The unwinnable loop
The first arcade games had no endings. Space Invaders, Asteroids, Pac-Man — all were designed as loops that escalated in difficulty until the player was overwhelmed. There was no credit sequence, no victory screen, no resolution. The game ended when the player's lives were exhausted, and the only record of achievement was a number on a screen. This design was not a primitive limitation. It was a deliberate economic choice: a game you could finish was a game that satisfied you, and a satisfied player had no reason to spend another quarter.
The loop design was also a statement about what games were: not narratives with beginnings and ends, but skill challenges with no ceiling. The point was not to finish. The point was to see how far you could go, measured against other players' scores on the same machine. The high score table was the game's memory — the accumulated evidence of everyone who had played, ordered by the depth of their engagement with the game's challenge.
The consequences of this design for how games managed failure were specific. Losing in an arcade game was not an obstacle to finishing — there was nothing to finish. It was the endpoint of that session, the measurement of how well you'd done. The "game over" screen was simultaneously a failure message and an achievement record. It told you how far you'd gotten, which was the only information the game's structure could give you.
Lives: the first abstraction
The lives system — you have three (or five, or nine) opportunities before the game ends permanently — was one of the first significant design abstractions in the history of games. It acknowledged that losing once was not the end; you had additional attempts. It built a sense of dread into the design — losing a life mattered more than losing nothing, because losing enough of them ended the session. And it gave players a sense of how they were doing relative to expectation: two lives remaining is fine, no lives remaining is desperate.
The lives system also served the arcade's commercial interests elegantly. Three lives gave most players enough rope to get interested in a game before failing — enough play to make you want to try again. Fewer lives made the game frustrating for beginners; more lives allowed too-comfortable play that didn't extract coins at the optimal rate. The standard of three lives was an economic calibration as much as a design one.
The tension within the lives system — between giving players enough chances to stay engaged and not giving them so many that they stopped spending — produced specific game design decisions. Donkey Kong gave you three lives. Pac-Man gave you three. Space Invaders gave you three. The standardisation across games from different manufacturers suggested that operators and manufacturers had converged on three as the right number through trial and error — the value that balanced engagement with coin extraction most effectively.
The continue and its ethics
Gauntlet (1985) made the economics explicit with its continue system: when your character's health reached zero, a nine-second countdown appeared. If you inserted another coin before the countdown ended, your character was restored with full health. If you didn't, the game ended. The countdown timer created urgency. The restoration of full health created desire. The mechanics were designed to make the decision to continue feel impulsive — you didn't want to lose your progress, you were right in the middle of something, the dungeon was almost cleared.
The continue mechanic spread rapidly through the mid-1980s arcade game landscape. It was commercially rational — extracting additional coins from players who were already engaged was more efficient than requiring them to start again — and it allowed more complex game structures. If you could continue from where you died rather than starting over, you could design a game with a narrative arc, with a final boss, with a sense of progress toward completion. The continue made the unwinnable loop less commercially necessary and created space for games with actual endings.
The ethical dimension of the continue — the deliberate use of emotional momentum and the sunk cost fallacy to extract money from players who were about to leave — became more visible as the mechanic migrated to home games in the form of lives systems that eventually became game-ending "game overs." A home game that gave you ten lives and then sent you back to the beginning was not extracting coins; it was simply making the game hard. But the design logic was identical. Difficulty was the mechanism that justified the continue, and the continue was the mechanism that generated revenue.
Saves and the democratisation of completion
The battery-backed save in The Legend of Zelda (1986) was a design decision as consequential as any in the medium's history. For the first time, a home game allowed players to preserve progress and return to it. You didn't have to complete the game in one sitting. You didn't have to use a password system — which preserved some progress information but was cumbersome and could be incorrectly transcribed. You could play for an hour, save, turn off the console, and return the next day.
The save changed what kinds of games were designable. A game that could be saved could be longer, more complex, and more narratively elaborate than a game that had to be completable in one session. The JRPG — Dragon Quest (1986), Final Fantasy (1987), and their descendants — is a genre that exists because the save exists. A forty-hour game experience requires that players can pause and resume without penalty. The save made that possible.
It also democratised completion. Before saves, completing a complex game required either extraordinary skill (to finish in one session without dying too often) or extraordinary commitment (to leave the machine running or to use passwords carefully). After saves, a patient player of ordinary skill could complete almost any game if they were willing to spend enough time on it. The relationship between difficulty and completion shifted: difficulty became one variable among many, rather than the primary barrier.
The modern settlement
The contemporary games industry has largely settled on a position that would have been unrecognisable in 1980: difficulty is a designer's tool, not a commercial mechanism, and the appropriate level of challenge should be calibrated to the experience the designer wants players to have rather than to how long it takes to complete the game or how many coins it extracts.
This settlement is not universal. The "soulslike" games — Demon's Souls, Dark Souls, Elden Ring — have built a major commercial identity around high difficulty and the specific satisfaction of overcoming challenges that initially seem impossible. These games are commercial successes despite — or because of — their explicit rejection of the accommodation philosophy. Their difficulty is real, not a coin-extraction mechanism, and players who succeed experience something genuine. But they exist within a broader industry that has largely moved toward accessibility options, multiple difficulty settings, and the understanding that different players want different experiences from difficulty.
The line from Space Invaders' unwinnable loop to the difficulty slider in a 2024 action RPG passes through the lives system, the continue, the save, the checkpoint, and the multiple difficulty setting. Each step represented a deliberate choice about who games were for and what they were trying to do. The evolution was not inevitable. It was a series of decisions made by designers responding to players, operators, and the commercial realities of whatever market they were working in. The history of difficulty is, in this sense, a history of the industry's changing relationship with its own audience.