Narrative Games Redux

If you're not familiar with the basic argument about the role of narrative (story) in games, this essay is not a good place to start. This essay is about a specific approach.

I'd like to talk about the idea of adaptive narrative. That is, a story which changes as you interact with it, and not with a dialog tree.

If you're like me, you've played Heavy Rain. If not, a quick summary is: the game has no real gameplay, it's mostly a giant branching path story where you try to figure out who the killer is. Spoiler: it's the one person who's definitely not he killer.

As games go, Heavy Rain is a bit of a bust, more like an interactive movie where you can choose which characters die, when.

If you're like me, you've played the old Blade Runner adventure game. There are several games where you are investigating a crime in something like real time, and the NPCs are also advancing in their own plots at the same time. Blade Runner is simply the oldest one I know about. Like most of these games, the way you can interfere with other characters' scripted timelines means that there are a wide variety of endings. Like Heavy Rain, it features not a whole lot of actual gameplay... but it's an adventure game, so that's par for the course.

If you're like me, you've played Catherine, known for its combination of box-pushing puzzle sections and fan service. If you haven't played Catherine, you may be unaware of its bar and nightmare plateau scenes, where you can talk to people, let time pass, and in general feel like you're impacting on their lives. Catherine has gameplay, but it could be replaced with any gameplay you care to name and nobody would notice: it has no relation to the story progression.

In my long history of failing to create adaptive narratives, I've tried many things. But I'd like to consider an approach based on these three games.

The core of the idea is that there are other characters, and they do things in some kind of (probably scripted) timeline. While time doesn't have to advance at a constant march, the point is that their stories advance when yours does.

There are three kinds of "inflection points" I'd like to consider. That is to say, there are three ways I'd like to let the player interact with the NPCs.

1) When the NPC's timeline brings them into collision with the PCs timeline during an event sequence. This could be happenstance (they rob a bank you happen to be at) or it could be that one of you is seeking the other out on purpose. Either way, the interaction offers a clear chance to change the trajectory of one or both characters. This is a common method used by branching plotlines today.

2) When the player, in his free exploring, sticks his nose into an NPC's life. AKA "sidequest syndrome". To be honest, I don't like this this method. Most games that use this method do not use it to radically alter anything, they are primarily subquests or only affect very minor characters. For example, the romance plots in Mass Effect, or the way you can try to free the slaves in any Jedi game ever. This is mostly distinct from type 1 in that they rely on the player exploring to the right physical place and then putting his life on hold to do some subquest.

3) Lastly, a subset of NPCs (and the player) can gather into one place when they are "off duty". These are quiet times in everyone's plot lines, and serve much the same role as a safehouse in a horror game.

In these sections, you can talk to a lot of different NPCs, perhaps in phases as the "off duty" time wears on. You will get to know them better as characters, and form a bond with them. There is also the option of interfering with their plotlines by giving them items, advice, or introducing them to other NPCs.

It's this "off duty" section I'd like to see more of.

Right now, most of our adaptive narratives are set around the idea of giving the player an illusion of freedom while, at most, offering them two choices. Basically, if you can imagine it, the current method is a lot of parallel lines that occasionally jitter slightly to one side or the other as the player bumps them.

The "off duty section" method is more like if those parallel lines converge - three meet here, five meet there. Then they separate and go back to being parallel. This has a lot of advantages, in theory:

A) Keeping up with the Joneses. Because the off duty sections are reliable as a heartbeat, the players can get updated on how the plot line of each character is progressing. It serves as a regular and centralized way to keep up to date with everyone. This allows for many plot lines to be active at once, and for a much longer period of time, rather than the short bursts of sidequests most games use.

B) Deflection as gameplay. By bringing everyone together, you can weave the plot lines in a regular fashion, such that the momentary lulls appear at more or less the same time for every character. The player then has to choose which characters will receive his resources and help, and in what way. So you take an ordinary set of simple branches, and suddenly there is choice and gameplay.

C) Closed-world. This system works great in a non-open-world game, where the player is not allowed to randomly wander into the NPCs' normal lives. In this way, a lot of the scenes and activities of the NPCs can be "just talk", rather than actually being represented in the game engine or with cutscenes.

That's my thinking. You?

Tactical Depth

I've been having a fairly in-depth conversation among my G+ game design circle that's got me thinking about some of my instincts. In this case, I am making a casual tactical game, and rejecting or accepting various suggested improvements.

One of the things which arises fairly frequently is that I reject something that sounds reasonable. For example, because this game features retiring characters when they reach a certain level, you'll have characters in your roster that are all across the level spectrum. A level 30 knight and a level 3 pistolier.

To make it so these characters work together well, there are two basic options: a variety of techniques which allow characters to work effectively together even with low-level characters, or a mentoring system where the weaker character is artificially improved while within a few tiles of the higher level character.

Of these, I choice the mentoring system. I did this not because it had the most depth, but because it had the highest depth to headspace ratio.

What I mean by that is that it's very easy to remember "weak character within 2 tiles of strong character", and the resulting dynamics are quite interesting.

On the other hand, even though the tactical options may be even deeper using work-together abilities, they require a lot more memory. Now the player has to remember "strong knight A can click on weak knight B to enhance stat C if within D tiles", and "strong pistolier E can stand near weak knight B to automatically fire on enemy C if they attack using ranged weapons..."

In a casual game, you have to keep the headspace as low as possible, because people will only play it for short amounts of time and won't be able to track so many highly varied pieces. Because of this, I've cut a lot of complexity out of the game. The idea is to keep the core play tactically deep but not using up much headspace: two players who have never met should be able to simply swap cell phones and play each other's battle scenario from the midpoint and immediately understand most of how the team operates together. No difficult-to-grasp swarms of details like you get in Disgaea or Civilization.

"Deep play, not complex play." That's my motto here.