Summary
Randomness is a design tool — neither inherently good nor bad. Hiwiller uses the analogy of salt: sometimes you want a dish salty, sometimes it ruins it. The designer’s task is to understand where randomness serves the experience and where it does not, then use randomness mitigation techniques to calibrate the skill/luck balance their audience expects.
(Hiwiller, Players Making Decisions, see source-players-making-decisions)
Key ideas
The skill/luck spectrum
Games sit on a spectrum between two poles:
Completely random (LCR, War, Baccarat) — the winner is determined entirely by random process. No skill can improve outcomes. Famed sociologist Roger Caillois wrote that in games of pure chance, “destiny is the sole artisan of victory, and where there is rivalry, what is meant is that the winner has been more favoured by fortune than the loser.” Adults find it difficult to derive satisfaction from pure-luck victories because the outcome was not settled by anything they did.
Completely skill-based (Chess, competitive Darts) — the better player should always win. Even here, total determinism is impossible: Chess’s Elo system acknowledges that a novice (1400) will defeat a master (2000) approximately 3% of the time. Full skill-based purity removes drama and creates an insurmountable gap between skill tiers — if I know I have no realistic chance against an Olympian in a sprint, why race?
Most games sit between these poles. Blackjack adds one decision to pure-luck Baccarat — whether to take another card — and that small injection of skill creates a substantially more satisfying and popular game. Magic: The Gathering is skill-dominant but randomness provides enough chance for a weaker player to win 9% of the time against a highly-rated opponent, which sustains hope and engagement across the player skill spectrum.
Why randomness is useful
- Drama: Unexpected outcomes sustain tension. A game where the better player always wins is ultimately solved.
- Replayability: Random elements produce different game states each session.
- Inclusivity: Randomness compresses the effective skill gap, allowing weaker players to remain competitive and hopeful.
- Pacing: Random events can punctuate and disrupt predictable rhythms.
Why randomness is problematic
- Reduces player agency: Outcomes not connected to player decisions undermine the sense that skill matters.
- Perceived unfairness: Players who feel randomness cost them a game they “should” have won disengage.
- Frustrating for skilled players: High randomness in strategy games can feel like it trivialises their skill investment.
Player perception of fairness
Fairness is the key factor in how players experience randomness. De Koven (The Well-Played Game) observes that “fairness” in games is invoked as protection — a way of claiming the right to win. Hiwiller notes two conditions under which randomness is perceived as fair vs. unfair:
- Fair randomness: Uncertainty affects all players equally; no one can gain a large advantage from lucky outcomes alone. Agricola’s random setup is perceived as fair because all players receive random elements in the same way.
- Unfair randomness: Randomness advantages one player disproportionately or rewards poor decisions.
Sid Meier’s Civilization finding: Players with a 3:1 army advantage expected never to lose a battle; players with a 1:3 disadvantage expected to win 25% of battles. The magnitude of the advantage is identical, but players’ fairness expectations are asymmetric. Designers must account for this psychological asymmetry, not just the mathematical probability.
Mitigation techniques
When the design calls for reduced randomness — particularly in skill-oriented or strategy-focused games — several techniques are available:
Setup randomness, not play randomness
Limit randomness to game setup rather than play. This is the preferred pattern in many European-style (Eurogame) board games:
- Agricola randomises the initial card deal. From that point, play is deterministic (except for action order). Players learn the random starting conditions and exercise pure strategy within them.
- This gives all players equal random information to start, then rewards skill for the remainder.
Drafting
Instead of a random deal, players choose from a pool of available options. Tournament Agricola uses drafting to eliminate power asymmetries from the initial deal while retaining the variety that randomness was providing.
Card-based dice replacement
Replace dice with a shuffled deck containing die-face results in exact expected frequencies. Settlers of Catan players use this variant: if 7s should appear one-sixth of the time, one-sixth of the cards show 7. This preserves the statistical distribution while eliminating runs of bad luck. Players who chose a position paying out on 2 can no longer be unfairly punished for several sessions by dice that never rolled 2.
Elo / probability communication
For competitive games, communicate win probabilities explicitly. The Elo system in Chess gives players realistic expectations before play. This does not reduce randomness, but it reframes perception: an upset becomes evidence of skill compression rather than proof the system is broken.
In practice
Unity implementation notes:
- For random loot or drops: use variable ratio schedules deliberately (see reward-systems); implement pity systems (guaranteed drop after N failures) to prevent extreme bad luck
- For procedural generation: limit randomness to level generation (setup), not level mechanics (play)
- For competitive balance: test win rates across skill tiers and target the skill/luck mix appropriate to your audience
Design questions to ask:
- Who is my audience, and where do they sit on the casual/hardcore spectrum? (Casual players tolerate more randomness; hardcore players often want skill to dominate.)
- Is the randomness giving all players equal uncertainty, or is it advantaging some positions over others?
- Where in the game does randomness occur — at setup, during play, or at resolution?
- Can players make informed decisions about random elements? (e.g. knowing the deck state in Ticket to Ride)
Evidence
“Randomness for the game designer is like salt for a chef. Sometimes you want to prepare something a bit salty. Sometimes adding salt only hurts the dish.” — Hiwiller, Ch. 11, see source-players-making-decisions
On Civilization: players with 3:1 advantage expected never to lose a battle; players with 1:3 disadvantage expected to win 25% of battles. Asymmetric perception of fairness that the mathematical probability does not reflect (Hiwiller, Ch. 11).
Keith Burgun (Game Design Theory, 2012) makes the strongest case for randomness as a structural necessity in single-player games: without randomness, single-player games degrade into memorisation puzzles (like Castlevania or Super Mario Bros.) or execution contests (like Guitar Hero). “Without another human mind in play to throw you off, some kind of random information is required to preserve ambiguity.” He cites the original Tetris as a game that maintained genuine decision-making through its random piece generator — and argues that the 7-Bag system introduced in modern Tetris destroyed that quality. His corollary: randomness in multiplayer games should be used sparingly, because human opponents provide variance more rewardingly than a deck of cards. (Burgun, Game Design Theory, Ch. 2, see source-game-design-theory)
Roger Caillois (Man, Play, and Games, 1961): in games of pure chance, “destiny is the sole artisan of victory… the winner has been more favoured by fortune than the loser.” Satisfaction in purely random games requires social or ritual framing, not skill expression.
Implications
- A small injection of skill into a luck-dominated game substantially increases its appeal to adult players (Baccarat → Blackjack).
- Setup randomness paired with skill-based play is a robust pattern for strategy games; it provides variety without undermining skill expression.
- Perceived unfairness from randomness is as damaging as actual unfairness — design for perception, not just probability.
- Pity systems and variance reduction mechanics (card decks replacing dice) address the difference between theoretical and experienced probability distributions.
Open questions
- Does making randomness explicit (showing the player the probability of each outcome) improve or damage the experience? Some players find this engaging (poker hand odds); others find it immersion-breaking.
- Competitive games often trend toward reducing randomness over time as the player base becomes more skilled. Should games designed for long competitive lifespans start skill-heavy or allow skills to emerge?
Related
- game-balance — Randomness as one dimension of balance; variance affects competitive balance
- patterns-puzzles-chance — Chance mechanics: dice, loot, procedural generation, outcome/strategic/narrative uncertainty
- meaningful-decisions — Blind decisions are a form of randomised choice; effective randomness supports rather than replaces decisions
- reward-systems — Variable ratio schedules and loot design; the engagement and ethical issues of randomness in rewards
- flow — Randomness affects the challenge/skill ratio; too much can push players into frustration or boredom unpredictably
- game-theory-fundamentals — Mixed strategies in game theory are formalised randomness; Nash equilibria can require randomised play
- burgun-taxonomy — Burgun’s argument that single-player games without randomness degrade into puzzles or contests
- inherent-vs-emergent-complexity — Randomness as a source of emergent complexity that prevents solvability
- source-players-making-decisions
- source-game-design-theory