One scheme to world-building

What makes this game fun? That's the question I want to ask when I've finished a game (or as I'm playing it). I recently re-played Fallout 2, and I noticed there's a lot of small quests which piece together to form a tapestry. Individually, the elements are not complex (deliver this item to so-and-so, talk to [person] about [subject]), but they compose together like lego blocks.

This post will discuss one way to world-building, which is necessary to coming up with quests and a game-plot. This is more a "review article" than anything innovative on my part.

Levine's "Stars" and "Passions"

Ken Levine has a great talk on "Narrative Legos". Basically, each location (village, fort, etc.) has around a half-dozen named characters ("Stars"), where each character has three "Passions". A "Passion" is what a character cares about relative to what the player can impact upon; it is transparent to the character, and responsive to the player's action. Effectively a "Passion" is a "bank account" for a Star (or a number between -10 and +10, starting at 0); helping a Star with their Passion results in that Star's invest positive points into their "account", thereby improve the Star's opinion of the player.

When a Star's opinion of the player (formed by combining the three Passion scores together somehow, e.g., adding them together, taking their geometric or harmonic mean, or whatever) reaches a certain high point, it unlocks certain bonuses. Blacksmiths offer additional bonus gear, Clerics offer additional services, etc. Conversely, if a Star's opinion diminishes, services cost more or are outrightly refused.

But two characters could have conflicting passions. This is a goal for writing a plot, because conflicting Passions for Stars means the player is thrust into a zero-sum game...which is fun for the player, and encourages the author to say, "Yes, and I can work this into another story arc!"

Example. Consider a simple fantasy setting. The player is tasked to fetch a McGuffin for the elf village. Elves hate Orcs (and Orcs hate Elves). A "Passion" shared by Elves is "Hurt Orcs".

Now, one of the Stars — Romeo the Elf — has a Passion: he loves Juliet the Orc. (Romeo the Elf's "Hurt Orcs" has the caveat "...except Juliet the Orc".)

For the quest (fetch McGuffin for the Elves), the player needs to take along a character from the elf village. Along the way, the player and companion are ambushed by Orcs, including Juliet the Orc. If the player takes Romeo the Elf, then Romeo will help kill the Orcs except for Juliet; the player is forced to choose what to do, kill Juliet the Orc, or run away somehow? And if the player takes along anyone else, then how does the player avoid killing Juliet the Orc?

Such is one way to have two different Passions interact with each other in a surprising way. It forces the game author to consider how to reconcile conflicting passions with "no wrong answer".

The number of characters should be around 5 or 6 to avoid over-burdening the player, and the number of passions should be 3 to avoid accidentally conflicting with too many other character passions (rendering the game unplayable accidentally).

There's a lot to digest here, and how we implement varies considerably. If we formalize a "passion" using a data structure (a few counters) and functions, we could [should] test the rewards and punishments are triggered properly upon shifting an NPC's opinion of the player. But this is just the observer pattern.

Exercise. Look at Fallout 1 and Fallout 2 (or any RPG you enjoy). For each town, write down who are the Stars and what are their Passions.

Begin with a Map and Needs

I've been playing a lot of post-apocalyptic games recently. I found it helps to begin with a map and asking myself, "How does society reproduce itself? Who produces what goods, and how are those exchanged among the cities producing them?" Looking at a real-world map, I can pick out a few cities that survived the apocalypse, think about trade routes, goods produced [at least food, water, armor, weapons, clothing], and this organizes and frames the thought-process about factions, their aims and beliefs.

These questions lead to more interesting power dynamics: multiple factions within a city disagree with how to distribute the goods, who to trade with, what to prioritize. How far are these factions willing to go to enact their policies?

Example. In one post-apocalyptic game, purified water is rare. Growing crops or raising livestock requires purified water. So for the survival of humans, purified water is absolutely critical (otherwise agriculture is impossible). But only one town has a prototype water purifier that can filter out radiation. As the only source of purified water, traders flock to the town to trade for the water. This makes the town a target for raiders, since traders regularly go to it. Fearing the raiders may damage the water purifier, the town has built sturdy fortifications around it...which has created an "inner city" within the fortifications, and an "outer city". One faction wants to keep outer city dwellers out as second-class citizens, another faction wants to keep the status quo. How far will the first faction go? Will they collaborate with raiders to drive out their rivals?

This "first draft" of an idea for quests sounds promising, but it has a problem: it's close to a "Jesus versus Hitler" morality choice. We need to make the status quo faction morally flawed somehow. There are many ways to do this — the status quo faction could support slavery, be violent imperialists refusing trade with cities that don't submit to their rule, etc.

Now that we have some idea of the factions and issues at hand in this town, we can apply Levine's "Stars" and "Passions" idea to further refine the idea. Each faction has a leader, but there's also the "head mechanic" in charge of maintaining the water purifier and its bottling, as well as the town doctor and the town trader. Thus we have our five stars for the town.

A town that produces nothing except coordinates trading (like the Hub in Fallout) has its own unique concerns. Just to rattle off a few:

• If a single partner (call it McGuffinsburg) is its sole source of McGuffins, then the internal politics of McGuffinsburg is a concern of the trading hub.
• Raiders are a perpetual problem.
• Power dynamics among competing traders; when there's little or no law, it gets quite cut-throat.
• Arms dealers in particular encourage secret agents to incite war between two neighboring powers, to profit from selling arms to both sides.
• Any of these could be reversed (McGuffinsburg exerting power over the trade hub, raiders as sympathetic figures, lawkeepers trying to maintain order, etc.)
• Any two or more of these could be combined.

One bit of advice: just as each Star has 3 Passions (for the sake of foregiving the player for acting unfavorably against one Passion without alienating the Star, thereby cutting off a potential quest-giver), we should take care to have several sources for each commodity. This is a rule-of-thumb, not a Law: sometimes, it's fun to have a single provider for a good, which causes tension and drives the plot ("We need to finish the [apparatus] to provide [goods] to save the world").

Remark (On Economy). At this point, it may be problematic determining the value of goods in the game. I'd advocate a Neo-Ricardian framework for determining the value of goods, if we are really concerned about a sensible economy. This works well for the settings I'm interested in, it works well. Each firm has its own methods-of-production and this determines the value of goods.

As far as the goods considered, we can consider every conceivable item in the game. This is hard. Instead we could start with some basic goods — grain and iron suffice for a first pass, we could successively add more as needed, like wood, weapons, coal, ammunition, etc. Traders add a markup to the value of the goods they transport in their caravan, which has been studied by Robert Vienneau.

Arguably, this doesn't need to be maintained. Prices could be determined as the sum of the cost of inputs, plus some markup, along the lines of Kaleckian economics...because if you aren't careful with your economy, your model may trigger hilarious/player-unfriendly results.

Note: although I have been thinking about a game in a post-apocalyptic setting, there's nothing preventing us from applying it to any setting. It's just a little harder if we have no map. (Post-apocalyptic settings let me be lazy, and use existing maps.)

Recap. So far, we have considered using the economy as a way to organize towns and factions. Each town produces some goods, and need to trade with each other to survive. This leads to identifying factions within towns, tensions between towns, and problems to be sorted out. It also leads naturally to using Levine's "Stars" and "Passions" to further refine the game. Both provide natural motivations for quests.

The map gives us a way to visually organize factions, consider how society sustains itself, the relationships between different towns and factions. Such considerations naturally give rise to Stars and their Passions. Altogether we have not formed a plot, but the fertile grounds for a plot as carried out through quests.

We have thus the basic process of world-building naturally give us ideas for quests and plot-lines. The "bottom up" approach with Levine's Stars and Passions combines well with the "top down" approach of drawing a map, determining the villages and towns, coming up with economies, inserting Stars and their Passions within each town, generating factions, and so on.

How real is the economy?

We need to decide how much detail the economy needs. This can serve a variety of purposes: just fluff, determine the actual value of goods, or as grounds for certain quests. Let's consider this last point in particular.

Missed Example 1. I was pleasantly surprised in Fallout 1 to learn the "gun runners" use historically accurate method of sulphur-less gunpowder production (using animal dung and urine). Although just lore and backstory, it could have served as a quest-line for players with high enough Science to suggest more modern methods of gunpowder production (leading to superior ammunition produced for the user, from gun runner traders alone).

This could also have led to the quest of finding sulphur sources (or other chemical sources). If using pre-war stockpiles, then there could be a limited amount of "superior ammunition" produced and provided to the player. If trading for chemical sources, then this could serve as grounds for multiple ways to obtain it: negotiate for trade, kill for obtaining it, etc. Perhaps there is some mistrust between the producer of the chemicals and the gun runners, the player needs to negotiate this (either by using a middle man caravan, thereby increasing the price of the superior ammunition, or by helping the factions bury the hatchet).

All told, I'm glad Fallout 1 didn't do this, because there's a time limit to the game, and this would have been a huge time sink.

As for modeling the value of commodities using economics, this has problems in the real world using textbook economics. As far as world building cares: value matters for trade routes, and for player bartering with merchants. For trade routes, we only need them to feel "about right" (e.g., goods expensive in one town is imported from a town where those goods are cheap, don't import goods when "domestically produced" versions are cheaper than the imported version, etc.).

Suggestion: Update the world map to reflect trade, specifically roads and ports are built (and improved) over time to better facilitate trade between partners. The quality of roads and ports reflect the trade power between partners.

For the player, accuracy can conflict with fun. When this happens, always side with fun (unless we want to create a constraint for quest lines, e.g., iron shortage causes more expensive equipment...motivating the player to, y'know, fix that shortage). The real concern is that the player has access to equipment matching the challenge. We don't want to give the player overpowered gear too early, nor do we want to force the player to have access only to mediocre equipment.

Economy for lore, well, this provides the grounds for quests. Town A wants to open trade with town B since B produces silk cloth. This isn't reflected in the goods sold in either town, but provides a new quest-line. Alternatively, if B were instead the sole source of iron, then town A could supply only, say, bronze equipment. This combines lore and player experience (which is good and desired: the player should experience consequences of their choices).

In short, put yourself in the world, and ask yourself, "What goods would I have access to? How would I get food? Water? Who produces them? How would I get them? What about luxury goods? Or equipment?" Answering these questions require us to consider the towns further, and increases the realism. It's not enough to have farms randomly placed around towns (Fallout 4 tried to do that): we must also consider the infrastructure, the shortages, needs, scarcities and abundances. This leads us to consider how towns interact, how factions within a town interact, and helps us build a world.

Generating History and Culture

History, Myth, State formation. If we consider how these Stars and villages interact, we can generate a history from conflict. Items of our Stars become revered artifacts. Music and art memorialize these events. Myths emerge from misunderstanding or deliberate lies. Rivalries build up, grudges between Stars and factions emerge over time. Villages band together, forming quasi-states, which dissolve under stress and strain.

Governments. It's worth noting that we could borrow liberally from history. For example, the Polish–Lithuanian Commonwealth had a unique form of government that is seldom discussed...it could easily generate problems to be remedied for plot-line. The ducy of Venice picked its leader through a lottery (well, a lottery picked an electorate, who then picked another electorate, and so on — the convoluted process of lotteries and indirect elections resulted in a new duke).

Religion. I don't have much to comment on here. Post-apocalyptic games tend to seldom discuss religion, and the fantasy games I play have similar pantheons. One thing worth considering, the original Rogue had "daemons" responsible for updating the health, etc., for players. I don't think anyone has made these daemons the Pantheon for the game, which could lead to interesting gameplay: praying to daemons leads to temporary buffs, destroying temples related to a daemon results in temporary negative buffs, etc. Or it could have the same effect as praying to the laws of physics (i.e., nothing noticeable).

Communication. And the most underappreciated point of consideration: it takes time to communicate these events. When an event occurs, news of it spreads through caravans and travelers. Spreading information takes time, and plans revolve around information. This is a challenge to code up, because we no longer have a simple observer pattern to update dialog and quest lines.