Why can't i get into CYOA or restricted parser games?

I’m just sort of curious whether anyone else loves general parser-style IF but has never found a choice-based twine-style game to be entertaining. I think for me it comes down to choice-based games having an even harder time casting an illusion of choice for me. When there’s literally only one thing to click to advance the story about the half the time I click, it ends up feeling a lot less like choices and a lot more like a stylized way to turn the page.

So, I guess I’m curious whether it’s more common to feel the same about both genres or more common to strongly prefer one to the other.

I’m also curious about whether anyone has cast a systemic/academic eye at the differences between the two genres. I haven’t thought about it but I assume they’re more or less functionally equivalent, right? Obviously, I could implement a CYOA in Inform and though it might get a little awkward, I suspect I could model all of Inform with a point-and-click choice-based interface. But it seems like there are many differences in they ware they are actually used. Like if I looked at the top 50 Inform games and the top 50 twine-style games, what kinds of differences would i find? I’m thinking of things like overall length, tone and genre. Number of puzzles. Types of puzzles. That kind of stuff. Does anyone know of any work in that vein?

Puzzles, I think, benefit from a feeling of originating the solution yourself; the best puzzle games try to make the players feel smart for solving the puzzle, “on their own.” Point-and-click adventures can undermine that feeling by enumerating the options (IMO regardless of whether they’re P&C text adventures or P&C graphical adventures).

But if you’re designing a puzzleless game, there’s much less to be gained from the parser and a lot more drawbacks. Authors can enumerate options on the screen to offer the player actions that would be infeasible in a parser game, especially options that are wildly different from turn to turn, options that take weird/unknown amounts of time, and social actions. (There’s a reason many parser games fall back to a choice-based or restricted-parser interface for NPC dialogue.)

So, if you love puzzles, you may be let down by restricted-parser games, choice-based games, or anything P&C.

I have mostly fragmented thoughts about this, and I’m sure others can pipe in much more coherently. All I can give are example games…

but Twine allows the author to alter the appearance of the text relatively easily, so choice-based games, while removing the illusion of choice, can easily use text modifications to aid in conveying the story. Things like the text moving away from you when you try to click it, or a change in background colour, may well aid in, say, conveying a sense of hopelessness or signalling a change of setting. This is especially evident in games like Compound Fracture or My father’s long, long legs.

You’re absolutely correct in that either format can be used to emulate the other; It was a Dark and Stormy Entry (ifdb.tads.org/viewgame?id=wqrlx3wvtmkjds9q) and (apparently - I haven’t got round to it) Fair, an entrant in this year’s IFComp, both use choice-based input couched in a parser form. On the other hand, we have works such as The Axolotl Project and 16 Ways to Kill A Vampire in McDonald’s, which are both Twines, but use a location-based world model, such as you would find in an average parser game.

1 Like

Parser systems are, by default, a parser plus a systematic world model. A systematic world model is well-suited for puzzles because it lets the player explore combinations and work out the system.

To build that kind of puzzle game in Twine you have about how you’re going to implement a systematic model and then make that kind of exploration available. Which is possible, obviously – it’s just not built in.

I enjoy both parser-style and CYOA-type games in general, but there’s a certain subset of CYOA games that I almost always find deeply unsatisfying. I only really figured out why recently–until then I would have only been able to describe it as ‘Twine games, but not all of them’. Turns out it’s related to the way choices are presented to the player. You might put them as explicit choices at the end of the paragraph, like so:

The air in the old house is thick with dust, and the ancient wood creaks and groans. Mice–at least you hope it’s mice–scurry around in the walls. As you step inside, you spot a bright line of light beneath the closed kitchen door.

A> Investigate the kitchen.
B> Investigate the noises.

This the way that a traditional Choose-Your-Own-Adventure book, or a Choicescript game, would do it. It’s also possible to do it that way in Twine. But a lot of Twine games do it like this:

The air in the old house is thick with dust, and the ancient wood creaks and groans. Mice–at least you hope it’s mice–(scurry around in the walls). As you step inside, you spot a bright line of light beneath the (closed kitchen door).

–where the parentheses indicate hyperlinks leading to the same passages that the choices above would lead to. In terms of the world model, it’s functionally the same, but something about it destroys the second-person illusion of taking part in the story, to me. Maybe it’s because what the hypertext will lead to isn’t always clear, or maybe because it breaks the flow of reading to have to read the passage once and then go back and look at it as a whole. I’m not sure what it is, but it’s the reason I generally skip almost any game marked as ‘Web’, because 95% of the time my only reaction is ‘oh, it’s another one of THESE’.

It is in fact actually possible to do some really cool and creative puzzles without a parser, or with a very, very limited parser. Detectiveland is a good example from this year’s competition–the underrated Grandma Bethlinda’s Variety Box from last year’s is another (it used a parser, but the range of actions was so limited it could have been implemented without one.) The best puzzles from these games made timing a factor–when there’s only so much you can do, requiring you to do it at just the right time is a way of adding complexity. The robber puzzle from Variety Box was probably one of the best puzzles of the year, and all the player got was a choice of three buttons to push.

With choice narratives, I have an ADD problem with links in the text. I prefer when clicking a highlighted thing provides more description, just like an EXAMINE command, and then doesn’t go anywhere. My problem is I read as far as a clicky thing and then I can’t resist before going on. Some games don’t give indication what interaction the click will be - more info, or going somewhere or picking up the bucket and putting it on your head.

I’m really enjoying making hybrid games with the Hybrid Choices extension - I get the world building and object handling of the parser, but I can switch to a menu to avoid guess the verb for complicated actions and for conversation.

Clearly communicating choices to the reader is an important authoring issue. I like this blog article by Ashton Saylor that he wrote about choices in gamebooks. He outlines some negative choices and some positive choices. For instance, a ‘Which Door’ choice is one where you don’t have enough information to make a choice, while the opposite end of the spectrum is a ‘Cake or Death’ choice where the best choice is obvious. The middle ground is where the interesting choices will be.

As to the flow issue created by inline hypertext, I am curious whether people think the interface of Texture solves this issue for games written using it. In Texture you make choices by dragging and releasing an interaction word over a hyperlink and these can be inline or end of paragraph. However, these hyperlinks are hidden and not clickable until you start dragging one of the floating interaction words at the bottom of the screen. While intended for a mobile interface, I think Texture’s interface does help a bit with maintaining a reading flow because interaction targets aren’t indicated until you start dragging the related interaction word, which you would be reading last.

I am reminded of the Adventure Book authoring system, which combined choice with a limited parser, of which Kingdom With End by Shannon Cochran is good example. It is a CYOA with a parser-level complex world model which takes good advantage of its dual CYOA/Parser nature. I’ve not played it myself but Jon Ingold and Emily Short both discuss the game here and here respectively. I find AdventureBook’s keyword inventory system to be very interesting because it provides for some out-of-the-box opportunities - if you have an item and think you are at a page where you can use it, you type it’s name in the parser. If you can use it, you are taken to a new page as if you had made a choice, thus allowing for discovery inside a CYOA game.

1 Like

This is a really interesting thread. I’ve written both types of game. I’ve tended to find that writing a parser-based game is more like world-building, i.e. geography based, whereas writing a choice-based game is more like screenwriting, i.e. history based.

Friends of mine who don’t really “get” parser-based IF are always encouraging me to convert my parser games into choice games, so that they can play them without the inconvenience of having to type or think about what to do next. But to me they’re chalk and cheese.

“Hybrid” games, whether choice-based games which attempt to emulate the map-and-inventory of parser games, or parser games which limit the functionality of the parser or make the range of available actions obvious, each seem to me to be working against their own strengths. I appreciate that they’re experiments, but playing them always makes me feel uncomfortable, feeling that the author made the wrong choice of medium.

I completely agree with StephC about the use of hyperlinks in choice-based games. In my own Twine games I use both choices-at-the-end-of-the-paragraph, in the form of lists, and links embedded within the paragraph, but I try to make it obvious that the links are choices of actions, rather than just highlighted words. I find the recent trend of adding a chunk of text to an existing paragraph really irritating, as it forces me to skim what I’ve already read in order to find the new information. I wish that authors using this technique would at least make the new text a different colour.

Dan is on point. It’s all about the “feel” of solving puzzles or the feeling that you have to find the puzzles before you can even begin solving them. You have to learn how to learn within the world model. It’s just an odd feeling you can’t get with anything else.

To me, a parser game can lend itself to a sort of “physics mode” that might exist in a platformer. Suppose a puzzle solution to get over a wall, and the author defines the PC must be eight feet off the ground to do it. The real solution is to find a rope and climb up to eight feet, but instead the player collects chairs and stacks them up to get eight feet high. That is something that would only happen in parser.

I thought OPPOSITELY OPAL was a great example of a game that only works in parser, since the player must learn how to cast existing spells backwards and discover the vagaries in doing so.

I’d like to see the apps who run parser games integrate hybrid concepts as standards. Some common conventions to expose room-specific hints (choices) and walkthrough that a user could choose to disable as they see fit.

I understand the down-side of a CYOA focus, that authors may not fully test all the possibilities for a free-form parser player - but I also think a standard convention for “hybrid” would bring more players into parser IF realms. There are whole generations of youth who may never learn to use the traditional keyboard. For that matter, handwriting may be on the way out :wink:

I’ve always thought parser games have a bit of a parallel with physics-based puzzlers: You can’t always possibly predict how the bits will fall, and unexpected surprising things can happen, even if they are bugs.

I experimented with programming NPCs by giving them objects that had code for tasks they would carry out. I wanted to test that a character would respond to a noise in another room but only after he had finished his first task. When I tested, I triggered the event and went to the room to wait for him to arrive. Somehow he had beaten me and was already there. Dammit. Was he teleporting somehow? I tried several times with the same result and thought there was a problem until I turned on ACTIONS. My character was taking a shortcut by climbing out the window and down some vines instead of the indoor path I expected him to take: he had instructions including “the quickest way from A to B” and had better knowledge of the map than I did! The NPC was carrying out the directions faithfully, but since it had not occurred to me he would optimize his path it was an eerie moment of seeming-autonomy.

I also LOVE when I read a transcript and a player figures out a puzzle solution I didn’t think of.

Spoiler for One Eye Open follows on this topic.

[spoiler]There’s a laundry chute in One Eye Open that leads to the basement. Once you get to the basement, you can unblock the elevator and then use the elevator to get between floors.

One of my playtesters observed that it was faster to jump down the laundry chute than use the elevator, and promptly set about jumping down the chute whenever he wanted to go from floor 3 to the basement. I had assumed people would always take the elevator once it was available, so this broke the game.[/spoiler]

It would have been a great optimization, if I’d only thought of it first!

I feel like that unpredictability is more of a feature of the kind of world model you’re using than of parser v. text, though. There are a lot of parser games that don’t particularly allow for unexpected solutions and unpredictable NPC behavior, because those aren’t programmed in.

It’s admittedly easier to get this going in parser games than in choice games at the moment, because as zarf said parser development systems come wrapped with world models that make these things easier to create and as far as I can tell existing choice-based systems by and large don’t (though Versu did, alas, and I haven’t taken a look at Raconteur). And parser games do, I guess, require a bit of systematicity in that any command can get entered anywhere, though probably the default way of accommodating this is that a given command doesn’t change the world state unless it’s entered under the right circumstances (“dig” tells you “You uncover nothing” unless in the room where something’s buried) or that it makes the same state change no matter where it’s done (“turn on lamp” causes the lamp to be lit, whether you do it in the Loose Crawl or the Treasure Room).

Surely you are right, Matt, but the parser allows you to model things more complicatedly than just what’s built in. That’s what I mean comparing it to a physics platformer. The author may implement a weight system and intend for the character to seek out a banana, an apple, and a grapefruit to solve a balance puzzle but not realize that the heavy idol the player already has can hold the pressure plate down by itself just fine. Happy accidents are part of the joy of parser games!