The value of narrative feedback for roleplaying

WARNING: Fallen London spoilers for the early game

A while ago while playing Sorcery! a thought occured: A significant part of roleplay might be the acknowledgement that you are, in fact, roleplaying. In TTRPGs this role is performed by the GM and fellow players, who will either acknowledge your decisions directly, but also account for them in future interactions with your character.

I.e. if you are a bad person initially, but make the effort to become better, everyone might acknowledge your attempt at a redemption arc. Or, if you make a snide comment towards an NPC, the GM might choose to refer to this at some point.

TTRPGs, in this sense, have a very broad space of potential action, wherein each action taken is both a mechanical action and a narrative one.

For most digital RPGs, however, this mechanical and narrative space is almost entirely split with some overlap in places, and where they do overlap they are usually limited.

Take Skyrim. You can do a lot in Skyrim. A lot. But almost none of the choices you make during play feel like narrative considerations. Usually you simply wander and act without much thought. Even if you did choose to roleplay (and I know some people do) the game does not help you in this.

You are the only curator of your character, the only arbiter of their consistency. If your attention lapses, the character deflates and leaves your carefully crafted mould of headcanon, and instead becomes the generic fantasy hero the game wants you to be. Because why wouldn’t your alchemist also rise the ranks of the Mage’s College, become the top thief, a feared assassin and also chief amongst the warriors? The game does not help you craft a character.

Now, Fallen London also suffers from this in part (as was pointed out to me in another thread). But, there is one storylet which is continuing to stick with me: The search for the Comtessa.

Initially you are hired by the Comtessa’s father to learn where her daughter went. The investigation concludes with you finding the Comtessa being converted into a clay man/statue and the clay man beside her claiming she wanted this. The game does not give you enough information to make a good choice, but you have to make one: leave the Comtessa to her (perhaps desired) fate, or smash her.

You make your decision, finish the quest line, done. You roleplayed, hurrah. However, this is where the genius part comes in: throughout the game, you keep getting the occasional callback to the choice you made. There’s no significant mechanical effect to the decisions you make here, but you are reminded. Over and over again, the game reminds you of the fact you made a decision, and you did so based on a hunch and what your character would have done.

This, I feel, is very effective narrative feedback. The game acknowledges your decisions without judgement, but I, as a player, feel seen. I feel like I’m not the only one crafting the story of my character.

I wonder then, what are other effective ways of doing this? I’m currently building an RPG of sorts, and wish to remind the player their roleplaying decisions are seen. But not just their option-menu decisions. No, I think a designer can turn simple mechanical decisions into roleplaying decisions as well. How? By means of narration.

I.e. if you choose to steal from a shop, the people might not know, but the narrator remembers and occasionally refers to this. You wander in the wilds for days? NPCs should start calling you mad, or express some other sentiment. You choose to join all the guilds? I think the guilds should have an opinion about that, one they really should be reminding you of.

Is that the right way to go about it? What are your thoughts?

4 Likes

Yeah. Any decision the game can record in some fashion can be reacted to by the game. I don’t know exactly what the format of your game is, but in a parser game, every command entered can be used as data for this kind of thing.

Is your narrator an actual narrator? Or do you just mean a narrating voice unattributed to anyone? Either way, all you need is an outlet for text reflecting on what the PC has done previously and how it’s affecting the present. The outlet could be a narrator. In my WIP, it’s the narrative voice of the PC themselves, since it’s written in first person.

-Wade

3 Likes

Keeping track of player’s prior choices can be also component of IF, and not necessarily of the choices. For example, in my WIP a puzzle starts in two different location, and albeit I coded it starting-point-agnostic, I can replace the boolean flag which avoid starting twice with a var noting where the puzzle started, allowing referring to where the puzzle actually started.

Recording the player’s choices and altering the narrative on the basis of it is the signature selling point of Dragon Age series.

tangentially, there’s those crazy “meta-RPG” (lacking better terms) like Knights of pencil and papers where you play the role of… GM, and the rather unusual Voice of Cards, where the narrative voice is in style of a GM.

Lastly, on the pure mechanical action, I can’t hide that I really like the “visible die rolling” in Baldur’s Gate III which (or whose) manages, at least for me, to bring back the tension on the RPG table when Alea jacta est

Best regards from Italy,
dott. Piergiorgio

1 Like

So, this is very much what I’d like to think interactive fiction in any form or shape does. Not only tell you a story, but acknowledge the choices you make throughout as well. However, I often find choice moments only extend to the moment itself, or if they have long term consequences, this is usually just by updating a visible or invisible tracker of some kind which the game gives no real indication is in play.

In other words, a relationship or morality tracker being examples of visible ones (Fallout, Mass Effect and Knights of the Old Republic do this) which, to me, often feel like ‘tainted’ roleplaying choices, or an ‘endgame decision’ tracker as an example of an invisible one. Baldur’s Gate 3 implements this quite effectively, where each companion can be influenced in some way to pick a certain endgame decision you are unaware of at that point. The problem is, once you’re aware it can be very tempting to just pick the “right” options.

Which brings me to Dragon Age, because while it did what it did very effectively at the time, I would say it maintains a very small narrative space you can actually influence, and while it has both visible and invisible trackers, the choices often still feel self-contained in the moment. In fact, many of the choices, while admittedly often more complex, lean on very black-and-white underpinnings of morality, which are then complicated by mixing them with obvious good consequences and obvious bad consequences.

Which, now that I write this out, is possibly another dimension to this. In some way, in order for those callbacks to actually feel meaningful, black and white morality is an obstacle. Instead, a decision like the Comtessa’s relies more on Blue-and-Orange/Bacon-and-Necktie morality. There isn’t an obvious good or bad. Both outcomes in that storylet are bad. Instead you’re very much choosing which bad. You are expressing your character’s values more than you are fishing for a particular outcome.

Also, yes, completely agree. Though I do have some ideas about dice rolls in digital RPGs which I started another thread about and their relationship to narrative agency.

1 Like

I suppose if you are writing in first person, with a strong characterisation of self, the PC as narrator could work. However, I feel that when it comes to creating roleplaying space, an external narrator is a much greater boon, since they can paint the situation for you, without taking agency away from the player in shaping the character.

Instead, I think a detached voice like this is much more effective at giving the player an opportunity to shape their character. By which I mean, the narrator describes a scene, and then prompts the player to assert how the character responds/feels/thinks/acts in that situation. This way the narrator does not run the risk creating a false justification of those character actions.

Just acknowledging what the player did before builds immersion, so that alone is great. But having those earlier decisions have (apparent) consequences, like affecting how NPCs interact with you based on what type of character you seem to be playing is a tantalizing idea.

Prior decisions could affect what solutions are available as you try to solve a puzzle. Consider a situation where you need information from an NPC who’s desperately poor. If you’ve developed a reputation as a kind, generous character, perhaps the NPC tells you what you want in exchange for a small handout, but if your reputation is that of a bloodthirsty sociopath, a hint of a threat might be all it takes. But turn it around, the killer you offers a handout, and the NPC runs away in fear. Or the kind you threatens, and the NPC laughs, knowing full well you’ve never hurt anyone.

In a long-term, it’s-never-going-to-go-anywhere experiment in parser fiction, I’ve toyed with giving a character their own mental model of the world. As they observe and experience portions of the objective world model, their mental model grows and changes. That has a few benefits.

First, it helps prevent inadvertent disclosure of secrets the character hasn’t yet discovered. The world model knows that black bird is a raven, but the character currently thinks it’s a crow. Until they learn otherwise, the narration refers to the bird as a crow, and the parser will feign confusion if the player prematurely refers to it as a raven.

Second, it allows the player to query their own model.

>WHERE ARE MY KEYS?
[checks scope in world model] They aren't here.
[checks player’s mental model] The last time you saw them, they were in the bowl by the door.

The response is correct, even if the player’s roommate has since taken the keys from the bowl.

Third, you can prime the player’s mental model with knowledge that they should already have. If the character works on a large farm, and they want to go to the barn, the character should know where it is, even if the player hasn’t yet seen it during play.

A typical world model has a “visited” flag for each location to note whether the player has been there before. The raven (oops! crow) knows which locations it has visited, but it doesn’t know whether the player has been to a particular location unless it actually observed the player there. My idea is to take this to extend this beyond “visited” to everything that to general independent models of the world.

Overkill? Almost certainly. But I think it could enable this idea of narrative feedback and consequences based on a player’s history of decisions.

4 Likes

Concur and agree. Indeed in the extensive, perhaps even surpassing Trinity, array of synonyms in The Portrait, a certain word related to civil status is absent… :wink:

In general, indeed the entire narrative of Isekai rotates around the PC’s perception (“mental model”) of the world, which changes often…

Best regards from Italy,
dott. Piergiorgio.

2 Likes

I wonder if there is an established IF mechanism/technique to determine your intentions for particular actions, like stealing or murder. Are you Robinhood or Jason Voorhees? Does it even matter?

I like to role-play in Skyrim. In a way, I think I like it that the game doesn’t assume anything about my actions. I know why I did what I did. I don’t really need an NPC to acknowledge it so long as there is a lot of freedom. But I think I would be turned off if the game reacted to my decisions without appreciating my intentions.

Food for thought.

1 Like

When you asked that, I immediately thought of The Baron. As I recall, it’s pretty linear, but with frequent choices about why you’re doing what you’re doing. Short of explicitly asking, I can’t think of a good way to do it.

2 Likes

My most disliked feature of Fallout 3 was the karma system. It forced you into an arbitrary good/evil system with no concern for intent or circumstance.

1 Like

In all fairness, real-life reputations seldom account for one’s intentions and even telling people why one did what they did can be ineffective at shifting public opinion of one’s actions. So, at least in that regard, NPC reacting to player actions without mechanism to account for player intention is truth in television.

Also, it’s discussing the subject in terms of a specific work of fiction(specifically, the Manga One Piece) and how it handles its character writing, but the following YouTube video I listened to last night goes deeper into the idea of a person’s reputation versus their actual character:

Of course, computer NPCs tend to be even worse at nuance than real people since the dev can only code reactions to patterns of player behavior they anticipate unless they’re willing to hand that aspect off to AI that can be unpredictable and likely to produce responses less coherent than those a human would write. And of course, trying to account for things like who knows what the player did(e.g. murder the owner of the general store and everyone in that town will probably know within a day, but people in the town on the opposite side of the mountain might never find out), who recognizes the player as the one who did it(e.g. if you were dressed in drab, ratty robes with a black hood over your face when you killed the store owner, people might not recognize you in your formal knightly regalia as the murderer), different NPCs having different reactions(e.g., the assassin’s guild might try to recruit you, the city guard might try to arrest you, the beggar in the street might beg for mercy), how much an NPC knows about the situation(e.g., someone who thought the shop owner was upstanding might condemn you, someone who knew the shop owner was secretly running the local slave trade or selling dangerously poor quality goods at premium prices to maximize profits might think you justified), etc. only adds to the combinatoric explosion a good GM can improvise their way through at the table, but which a computer game dev has to anticipate and make compromises on how well developed/responsive NPCs are and more often than not errs on the side of only a few NPCs being properly developed and having a limited range of responses, if not being entirely static.

2 Likes

There’s the classic dialogue menu format:

Yes, I cut down that cherry tree.
I never cut down that cherry tree!
I never cut down that cherry tree! (lying)

It’s not a declaration of motivation, but it speaks to intention in a neat way.

3 Likes

I was talking with a friend about this the other day, and I definitely agree that a game should not explicitely assume your intentions. But, merely pointing out that you spend a significant amount of time in the woods is already something worth commenting on I think.

Nonetheless, moments where you are asked about your intentions might be worth it. Now, I admit I personally struggle to roleplay in Skyrim as I believe I mentioned, but developing a reputation or a character explicitely rather than in my head feels like a valuable idea to explore.

Which means I might have to actually give this one a try to see how much of a ‘roleplay- experience’ this helps invoke. Perhaps linearity is not even that bad.

1 Like