Defender and the philosophy of difficulty
Eugene Jarvis joined Williams Electronics in 1979 as a programmer and was given the task of designing a new arcade game. What he produced — Defender (1981) — was by any reasonable measure the most complex arcade game that had existed up to that point. Five buttons (fire, thrust, reverse, smart bomb, hyperspace) plus a joystick. A scrolling world that wrapped around a planetary surface. Humanoids on the ground who could be kidnapped by Landers, transforming into Mutants if the Lander escaped the top of the screen. A radar display showing the full horizontal extent of the world. More distinct enemy types than any previous arcade game.
Most game designers, when they examine this list of features, see a design problem: too complex, too demanding, will lose casual players. Jarvis saw it as a design goal. He had watched players at Williams' factory during development and noticed something specific: the players who were willing to invest time in mastering a complex system were the most committed, the most engaged, and the ones who spent the most money over time. A game that filtered out casual players but created dedicated ones might earn more per machine over its lifetime than a game that everyone could play comfortably.
Defender was the highest-grossing arcade game of 1981. Jarvis was correct.
Robotron: 2084 and the twin-stick invention
Robotron: 2084 (1982) was Jarvis's next game, co-designed with Larry DeMar. The design problem Jarvis had set himself was: what is the most intense arcade experience possible? His answer was a game where the screen filled with enemies in greater and greater numbers, where the goal of saving the last human family added moral weight to the carnage, and where the control scheme made movement and attack fully independent — one joystick for movement, one for aiming in any of eight directions.
The twin-stick control scheme was, as far as anyone has been able to determine, Jarvis's invention. He has said in interviews that it came partly from his own hand injury — he had hurt one hand and found himself thinking about games where both hands contributed equally rather than one hand leading. The result was a control interface so natural that players report feeling it becomes instinctive within minutes, despite having no prior reference point. Every twin-stick shooter since — Geometry Wars, the Halo games, every modern action game with right-stick aiming — descends directly from Robotron's control scheme.
Robotron was also, by design, a game you could not win. The waves escalated without limit. The question was never "can I finish this?" but "how long can I last?" This distinction — between a game with an endpoint and a game with a score — shaped an entire philosophy of arcade design. Jarvis believed that the unwinnable game was more honest than the winnable one: it told you clearly that you would fail, and measured only how well you managed your failure.
The comeback: Smash TV
Jarvis left Williams in the mid-1980s and founded Vid Kidz with Larry DeMar, a design consultancy that produced Stargate (1981) and Blaster (1983) for Williams. He returned to Williams as a designer in the late 1980s and produced Smash TV (1990), which is perhaps the clearest statement of his aesthetic philosophy.
Smash TV was set in a dystopian near-future game show where contestants fought for their lives for prize money and consumer goods. The prizes — announced in a booming game show voice ("A BRAND NEW CAR!") — were the reward for slaughtering rooms full of enemies. The game was a twin-stick shooter in the Robotron mould but with a specific satirical edge: the violence was entertainment, the spectacle was the point, the absurd consumer goods prizes were a comment on American television culture of the 1980s. Jarvis has spoken about Smash TV as a reaction to the culture he'd watched develop around his own games — the arcade as spectacle, violence as entertainment — and a decision to lean into it rather than apologise for it.
What Jarvis understood that others didn't
Jarvis's consistent insight across Defender, Robotron, and Smash TV was that difficulty, properly calibrated, was not an obstacle to engagement but a source of it. This sounds obvious now, when game design theory has developed a substantial vocabulary around challenge curves and skill development. In 1981, it was genuinely counterintuitive — the dominant industry assumption was that accessibility was always better, that any player who couldn't quickly play a game was a lost customer.
Jarvis understood that arcade games were social objects as much as games. The player who had mastered Defender demonstrated that mastery publicly, in a shared space, to an audience of observers. The difficulty was what made the mastery visible and meaningful. A game everyone could play well produced no hierarchy of skill to admire. A game that was genuinely hard produced experts whose skill was spectacular to watch — and spectacle drove other players to the machine.
His games were, in this sense, designed to be watched as much as played. Defender and Robotron both look extraordinary when played by experts — fluid, fast, almost balletic in the way skilled players move through the screen. The difficulty was a filter that selected for players capable of producing that spectacle. The coin drops from observers who wanted to try to do the same thing were part of the business model, consciously or not.