Content As Reward

tim-cain game-design player-agency procedural-generation content-design rpg-design

Abstract

Problem: Games rarely offer substantial content that reacts to player choices — why is "content as reward" so uncommon despite being beloved by players and designers alike?

Approach: Tim Cain defines content as reward, explains its appeal, identifies the economic and production barriers that prevent its adoption, and proposes procedural generation as a practical solution.

Findings: Content as reward — where major game content unlocks based on player agency — creates deeply reactive experiences but dramatically inflates production costs, since developers must build content most players will never see. Procedural generation of encounters, dungeons, followers, loot, and quests can deliver similar reactivity at manageable cost.

Key insight: The real enemy of reactive, player-driven content isn't design philosophy — it's economics. Procedural generation, done carefully, can bridge the gap between the reactive games designers want to make and what budgets actually allow.

Source: https://www.youtube.com/watch?v=ZpGbO9qk4qw

What Is Content As Reward?

Content as reward means game content that opens up as a direct consequence of player choice and player agency. Not just minor mechanical unlocks like picking a lock because you have the lockpick skill, but major content: storylines, maps, companions, encounters, and entire quest chains that exist only because of how the player chose to build and play their character.

Tim gives examples of what this looks like in practice:

  • Exploration-driven content — stumbling across a cave with no quest leading you there, just because you went off the beaten path
  • Skill-gated content — master-level skill unlocks or perks requiring significant point investment that open up new gameplay
  • Companion availability — companions whose recruitment (or departure) depends on player attributes, skills, and story choices
  • Branching story content — choosing a faction and getting unique content as a result

Tim reveals that The Outer Worlds originally had more companions, but some were gated behind player choices — your leadership skill might be too low, or you'd chosen the Board path and a companion would refuse to join (or leave the party entirely).

Why You Don't See It Much

The core problem is economic. Content as reward means content some players will never see, and that's expensive.

Tim lays out the math with a simple example: suppose 20% of your game branches into three paths based on player choice. The base game is 80%, but you now need three versions of that 20% — totaling 140% of a normal game's content. You've just committed to building a game-and-a-half.

The costs compound across every discipline:

  • New maps need level designers, lighting, encounter design, testing
  • New NPCs and companions need voice-over recording
  • Every branch needs full QA coverage
  • Perhaps 80% of players will never see a given piece of branched content

When games cost a couple million dollars, going from $2M to $2.5M was tolerable. When games cost $100M, asking for $150M to make content most players won't see is a very hard sell to producers, directors, and publishers signing the checks.

The Solution: Procedural Generation

Tim acknowledges this won't be popular with everyone, but argues that procedural generation is the practical answer to making content as reward economically viable.

His vision spans multiple game systems:

  • Encounters — procedurally generated based on player actions. Did a prison break by blowing a hole in the wall? Escaped prisoners now appear in encounter lists. Opened a gate to the Nether world? Demons everywhere — until you close it.
  • Side dungeons — randomly generated caves and complexes to discover through exploration. On replay, they change. If one is locked and you lack lockpick, you can keep exploring and find something else instead of hitting a dead end.
  • Followers — instead of being limited to six hand-crafted companions, players could visit guilds and recruit procedurally generated followers matching their needs ("I need an illusionist," "I need a fighter with thief skills who can backstab").
  • Rare loot — randomly generated with characteristics that might perfectly suit your particular build, creating exciting slot-machine moments on every powerful kill.
  • Quests — once dungeons, encounters, and loot can be generated, quests follow naturally. An NPC says their grandfather had a treasure map, or their mother disappeared into a dungeon — the dungeon, its encounters, and its rewards are all generated to match.

The Arcanum Precedent

Tim points to Arcanum as proof this approach can work. The game used procedural dialogue generation, and he believes the same careful approach applied to broader content systems could produce deeply reactive games that feel personal to each player's choices.

The Counterarguments

Tim directly addresses the skeptics:

  • Some players find procedural content sterile and repetitive
  • Some developers share those concerns — and find procedural generation threatening job-wise
  • It's easy to procedurally generate things, but hard to procedurally generate good things
  • Hand-crafting content and shipping a linear story with slight reactivity is far easier

But his core conviction remains: if developers invested in careful procedural generation, the result would be far more interesting, reactive games — games that genuinely reward players for the characters they build and the choices they make.

References