Design Discussion (3/2017): Nonlinearity

[Ferdielance]

Unofficial Case Design Discussion of the Month: March 2017

What is the Unofficial Case Design Discussion of the Month?

It's a place to discuss specific challenges we face and decisions we make when we design cases.

Currently, to the extent that AAO members discuss design, it's limited to private chats, if it occurs at all. I propose to address one topic each month in a dedicated thread.

How does the discussion work?

Each month, I'll post a problem or challenge in the form of a short post followed by a series of starter questions. I won't give my own thoughts until a few people have replied. You don't have to answer the starter questions directly, though! They're just there to get discussion started!

Ground rules:

1: Do not cite fancases as examples of a bad design decision unless 1) the author(s) have given permission to open up their cases to critique in this thread, or you are the author yourself, AND 2) you can make a positive suggestion as to what that case could have done differently. It's easier to criticize than to solve problems, and we don't want this thread to descend into some kind of public shaming ritual.

2: Commercial cases (such as the AA games) don't require permission from the author for constructive criticism here. However, the rule that no problem should be cited without a positive suggestion still applies.

3: However, it is encouraged that you point out fancases that have handled a design problem particularly well, and set an example we can learn from. There is no need to ask permission for this.

4: Avoid gratuitous off-topic content. For my own part, I promise not to spam pigeons.

5: Of course, all of the usual forum rules apply, too!

Topic for March 2017:

Nonlinearity

The canonical Ace Attorney games tend to be highly linear, especially in the courtroom. There are few alternative paths, and while one might rearrange the order of events in an investigation slightly, the trials tend to be tightly restrictive. In a few cases, there are different ways or orders in which evidence can be presented, but this is not usual. Even the most striking player choice of AA, the final decision in 2-4, results in relatively little changed dialogue.

In contrast, some visual novels offer much more extreme branching. Some of the earliest mystery VNs, such as the Portopia Serial Murder Case (1983), include relatively free exploration and endings. Kamaitachi no Yoru (1994) (Banshee's Last Cry in the Aksys localization) goes so far as to give the player many different chances to solve the mystery. The longer the player takes, the more convoluted and bloody the mystery grows.

However, nonlinearity necessarily involves a tradeoff. Branching creates a great deal of extra work for the author, and highly branched works are harder to test. The developers of Ace Attorney were quite clear that pressures of time made true nonlinearity in an AA game unlikely.

But that restriction doesn't apply to AAO...

Some Possible Starter Questions:

* Would you like to see more nonlinearity in AAO cases?
** What benefits might nonlinearity bring?
** Are there any major drawbacks to consider?
* What sort of nonlinear structures might be worth considering in AAO cases?
** Example: One extreme nonlinear design would be to branch a game very early into three entirely different plots. A less-extreme one would be to have key decisions at multiple points, and track the player's choices to affect the ending.
* Have any AAO cases used nonlinear design especially effectively?

[TheDoctor]

One problem I can think of is the frame limitations. A full length case is going to be somewhat difficult to make non-linear because of the number of frames involved in creating branching pathways. If you can create a case, make it non-linear, and somehow keep it under 5000 frames, that's pretty impressive.

[Blackrune]

For trials, I think they're most efficient if mostly linear. Adding too much nonlinearity could require adding evidence and information that's only relevant in one path.
Especially if they converge back together later on anyway, it could be difficult to keep both paths relevant and of equal value.
Of course slightly changing the order in which things happen around a little is usually not a bad thing. It could prevent players from being frustrated for pointing something out a little earlier than they "should" have, although I don't think they'd feel satisfied if getting the solution early causes them to skip the entire game.

[Gamer2002]

Canon AA games are linear. That birds version of AA does have some non-linearity, but I didn't play it.

For AAO it requires a lot of work to make a game that involves multiple paths. You can, if you plan things out, prepare dialogues and events that happens on more than one path, direct the player to the prepared section and direct him back to his current path. But things can get messy and everything becomes really hard to playtest properly.

But you don't have to either have one, always same, path, or multiple different paths. You can keep both benefits of linearity (easy to plan, implement and test) and giving player meaningful decision (variety of playthroughs).

Typical AA scenario - Phoenix, for whatever reasons, has to decide which one, and only one, of two evidences he should take with himself. In canon AA you either make the good pick and use the evidence to proceed further, or make the bad pick and later return for the good evidence. This decision doesn't change the story at all, it only possibly halts your progress until you learn that the wrong evidence is useless.

But let's do things differently. You can't correct your decision later on, you keep your picked evidence. For fairness, you can hint earlier which evidence is more useful. But besides that, the player picks his evidence and progress further with the investigation - without anything immediately apparent making him realize if it the choice was right or wrong.

That way the player made a decision that will impact a future outcome, but currently doesn't affect his progress at all. If the choice was good, it will be easier to catch the killer. If the choice was bad, you will have to face some consequences of that. Maybe the wrong evidence also can help you, but in a lot more tricker way? Maybe you will have to ask for emergency investigation of the place where the other evidence remained? Or maybe you'll get a bad end? Of course, something as big as bad end should allow the player to have some degree of fixing his mistakes. Maybe the player will get one more chance to swap the evidences? Possibilities are endless.

But regardless, that way you can have a game that does remain mostly linear, but still forces the player to consider own actions.

[Tiagofvarela]

I'll start by stating that people are allowed to criticise any of my trials, which at the moment amount to one, one which is currently down for renovation.

I'll talk a bit about my meagre experiences with case development on this topic.
Nonlinearity was one of the key ideas I had for my fan case, The Laggy Turnabout. Ace Attorney cases have always been, exempting perhaps the order of investigations, quite linear. This is something that I wanted to try my hand at solving. As I came to find out, there's a reason they aren't nonlinear. Because nonlinearity is hard.

To start with, what is nonlinearity? I believe something is nonlinear from the moment you do not feel constrained by the limits imposed on you, and that you do not feel you are going down a train track set up for you. It should be an experience that kicks back when you kick it. It should react to the way you try to be oh so smart.

As an aspiring author, the amount of ways I tried to implement nonlinearity, and failed, are off the charts. In The Laggy Turnabout, there are some instances where the music changes depending on the choices you've made. Not only did I implement this in the most of asinine possible, such as adding each statement to the testimony twice, with a different music for each, trying to make sure all the press conversations led to the right statements (they didn't always), I ended up pretty much giving up on the idea. Not exactly an example of nonlinearity, but a pain to implement something this simple.
Another idea that didn't make it was my attempt at making the second part of the trial different, by changing who of the two members of the defence team actually defends. Technical issues such as attempting to garner the correct sprites and critical logic errors in my trial at this time made me give up on the idea permanently -- the idea I had for facilitating the divergence in the trial was made possible due to a logic error, after all.

What sort of nonlinearity did I manage to implement, then? The "cover up the plothole" kind.
A)Generally speaking, trials are not designed to have multiple solutions -- in fact, I'd say designing one in such a way would actually make it harder -- but if you're as bad at plotting as I am you will leave plot holes open. Plot holes you will find on your 17th playtest of the case. Essentially, the objective was to go through what I already had and try to come up with more solutions to it. More objections. As it happens, a lot of players interpret things differently, even if wrong. I had my betatesters tell me all their theories, even when they were wrong, and implement them into the game, as unique penalties or objections, for instance. Sometimes its even impossible to actually prove the wrong theory is, in fact wrong, which is when it truly can become nonlinear. Generally speaking though, I was able to wing any such instances and connect back to the main linear path.
B)Another way to add small measures of nonlinearity was to thoroughly study each prompt and present. In an ideal world, your players wouldn't present stuff at random, and actually have a logical reason to present something. If I can find those out, I can make the lawyer reflect the player's logic and become true avatars for them... gently refuting their logic with evidence pulled out of nowhere or something. Then all that needs to be done is assess whether the player logic is solid (no penalty) or if it has a flaw the player could have seen (penalty). Another way to simply do it is to just write random dialogue for each present, though that is time consuming and hard. Not even my masochistic self found it in him to implement my original idea to have each individual evidence present lead to a frame that mentioned what was presented and then transition into a generic "wrong evidence" dialogue.
C)The order and ways to present something also offer a measure of nonlinearity. A smooth nonlinear experience takes into account every moronic way your player goes about conveying himself, by not limiting to presenting a certain item at a certain location. Look for every roundabout way there is, and account for it. Sometimes they actually mean to present at the press conversation and not the statement. Sometimes it's the other way around. Sometimes you're forced to build some three different prompts just to make it all read smoothly without any weird connections. Like I was.
D)Finally, I think a very important point in nonlinearity, and by my standards the most important, is to have the case react to your player. Remember every choice he made. Bring it up later. Or don't. Depends on the player choices. Bash him for contradicting himself. Or for presenting the Attorney's Badge three times in a row. Or for pressing every statement like a poltroon.

The main drawbacks to creating a nonlinear experience are the amount of work it takes and the frames it consumes.
A) This method has the issue of requiring the new path to actually lead somewhere. Let's say that an alternate solution presents itself. Not only is finding all of them hard, sometimes the way you resolve the testimony is crucial for the next one. It's possible that the alternate solution won't let you transition onto the next testimony, or solve the case, leading to a dead end. Coming up with a way to make it work is often harder than just acknowledging the contradiction exists, but force the player back onto the main path.
B) Actually thinking up of every valid reason to present something is hard. It involves writing a bunch of extra content, that most people may not even read, for instance.
C) The main issue with implementing this sort of thing is testing it. You could have 20 players go through it and still create a broken path nobody notices. Trust me, I found one in The Laggy Turnabout, which has been released for almost a year, just a couple of days ago. Thinking up the ways a person could try to convey himself through is also hard.
D)Because you can actually lose at these games, some players are forced to reload, which can erase their previous variables. This makes a lot of different little touches be rendered useless, as the player will never actually get to activate them.
EDIT: It's also somewhat difficult to keep track of all variables, what they do and how you activate them, so there's that.

You may notice a pattern. A lot of this involves trying to understand and predict what your players will do or think -- and that's hard when not impossible. You'd need dozens of playtesters to truly encompass a majority of worthwhile alternatives. This leads me to a power fangames, and AAO, have: They can easily remain in constant change. As feedback comes in over days, months or years, new ideas and alternatives can be implemented. They don't have to remain static, as large games may need to these days.
As you read about someone's bokers theories and suggestions you can have a go at making the game react to them.

React. I find that is the key word that truly represents nonlinearity. Every choice you make creates a reaction. The scale of the reaction determines if it's nonlinear. Small dialogue changes versus actually changing the plot. Alternative ways to reach the same goal or one way to reach many goals.

By my definition of nonlinearity, simply creating a linear trial that at some point divides into three other linear paths is not a good example of nonlinearity -- even if impressive. I don't think I could create a trial with that many solutions.

[Blackrune]

D)Finally, I think a very important point nonlinearity, and my standards the most important, is to have the case react to your player. Remember every choice he made. Bring it up later. Or don't. Depends on the player choices. Bash him for contradicting himself. Or for presenting the Attorney's Badge three times in a row. Or for pressing every statement like a poltroon.

I did some of that in Curiosity of a Turnabout with the game remembering if you get something wrong here and there. Presenting the right thing afterwards can sometimes lead to fun. (like the protagonist saying "this is obvious", and someone reminding him that he screwed up completely just a few seconds ago)
I guess your "people will reload" problem with this is valid as most people wouldn't expect that getting something wrong will actually yield anything.
There's a non-AAO case series, Conflict of Interest, which also gives you a choice in one case that affects dialogue in the next case.
I wouldn't really call that nonlinearity, it's kind of just customizing/improving a still mostly linear experience. But it works well there.

[Acid Rain]

My thoughts on branching storylines has basically been summed up already, as I'm sure many people would agree. On paper, making a CYOA-type story seems awesome, but you'd quickly realize just how much work that would be, which is almost certainly why there aren't any (?) cases on AAO with branching storylines.

That being said, there are other ways to have non-linear gameplay aside from branching paths. For example, Zero Time Dilemma lets the player experience sections of the game completely out of order (as well as in different branches). Structuring a case (or literally anything) like ZTD probably wouldn't work very well, though, since ZTD had specific, complex reasons for having the game laid out like it is that would almost certainly not work in any other scenario. You could still use the concept in other ways, though. For example, you write a story that involves several different people's perspectives that are in some way connected. At some point in your case, the player is shown a "Character Select" screen where they would select a scene which would be told from the selected individual's perspective, and when the scene is over, the player is brought back to the "Character Select" screen to pick another part of the story to experience. In this way, the plot stays linear and doesn't become excessive, and the frame count would also stay basically the same; the only real change is in the way the player perceives the story. Depending on the order the player experiences the story, certain plot points may have more or less of an impact, and the player's opinion on characters or events may not be the same as they would be had they played in a different order. At first glance, this seems to be a way easier way to make your typical fancase "non-linear" than scripting multiple storylines.

Actually, has this been done before? If not, someone should steal this idea. It's nifty.

And for these topic in general, please feel free to use R E M as an example of good or bad game design. Not that I think R E M is very relevant to this particular topic.

[AceAttorneyMaster111]

I just thought of a really cool way to branch a trial: In a kidnapping case, the user plays as the victim/hostage à la JFA4. (This could also be extended to murder by playing as the victim pre-death.) However, unlike JFA4, the choices the hostage makes affect the game in some major way. This could be:

• Causing the kidnapper to kill the hostage because the hostage did something stupid like screaming for help
• Accidentally destroying key evidence, like causing a fire
• Escaping and being able to pinpoint the kidnapper
• So many other options here.

Pretty sure (I haven't fully read other responses yet) that the frame craziness this would cause has already been touched on, but this kind of fork would require a lot more work than a truly linear scene only giving the image of agency (RPG terminology) like JFA4.