentalities of puzzle-maker and puzzle-solver -- which, as a puzzle-maker and puzzle-solver (not just in Zendo, I used to work for the Journal of Games and Puzzle Design) I can completly understand and relate to without any incoherence -- solves a lot of that. The trad GM is putting situations in front of the players and they then react to and in those situations. The trad GM curates the experience in a way that's anathema for me as a roleplayer but understandable when I view it through the lense of puzzles.

    However. And my article was split in twain by a raging river called "However". And I went on to explain how even in the world of puzzle design, there are rules and covenants. This is where I brought it back to old school over trad. The first half of the article explains trad to me. The second half, then, hopefully explains old school to trad players. When Matt Colville and even Jeremy Crawford advocates changing encounters on the fly, I'm saying to them that now I do understand you a little better than I did a few weeks ago: you're not the opponent, you're the challenge-curator. You don't see it as cheating. But then I go on, in the same language of puzzles that helped me understand them, to explain why as a player, as a puzzle-solver, I do see it as that.
  • Guy, that's the sort of posts I want to see more of, from everyone, on S-G♥
    agreement with me♥
  • edited July 2018
    For the record, I was also agreeing! :)

    (And I have no particular attachment to either use of "trad"; I was, rather, hoping to clarify before anyone else decide to get into a fight over it. I'm fine with either use of that word, no problems here. I understand your dilemma in writing this and sympathize.)
  • Yeah, my text was unclear there. I used "trad and old-school" at one point. Confusing
  • edited July 2018
    It seems to me that, in theory, the game full of Goldilocks-zone puzzles (just the right difficulty) and the game full of simulationistic puzzles (whatever would be there per the internal logic of the fictional world) ought to be prepped, run, and played entirely differently.

    In my experience, however, not many of my groups want to play either of those extremes. Almost all of them want to play a mix. They don't want their sense of a plausible fictional world replaced by some transparently-rigged obstacle course, and they also don't want to grind away for hours on puzzles beyond their ability.

    Accordingly, I've gotten great mileage out of cheating at Zendo. That is, not always have one pre-defined rule which is the answer to the puzzle at hand. This allows me to tweak a puzzle that started off too easy or too hard, moving it toward the direction of satisfyingly fun. It also allows me to not sweat prepping for the Goldilocks zone (which, in my experience, is actually impossible -- I've never seen any GM churn out "just right" challenge after "just right" challenge).

    But wait! Doesn't my tuning every challenge mid-play result in that contrived feeling where a compelling fictional reality is replaced by a giant board game? Well, it would if my only rule was to tune every challenge for optimal satisfaction. Fortunately, I have an additional rule, which is to go in with a loose idea of the answer to the puzzle, based on firm rules about some of the relevant factors. The range of potential solutions may fluctuate, but the simplest logic behind them does not. The hard rule I refuse to break describes the fictional reality, not the puzzle solution. The puzzle solution then emerges from some combo of what that fixed rule allows and what seems like fun at the time.

    I respect the principled purity of process when GMs run pre-made one-right-answer puzzle challenges with complete objectivity. I just don't usually find them fun to engage with all the way to the end. As soon as the players stop making obvious progress, I'll advocate moving on to something else so we don't wind up in a grind. Because oh man have I played some grinds. Spending 2 hrs staring at graph paper is probably my least favorite use of an RPG session.

    All these trade-offs I'm mentioning are less necessary in Zendo, because Zendo has a specific procedure for testing theories and getting more information. Compared to an RPG puzzle, I think it's much easier to spot a Zendo puzzle that has stumped you: the second you struggle to think of a useful input to submit to Zendo's hypothesis procedure, you are now stumped, and continuing will be a grind. Perhaps more importantly, in Zendo you're less likely to get stumped in the first place, as the information to inform your hypotheses is unambiguous, and the process for testing them is unambiguous, and the feedback is unambiguous. When RPG puzzle design achieves this level of clarity, I think everybody wins.

    To move slightly in that direction, a friend and I have developed some in-fiction Zendo-like options for adventurers in Delve. When it comes to magical effects, you can string together some runes and see what the feedback teaches you, or you can sprinkle some magic dust and see what celestial symbols it forms. I haven't really brought this front-and-center in Delve for various other reasons, but it might be a viable way to approach every puzzle in a game with different goals.

    Would that be fun? Or is the RPG puzzle's appeal based in the very fact that (at least initially) it is ambiguous? That, yeah, solving it is part of the fun, but so is discovering it? "Here's one toolkit to use on every puzzle" might ruin that.
  • Have you literally cheated in Zendo or are you using a metaphor?
  • The second edition "Going Deeper"-booklet has some suggestions (p 16) for how to get out of grinds without cheating.

    Various puzzle systems have different ways to construct and run the puzzles -- crosswords are different from sudoku -- but the core point I'm trying to make here is a dialectic between

    A. Awareness that yes, there can be orthogonal-but-intersecting goals of "creating a good experience", a fair challenge etc on the dungeon master end, and "trying your best to navigate and succeed in the world" on the player end,


    B. An implied social contract on how the game is setup and run. The entire reason for me engaging with the magic mirror the way I did in the mirror story was predicated on assumptions based on how the labyrinth lord ran the world, how we as a group found out what was factual or not. In a "no myth" impro exercise I wouldn't have felt the way I did and I wouldn't have cared the way I did.

    That doesn't mean the DM was "out to win" or that the module writer was out to screw us. Weirdly enough, for all the insta-death traps in that dungeon that chopped our ten foot poles, killed some unlucky hired NPC scribes, was evaded by our care or planning, or directed towards monsters -- what always did kill us were when we got into combat, never the traps. The traps were difficult, lethal -- but fair.

    You said something a few months back about not knowing exactly how a puzzle in Delve worked but knowing enough about how Delve's magic worked to wing it. That disturbed me but I chalked it up to not being up to speed on Delve.

    Trying to solve a puzzle where the gamemaster is cheating is something I do not want to do. It seems to me like the least worthwhile use of my time that I have ever heard of. (And that's coming from me, someone who spends around 12h per day edition warring on various net fora, except the days where I run a local shikataza group where we literally just sit quietly and do nothing else, not even watch paint dry.)

    What's the appeal from the player side playing Zendo like that? Also it's against the rules of Zendo. Structures (koans in first ed) must be self contained and the rule can not refer to outside factors such as play length, room energy, grind emotion etc.

    Perhaps we're misunderstanding each other.

    For D&D, I don't advocate goldilocks puzzles. Our answer to the mirror was definitely not in the book. But the mirror's properties, that we used, were.

    Having cool interactable things/people/places in the sandbox is what I'm after. Not one-planned-outcome. I never want a particular outcome.

    Idk, you are confusing me now.

    I described two ways of cheating in Zendo, called them powerful because they allowed you to set the pacing and goldilocksness of the session with much greater control, but condemned them. They break the convenant between solver and maker.

    Have you played Zendo (as a solver) that way?

    I addressed -- and the Zendo rulebook linked to above adresses -- the 2h grind. That strawdoll is irrelevant here. There is always the option to go somewhere else. In the same campaign that mirror story happened, there were three levers that we never figured out. We just left them and gave up.
    There is no "forward". That's such a key to running D&D.
  • Accordingly, I've gotten great mileage out of cheating at Zendo. That is, not always have one pre-defined rule which is the answer to the puzzle at hand.
    OH MY

    David, I respect that people have different preferences. If we ever meet in person and you run Zendo with the possibility of cheating and you don't tell me up front that the possibility exists, I will still respect that people have different preferences, but since I'm telling you this now, I will no longer respect you. Sorry, I'd like to, but that's what'll occur, it's how my mind works, and I'm not interested in fixingchanging it. That's like... to me, that's like The Unforgivable Sin. Like giving someone makework and fooling them into thinking it's important and letting them build their identity around the "important work" they're doing.

    Here are two other options.

    1. Let me know you might cheat in order to increase everyone's fun. This is fine because it lets me stop playing immediately, no harm no foul.

    2. If one of your rules turns out to be too hard, apologize as soon as you notice and then give everyone the option of skipping this round. If they agree, reveal the rule, have a brief discussion about what quality of the rule made it way too hard so everyone learns something, let everyone frown briefly at you and then laugh together, and start a new round. If they disagree, let them be challenged and frustrated even though it's uncomfortable because "you caused this", and ask them again after about 1.5-2.5 more times around the table.
  • I'm open to the notion that I might have misunderstood Dave here
    Perhaps he doesn't mean literally

    Guy, I feel like I'm on an isolating path here. I'm pretty much never going to be able to play since I see so many other DMs as "sinners".
    And I've made mistakes on my own in that regard. Even this current campaign has been compromised:
    The party was heading to a jungle camp that their previous characters had visited. There had been enough PC deaths that this party was new. I decided that the camp would've been overrun by zombies. I wanted some change, I was looking through crosshairs…

    After the session, I felt like I had cheated. I should've at least rolled for it. Ideally, evolving the sitation at the camp -- which the party had barely explored during their previous visit -- with finer granularity and more randomness -- would've been better. The group threw me a curveball by going there again.

    I definitely broke my DM ideals there.
  • edited July 2018
    Also again, Zendo second ed (and the rulebook is free online, linked upthread, for those who have first edition) has more options for what to do when the rule turns out to be too hard.
  • Metaphor, guys. I've barely ever played Zendo, and never been the rule-maker. I thought this thread was about RPG puzzles. All I meant is that the thing 2097 described as cheating at Zendo is in fact an effective way to GM puzzles. I guess "cheat" was a distracting phrase there, because that implies a violation of formal rules, social contract, or both. What I was talking about is neither of those things. It's basically just "GM doesn't have one right answer in mind". No need to lie to the players about that. In my case, I just make puzzles that would actually by everyone's understanding of the fiction plausibly have multiple answers.
  • Yeah that's fine. Multiple answers, none particularly in mind, that's great. Get everyone on board with playing one game, actually remove their agency and force them to play another, that's not.
  • If there is a magic mirror or an endless rope or an immovable rod or a bottomless cup in a D&D game that is a sort of puzzle without One Outcome, part of the toolbox/sandbox style of play that I love.

    I don't understand how Delve's magical puzzles work so I shan't judge them
  • edited July 2018
    Get everyone on board with playing one game, actually remove their agency and force them to play another, that's not [fine].
    Of course. I don't think I've ever heard anyone on S-G advocate for that.

    I keep assuming we're past that and forgetting the need to clarify.

    In describing the way I play, I mainly wanted to drop a note about some upsides of incomplete prep. Like, I think "multiple answers" isn't controversial in itself, but "none particularly in mind" can be. The process of resolving to a specific answer (or failing to) is a bit more nebulous in that case. The GM rolling with what the players throw at them can make it harder for the GM to be impartial (as opposed to a prepped answer). Converging toward "what is the truth" can also be awkward for groups who prefer total transparency. If you opt not to see behind the GM curtain, then emergent puzzle solutions can appear seamlessly, but if you really need to see behind the curtain at all times, not so much.

    Not sure if all that is important to this thread or not.

    I do think that looking to Zendo for ways to tweak RPG puzzle designs and GM-player interactions in RPG puzzle solving is a neat direction to explore. My runes and dust are vaguely partway there, and have been a lot of fun, though often not all that clear.
  • My orig point tried to be: the Zendo moderator ("master" in 1e) is not trying to "win" vs the players but there is still "cheating". (They can make choices that help them reach one of their goal — goldilox game length — at the cost of their other goal, game integrity, and at the cost of the player's goals.)

    This was metaphor not for a particular type of puzzle design or even play agenda but as a counter argument to the trad crowd out there saying that "nothing the DM does constitutes cheating". Cheating just means something different for DMs and players. And different from game to game (and every table has their own "game" when it comes to D&D).
  • Oops. Apologies for the tangent, then. I latched onto totally different parts. Maybe I'll start another thread at some point about all the other cool stuff I see embedded in there. :)
  • Dave, I do think that's an interesting and relevant post but instead of thinking of it in terms of "answers" or keys that will singularly unlock a puzzle, think of it as "what are the facts". A Zendo puzzle doesn't presume a particular sequence of structures to be built. That part of the game is emergent and where the life of the game is. There are multiple routes to solving it.

    I usually don't prep "wallpaper", knowing I can wing that stuff. But sometimes that kind of color gets misread as a puzzle and we get into a weird situation. I had a magical effect on a building look like some sort of out-of-time running water. And that was a clue to the nature of this place and color and improvised -- when the players started actually interacting with the effect I knew I had messed up and should've prepped it.

    This reminds me of the woman with the secret map from our discussion years ago, remember? She'd look at her map secretly. You seemed to see it as "How will the players find her map?"
    I saw it as… she just has a map. She has these values, these experiences, this fear. I try to not think of outcome. I put the pieces on the board and let the players loose, ideally (and sometimes I mess up when it comes to reaching that ideal). If the players don't find the map they don't find the map.
Sign In or Register to comment.