Attribute Design

tim-cain rpg-design attributes game-balance character-creation fallout arcanum system-design

Abstract

Problem: How should RPGs design and balance character attributes — the permanent stats chosen at creation that define a character throughout the game?

Approach: Tim Cain draws on decades of experience designing Fallout, Arcanum, and other RPGs, reflecting on common pitfalls, clever solutions from tabletop games, and his own evolving philosophy.

Findings: Attributes are hard to balance due to range problems (too small = bonuses dominate; too large = individual points feel meaningless) and the fundamental issue of forcing choices before the player understands the game. Multi-attribute dependencies, permanent attributes with indirect bonuses, and penalties-vs-bonuses distinctions are key design levers.

Key insight: Attributes should be permanent with a small range, and nothing in the game should modify them directly — instead, items and effects should modify the things attributes affect, creating more interesting design knobs and better balance.

Source: https://www.youtube.com/watch?v=G_2kCbPO2sU

The Core Problems with Attributes

Tim identifies two fundamental criticisms of attribute systems:

Hard to Balance

Attribute ranges create a dilemma. A small range (0–5) means any in-game bonus is enormous — a +1 covers 20% of the entire range, and stacking two bonuses covers 40%. A large range (1–100) makes individual point investments feel meaningless — the difference between 17 and 18 strength is negligible. Tim recalls EverQuest's 0–1000 AC range where players would argue over a single point difference that was statistically irrelevant.

Blind Choices at Character Creation

Players must commit to attributes before they've ever played the game. They can't know whether their favorite weapon requires a minimum strength above or below what they've chosen. Even in replayable games, the first character faces this problem.

Penalties vs. Bonuses — A Subtle but Critical Distinction

Many modern games have replaced the traditional penalty/bonus model (stats start in the middle; going below gives penalties, above gives bonuses) with a "small carrot / big carrot" model where every stat level gives some bonus. Tim first saw this done deliberately in World of Warcraft and confirmed it with the developers.

This matters more than people think. Consider heavy armor that "negates dexterity bonuses" — in a penalty/bonus system, a clumsy character still suffers their penalty in heavy armor, making low dexterity meaningfully different from high dexterity even when bonuses are negated. In a bonus-only system, this distinction vanishes. Tim considers this the kind of subtle design lever that many designers either don't know about or deliberately ignore.

Balancing Strategies

Multi-Attribute Dependencies

Skills and derived stats should depend on multiple attributes, not just one. If dexterity feeds into many things, make its contribution secondary. Example: Disarm Traps could use both perception (+5 per point) and dexterity (+2 per point). This lets a "popular" attribute touch many skills without dominating them.

Tim admits his own dexterity designs in Fallout and Arcanum were overpowered — the same mistake he'd criticized about GURPS.

Self-Balancing Constraints

Several mechanical approaches help:

  • Higher cost to raise than to lower attributes
  • Limits on low attributes (e.g., only one stat below 8) to prevent dump-stat abuse
  • Shape-based allocation — instead of rules, give players a geometric shape (triangle) they can rotate, where the constraint is built into the visual. This appeals to non-number-oriented players while keeping the math underneath.

The Amber Diceless Auction

Tim highlights a brilliant tabletop RPG mechanic: in Amber Diceless, players receive 100 points and bid on attributes in a group auction run by the GM. The highest bidder for "strongest" pays their bid and has fewer points for remaining auctions. Players might intentionally drive up bids on attributes they don't want, hoping to drain rivals — but risk winning an attribute they didn't actually need. Tim finds this ingenious but acknowledges it doesn't translate to single-player computer games.

Going Attribute-Free

Some designers advocate removing attributes entirely. Tim doesn't oppose this but insists it requires a robust perk, trait, and flaw system as replacement:

  • Perks for positive capabilities (e.g., "Strong" increases carry weight)
  • Traits for trade-offs (e.g., "Big and Clumsy" — more strength, less dexterity)
  • Flaws for negatives that come with compensating benefits

Each subsystem is self-balancing: buying a perk means not buying another; taking a flaw grants something in return.

Concerns About the Fallout Series

Tim expresses direct concern about the trajectory of the Fallout franchise:

  • Fallout 3 removed traits (self-balanced character creation choices)
  • Fallout 4 removed skills (leaving only SPECIAL + perks)
  • Fallout 5 might remove SPECIAL entirely, leaving only perks

He sees a troubling trend of eliminating meaningful character creation choices.

Tim's Ideal Attribute System

If designing today, Tim would use:

  • Small range, permanent attributes
  • Nothing in-game modifies attributes directly — items and effects only modify what attributes affect
  • Example: Power armor doesn't give +Strength; it increases carry capacity. A hydraulic arms mod lets you meet higher minimum-strength weapon requirements — but the mod has weight, reducing the carry bonus from the armor itself

This creates interesting trade-offs with many tuning knobs rather than flat stat inflation.

The Value of Imperfect Heroes

Tim closes with a personal design philosophy: he loves flaws, weaknesses, and low attributes. Playing a hero who isn't perfect is more fun, more challenging, and more memorable. When games remove attributes or make all stats bonus-only, they lose this quality.

References