Thoughts about challenge design in IF

These are great questions. I think there is a great deal of “illusory” choice in all IF, and have commented myself that great parser fiction doesn’t have to model every possible input, but only the input that the player will type. So the trick is to guide the player by the nose into typing what you want them to type, without letting them realize you’ve done that. And part of doing that is creating authentic motivations for the player. They feel motivated to solve a puzzle in one game, or to respond in a way that is consistent with the nature of their character in a more high-brow “literary” narrative.


So far as the question of “what is fiction” goes, it’s been a big tent in print media since at least the 1960s. Like difficulty/challenge, it is a collection of things. Their individual importances vary by work. When people say “actually, Infocom’s interactive fiction wasn’t really fiction,” they’re holding 80s video games to a stricter standard than many 20th century authors and readers did traditional, printed work.

Re: experiences of failure, I don’t mind them as long as a) the rules are clear b) significant replay is not required c) it is actually interesting d) it is in keeping with the spirit of the game. Infocom often made deaths funny. While we don’t enjoy that so much today, I appreciate their efforts to maintain the overall tone and spirit of their work.

My interest in difficulty (or perceived difficulty, as I have said) is quite similar. I’m personally very bothered by puzzles aren’t a natural part of the game world. I don’t want them to disrupt my subjective sense of the game world by reminding me of their “gaminess” if that makes sense. The terms I’ve used in writing are “organic” vs “inorganic.”

The most egregious example I can think of is the opening half of Zork III compared with the second half. You start out with this intriguing, atmospheric region but wind up with knick knacks like the Royal Puzzle. To me it feels like something one might find on a coffee table. Zork III initially promises to be a subversion of adventure game tropes, but it doesn’t keep that promise. Another example would be the “coins on a scale” problem in Spellbreaker. I know some people enjoy these things, but they throw me right out of the game.

Even puzzle-heavy games can consist of very organic puzzles. Counterfeit Monkey, for example, is rather brilliant in this regard.

The thing I’m most interested in with my own WIP is having consequential choices without consequential fail states. You see a lot of that in choice games. The player is making decisions that shape a narrative. Figuring out how this works is a problem to solve, should the player want to, but everyone gets a complete story that isn’t truncated by You Have DIed or somesuch.


Expectation management comes into this. If you write a game called “Marooned on the island of a thousand deadly dangers” no one will bat an eyelid on ending the protagonist’s life a few times. If you go about it in this way you also keep the option of not conforming to those expectations without frustrating the player any more than they were in the first place. I for one only ever include the paths to the bad endings in the walkthroughs for my games to pleasantly surprise those players who find the good ones by themselves.


So many thoughtful and interesting responses.

Yes, my definition of “interactivity” was a bit narrow. I meant to add branching fiction as a way of achieving this – although it’s always going to be limited how many meaningful branches you can offer. And choices to e.g. style your character’s name, appearance, personality – that’s an interesting avenue, though not one I’m personally drawn to.

Mike, that’s interesting, your sailboat challenge example is something I try as well – I think of it as “vertical branching”; the story doesn’t branch horizontally in different directions, but the player can take the high road or the low road or the middle road to the same place, and the rest of the story has a different flavour depending which they take.

Dorian, I’m interested in your systems and terminology. If you don’t post them here, I’ll go searching for them!

“Songs in a musical” is a great analogy for puzzles in IF. I have the same emotional reaction to both – it’s rare that it’s completely seamless and that grates on me, except when it’s done well enough for me to buy into it, and then it’s magical.

The idea of making the player “feel smart” – I’ve been flirting with this idea, and it’s good to hear it from others. That gives me confidence to lean into it more. I want my players to succeed and act out cool, inventive, heroic actions, and I don’t want them to ever feel cruelly punished if they fail. But this leads to another question of mine (that might warrant a thread of its own): for the writers like me who like to insert challenges with consequences into their stories, how can we playtest our games to get the difficulty just right? I compare it to boardgame design, which I’ve done a little bit of, and which is a very iterative process. The first version of your game is always hopelessly broken and gets refined through many playtests. But with a boardgame you can send it back to the same playtesters again and again. With a game in a story you need virgin playtesters who don’t already know the solution… I feel like I would need at least a hundred (in batches of ten perhaps)… This is something I’ve struggled with in my IFcomp game.


This is a really complex topic and well worth discussing.

Different IFs, by the nature of the fictions they wish to express, will need to provide different answers. Options that can be valid in some contexts include:

  • Excluding interactivity other than moving the game forwards positively/not moving it forwards at all (that’s the type that would answer “Yes” to the “Is it true that to the extent we are invested in a story and its protagonist, we don’t want interactivity of this type?” question).

  • All happy endings (like the above, but without the “not moving it forwards at all” options. Even this can be made branching if there are different kinds of happy endings to experience).

  • “Failing forwards” (a type of IF where getting something wrong will recount the failure and segue smoothly into whatever part of the narrative follows on from that failure). Note this can be in an essentially linear or heavily branching IF.

  • Using state tracking so that failing can lead to different narrative outcomes later on (this can be extended to different types of outcome depending on why the failure happened and the surrounding circumstances).

  • Having a protagonist who is intended to fail at some things, and must fail to eventually succeed

  • Having a protagonist who is intended to fail at some things, but with sufficient flexibility in the program that the player can effectively “choose” where (and sometimes if) those failures happen

  • Having “bad endings” be a big part of the point of the game (perhaps to the point where the story only makes sense if some of them are experienced)

  • Diegetic hinting, where the hint system effectively adds another layer to the story. This may appear at the end of some failure paths, insert itself into the puzzle process or wait for the player to decide the hints are wanted.

  • A framing story which turns the failures into part of a broader narrative

Illusion of choice can be applied to any of these options. Light-hearted tones can carry off narrative-failure-less games more easily than heavier-toned ones, though some of these techniques would allow the latter to do this also. AI-generated branching fiction would need to have the same considerations as the manual variety, or else the choices generated could themselves reduce investment in the story and protagonist.

Another important thing to note: different degrees and tones of failure are possible. Not every failure need mean the narrative comes to a stop (although it’s hard to avoid this in games where many options are available at a specific moment, such as some parser games which large vocabularies). However, weaving meaningful failures into the narrative can strengthen the story significantly while also boosting interactivity and fun.

For Budacanta, I went with “protagonist intended to fail, but with enough flexibility so where and how can be selected”. (There is no “if” - without some failures, the narrative and characterisation would not be plausible). The nature of the story did not lend itself to a protagonist who succeeded at everything. The choices presented are what make sense to the protagonist, who has character-based reasons for these not always matching the options the player would consider in the same position. There’s also plot reason why most failures connect to the rest of the narrative without stopping it in its tracks. Yes, it’s a lot of writing and many players will never see half the text. However, it allows for a lot of options when integrating failure in a way that makes sense to players employing a variety of approaches to a given choice. It definitely has undo-and-retry and there are multiple ways to go through the early phases of the game (which is just as well as the early phases are all that are complete so far).

Every other option on this list likely has examples of IF that would count, and it wouldn’t surprise me if there were authors on this very thread who had used every single one of the methods in my list when authoring their IF over the years.


You’re pretty tough on death for a Resident Evil player :slight_smile:


1 Like

Yes Alianora, I see it as a whole toolkit of devices for the writer to create interactive experiences in IF. In The Princess of Vestria I used:

*Illusion of choice (hopefully sparingly enough that players don’t spot it!)

*State tracking (affecting A)future interactions/challenges and B)the story’s epilogue)

*Failing forward (but always with consequences via state tracking – including a “luck” counter that will run out)

*Ability to style/personalise the character, (a little, mostly via the epilogue)

*Death/capture/other catastrophic failure + forced rewind

*In-game hints

*Branching narrative (rarely, and always quickly rejoining the main narrative)

*Different ways to “win” at some challenges, including the game’s ending

*Timed puzzle

With this last one, I’m wondering if it was necessary to make the puzzle timed. A regular puzzle is also interactive, if not as dramatic. Is there a special term for a puzzle where the narrative is paused at that point until you’ve solved it, like in point-and-click games?


Ha! And FromSoftware games, too! Well, my unfinished game does have 40 unique deaths and fail states… so far! I enjoy a good death provided the conditions are right :sunglasses:

The first instance of an IF game using player actions to drive a global “clock” is Ballyhoo (1986), I think? I’d love to know if there is a specific term for it. I would likely call it “Ballyhoo time,” to go along With Zork (or ADVENT, if you like) time and… Deadline time is a bit clunky… perhaps just simulated time for a constantly ticking clock.

You might make a new thread with a poll: how do you feel about timed puzzles? It’s a good question!

1 Like

I was just about to make a comment about soulslike games here as well. Thing is, soulslike games, and a lot of modern roguelike/roguelites as well, is that they really don’t have a “fail” state. Or certainly not a “hard fail” state.

An awful lot of ink has been spilled over puzzle design in IF, but I think one of the fundamental underlying problems is the design assumption that the critical path through an IF title must be narrow and linear. Why not a game where the player can “fail” at literally every puzzle and still reach an ending other than an abrupt “game over”?

In very simple terms this can devolve into the “true” ending versus various grades of “bad ending”, but there’s nothing inherent in IF that demands that this must be the way it is.

1 Like

A lot of this depends on what “failure” is - if failure is death/game over screen it’s not so much a choice as a dead-end branch. Even if the game says “Do you want to go back and maybe not make that choice that ends the game again? Because it looked like you had a choice but you really didn’t…” That done repeatedly is infuriating. Especially if it’s something like I’m the detective solving the mystery but I can slip on a banana peel and die (and it’s not a game about pursuing the Banana-Peel-Killer…)

If failing means something else interesting happens, then that is agency and interactivity. It depends on how deep the author wants to implement the world. If you want to give players opportunity to make bad choices, hopefully the results of those bad choices aren’t just “And, you’ll need to load your last save…”

If failing means you miss lore or a side branch of the story but keep going, to me that’s catnip for a re-play. (If the player doesn’t make the decision to re-load from a save point.)


Lotta thoughtful responses here. Continuing on from the discussion of agency & failure - you can have total agency while never having failure, if you want.

Let’s say you’re looking for a job. You interview around, get a couple of offers at two very different companies. Is picking one or the other a failure? In some sense, yeah, you might consider it a failure if it turns out to be a bad fit, but in an authored setting, you could just as easily write that each job is a great, fulfilling, satisfying job. So, I mean, agency doesn’t have to be contextualized in a strict pass/fail sense.

Regarding “a crossword at war with a narrative” that seems most applicable to the adventure game style of puzzles and gameplay/story segregation; I think many games, especially choice games, simply have no crossword element, but remain interactive.


One of my favorite design ideas was from Storynexus - this is mainly in context of a QBN, but I like the philosophy:

Punishing Characters

Punish characters – they’ll love you for it. And it’s good for introducing a bit of strategy to your world.

When something goes wrong for a character – usually failing a challenge, but also choosing something that has a significant downside – you should mark that. You could take away some of their progress or resource qualities. Or you could introduce new qualities that are a problem and need to be managed (like Nightmares or Wounds in Fallen London). This is good, because it gives a sense of risk and reward. And it allows options like spending resources to reduce these menacing qualities.

It also means you get to be quite mean to characters if these problems get out of hand. That’s often fun to write, and it provides scars and war stories for characters. They won’t remember when you took a few actions of progress away from them, but they’ll remember being thrown into the weasel pits.


The formal name for games whose notion of time is like Ballyhoo “clock-based game”, where events are based on game clock. There are a number of games that have this feature. One reason it’s not more popular is that one has to be careful to allow enough margin for people to have reasonable opportunity to succeed at timed objectives - something that has proven difficult to do in a manner perceived as fair.

1 Like

I think you’re thinking of Deadline and the like, hence my suggested “simulated time” (Or, more fancifully, quantum or indeterminate game world). The player absolutely can miss a lot of stuff in those simulations. I don’t think people make many of those games anymore. Ballyhoo only advances the global “clock” when specific actions are performed by the player, for instance frightening the elephant IIRC. I don’t think there are any timed objectives in Ballyhoo, other than maybe the endgame and a few other encounters. however, those are based on local timers rather than on a global, ticking clock. I think Ballyhoo time remains a common approach. It was quite popular in the old Sierra games, though the triggers were often frustratingly obtuse.

I’m sure that someone will chime in if I’ve misspoken.

1 Like

That is also a type of clock game, it is simply that in Ballyhoo more of the actions don’t move the clock forward (there are some that don’t move it in Deadline either). The breakthrough Ballyhoo made was in designing a sufficiently forgiving and coherent design that the things that move time only move relatively inconsequential things - unless some action happens that move the plot forward substantially. Before this, “clock time games” tended to be either very simulationist (like Deadline and lots of Sierra games) or take out the clock altogether and just have things be entirely event-based with perhaps a decorative clock for setting evocation purposes (like a large proportion of IF involving time changes in the “independent revival era”).

1 Like

Just to understand: I can’t think of an instance in Ballyhoo where an independent clock changes anything, substantial or otherwise. Can you think of an example?

I can think of timers, but those are localized. There’s no global clock. A player might, upon entering a trailer, only have x number of turns to complete an action before being kicked out. But that’s a closed circuit. It starts upon entry and stops upon exit.

The clock is an illusion in Ballyhoo as far as I know, a rationale for global changes in the game world. I admittedly haven’t played Ballyhoo in a while (my playthrough is only through AMFV), but I can’t find evidence of an independent clock in my transcript.

E: sorry if this is off topic


I do think it’s useful to distinguish between “you need to be in the right place on the right turn count or you’ll miss it” (Deadline, Suspended, All Things Devours, Varicella, A Change in the Weather) and “time doesn’t advance until you achieve an objective in the game” (Anchorhead, Vespers, Counterfeit Monkey, it sounds like Ballyhoo also though I haven’t played it, and I think Plundered Hearts).

It’s not an IF example, but the second one is also what we see in games like Bloodborne, where most events happen in real-time (every second matters in a fight), but the sun doesn’t set until you beat a certain boss, and the moon doesn’t rise until you first visit a certain area. In other words, there’s a clock that’s affected by every action, and a separate clock that only advances with progress in the game.


Maybe I’ll make another thread. It’s an interesting subject.

1 Like

I think the first time this was highlighted in the post-Infocom IF world was Christminster. I think of it as Christminster-style time, anyhow.

(I played Ballyhoo before that but I don’t recall noticing the technique as a design idea.)


I’ve toyed with an IF game set over a particular time frame - namely, midnight to around 4 AM on New Year’s Day - that does have timed events and such, but it is very much a ‘play multiple times’ type of game. I usually dislike the sort of game that you can’t conceivably ‘win’ in a single playthrough, but the idea is that it is more about exploring a world (in the loosest sense, it’s set within the downtown of a city) and uncovering the ‘quest’ in the process. If you played to the end or died, the end message would be something like ‘… but that was not how the story went’, until you got the ‘actual’ story itself.
On a much more minor scale, I was also toying with how to implement having a jukebox and nightclub mention a particular song playing depending upon the time in the game.