Replaying Diablo

Abstract

Problem: What can modern developers learn from revisiting a 28-year-old action RPG?

Approach: Tim Cain replayed the original Diablo (1996) in 2024, analyzing its design decisions with fresh eyes and decades of industry experience.

Findings: Diablo's combination of procedural dungeons with handcrafted encounters creates lasting replayability. Its deliberately minimal feature set β€” no dialogue trees, limited NPCs, simple UI β€” proves that restraint in scope can produce a timeless game. Playing older games teaches developers to distinguish lasting trends from temporary fads.

Key insight: You don't need every feature in every game. Diablo shipped without conversation trees, deep NPC systems, or accessibility options β€” features that existed at the time β€” and it didn't need them. Studying old games reveals which design patterns endure (trends) versus which were momentary (fads).

Source: https://www.youtube.com/watch?v=AH0jo26TQ_c

What Still Holds Up

Tim revisited the original Diablo (640Γ—480, released 1996) and found it holds up remarkably well:

  • Visuals β€” still beautiful despite low resolution
  • Music β€” the original soundtrack remains excellent
  • Voice acting β€” small NPC cast, but fully voiced
  • Minimal HUD β€” just a bar at the bottom, with pop-ups for inventory and character sheet
  • Fun factor β€” still genuinely fun to play nearly three decades later

Procedural Dungeons with Handcrafted Encounters

The 16 dungeon levels are randomly generated each playthrough, with themed tiers (dungeons, caves, fire levels). Monsters, named creatures, and quests are also randomized, creating enormous replayability.

However, certain encounters like The Butcher use a hybrid approach: the boss always appears on level 3 in a recognizable, hand-designed room β€” but that room is embedded within a randomized level. You don't know it's coming until you stumble into it. Tim calls this an "interesting combination of procedural and handmade" and considers it a design technique worth studying.

Procedural Wonkiness

Random generation creates occasional oddities: stairs from level 7 depositing you right next to the level 9 stairs, empty chests, empty rooms. These are acceptable trade-offs for the replayability gained.

The Quest Reward System

Quests are randomized (except the final two). Completing a quest guarantees a unique item with up to six bonuses, compared to regular magical items that cap at two. This creates strong incentive to pursue quests.

The catch: there's no guarantee the reward suits your class. Tim's Rogue (archer) received a sword for completing the Anvil quest for Griswold β€” useless for a bow user. He sold it for 13,000 gold, which created three inventory stacks he had to dump on the ground in town, where it sat safely because nobody ever steals it.

Dated Design Decisions

  • No dialogue trees β€” NPCs talk at you; you can't respond, decline quests, or make choices
  • Gold overflow β€” gold stacks cap at 5,000 and consume grid inventory space, forcing players to dump piles on the ground in town
  • No accessibility options β€” no colorblind settings; item colors (white, blue, gold) were Tim's only real differentiator, and even he saw them as "white, gray, tan"

Why Play Old Games

Tim's central argument: restricting yourself to recent games limits both fun and craft development.

  • Playing only new games teaches you about fads β€” features that are popular right now but may vanish
  • Playing a range of old and new games teaches you about trends β€” what came and stuck
  • Fads are nearly impossible to predict; trends are much more predictable
  • If a design pattern from 5–6 years ago is still widespread, it's probably a trend worth adopting

Feature Restraint

Diablo proves you don't need every feature. Conversation trees existed in 1996 β€” Diablo didn't use them and didn't need them. Every feature should earn its place.

The Fallout Anecdote

When Diablo launched in late 1996, Fallout was still a year from shipping. Marketing came to Tim demanding he add real-time combat and multiplayer to Fallout, pointing at Diablo's success. Tim's response: he pointed out everything Fallout had that Diablo didn't. When marketing insisted, he did "what any good producer would do" β€” he drew up a new schedule and budget reflecting their requests and submitted it for approval. That was the last time anyone mentioned real-time combat or multiplayer for Fallout.

References