Perfect Is The Enemy Of Good

tim-cain game-development production optimization feature-cuts fallout outer-worlds arcanum temple-of-elemental-evil

Abstract

Problem: How does pursuing perfection in individual game features harm the overall quality of a shipped game?

Approach: Tim Cain draws on decades of experience across Fallout, The Outer Worlds, Arcanum, and Temple of Elemental Evil to illustrate the real costs of over-polishing single features at the expense of the whole.

Findings: Time spent perfecting one feature directly steals from others. The most successful approach is maintaining consistent quality across all features rather than having a few perfect elements surrounded by underdeveloped ones. Feature removal is harder than producers think, early optimization is usually wasted effort, and knowing when "good enough" is good enough is a critical production skill.

Key insight: A game where everything is consistently good will always be better than one with a few perfect features and several bad ones.

The Phrase and Its Meaning

Tim first encountered "perfect is the enemy of good" around the mid-2000s while working at Carbine Studios. The core idea: time spent making one feature perfect takes away time from making other features good. It's typically used as a warning against over-investing in any single feature when others aren't at parity.

Sometimes, though, Tim has seen it used as an excuse to cut features prematurely when teams can't reach the desired quality bar.

Why Removing Features Is Harder Than You Think

Tim pushes back on the notion — common among producers — that removing a feature is as simple as deleting it from the schedule. In RPGs especially, features are like threads in a fabric; pulling one unravels others.

The Crafting Example

Consider removing crafting from an RPG:

  • Crafting materials become purposeless junk items that only exist to be sold
  • Recipes found in the world reference a system that no longer exists
  • Item economy breaks because crafting was one of several sources for best-in-slot items at various levels
  • Rebalancing is required — those items need to be redistributed to merchants or quest rewards, which risks letting wealthy players buy top-tier gear too early

The lesson: cut features early, before they become deeply interwoven with other systems.

The Outer Worlds: Fallbacks and Feature Cuts

The Outer Worlds is Tim's go-to example of managing this tension. Many features were either dropped entirely or reduced to simpler fallback implementations during development. Tim maintained a large list of fallback features — simpler, faster alternatives ready to swap in when a feature couldn't reach the target quality level or couldn't be made bug-free in time.

Monarch's Lost Jungle

The planet Monarch was originally designed to be four times its shipping size — a lawless jungle planet meant to be a sprawling open sandbox for exploration, combat, weapon testing, and monster hunting with heavy respawning. It kept the "lawless" feel but lost the jungle and the scale, costing an estimated five to ten hours of play time in exploration and combat alone.

Early Optimization: A Programmer's Trap

Tim identifies premature optimization as a classic manifestation of the "perfect is the enemy of good" problem:

Speeding Up Code That Doesn't Matter

  • Optimizing functions that are rarely called has negligible impact on the game's performance profile
  • Using sophisticated sorting algorithms on arrays that will never exceed five elements is pointless — just walk through and find what you need

Reducing Memory Footprint Unnecessarily

  • If a function creates a temporary array, does calculations, then releases it to garbage collection, that's fine
  • Keeping a global array to avoid reallocation is unnecessary optimization if the function isn't called frequently, doesn't use much memory, and the garbage collector handles it well

Fallout's Footstep Perfectionism

Tim's most vivid cautionary tale comes from the original Fallout. The team spent enormous effort ensuring walking animations didn't "float" — that footsteps connected to the ground without sliding, a common problem in isometric games. This required:

  • Specific rendering approaches with particular focal distances and lens settings in Maya/Alias
  • Custom algorithms for how art was cut and placed in the game
  • Careful coordination between Tim and the artists

The result was something Tim estimates 95% of players never noticed, especially since contemporary games like Diablo didn't bother and nobody cared.

What Was Lost

The time spent on footstep perfection meant Tim never implemented a feature he had on the back burner: smooth straight-line walking. Instead of characters jerking along hex-by-hex paths (especially noticeable when moving vertically), clicking a destination would walk the character in a smooth line toward it. Characters might clip slightly through doorways or scenery, but the visual improvement would have been significant.

Tim notes that Temple of Elemental Evil solved this elegantly — 3D characters walk smoothly toward destinations with proper rotation, looking far better than Fallout's hex-locked movement.

Arcanum: The Feature-Packed Cautionary Tale

Tim calls out his own game Arcanum as an example of the opposite failure mode: a game "packed full of features" that never got polished enough. Too many things were being made perfect while others never even reached good. The result was a game full of underdeveloped systems.

The Takeaway

Whether you're a programmer, artist, or designer: you'd rather ship a game where everything is good — consistently good — than one with a few perfect elements, some good ones, and some bad ones. Knowing when to stop polishing and move on is one of the most important skills in game development.

Source: Tim Cain — "Perfect Is The Enemy Of Good"

References