What Was the First Video Game? Not Pong, by 25 Years

On 18 October 1958, a physicist named William Higinbotham wired a Donner Model 30 analog computer to a five-inch oscilloscope, set it on a table at Brookhaven National Laboratory’s annual visitors’ day, and invited the public to play tennis on it. Hundreds of people lined up. Higinbotham never patented the thing, because, as Brookhaven’s own history page records, the game “did not seem to be any more novel than the bouncing ball circuit in the instruction book”.

He was also eleven years too late to be first.

What was the first video game?

There is no single one, and every honest answer gives three. The earliest patent for a game played on a screen was filed in 1947. The earliest real-time game a member of the public could walk up to and play was Tennis for Two in 1958. The earliest that escaped its own building and got copied was Spacewar! in 1962. Pong, in 1972, is none of the above.

Each candidate falls to a different rule

The reason nobody picks a winner is that the candidates are eliminated by different tests, and no two people apply the same set. Laid out with the disqualifier attached, the argument stops being vague.

Candidate	Date	Who	Why it does or does not count
Cathode-Ray Tube Amusement Device	Filed 25 Jan 1947	Goldsmith and Mann	Player-steered dot on a CRT, but no computer and no memory. Targets were pictures placed on the tube face. Never built for sale.
Bertie the Brain, Nimrod	1950, 1951	Josef Kates, Ferranti	Real games on real machines, but the board was a panel of light bulbs rather than a video display.
OXO	1952	Alexander S. Douglas	Drew noughts and crosses on a CRT under stored-program control, on EDSAC, the single machine in the Cambridge Mathematical Laboratory.
Tennis for Two	18 Oct 1958	William Higinbotham	Real-time, two-player, with gravity and a net. Ran on an oscilloscope and generated no video signal. Dismantled in 1959 for parts.
Spacewar!	1962	Steve Russell and others, MIT	Digital, real-time, and copied to other sites. Needed a PDP-1, priced at US$120,000.
Computer Space, Pong	1971, 1972	Bushnell and Dabney, Allan Alcorn	First sold, then first to sell well. Both arrive a decade after the interesting part of the argument.

The one rule everybody does share is buried in the word itself. Under the early technical definition, a product counted as a video game only if it transmitted a video signal to a display, which rules out games that sent their output to a printer or a teletype and the machines that spelled their boards in light bulbs. Countdowns have a first-of-its-kind argument too: the earliest backward count is a Cambridge boat race in 1865, a Jules Verne cannon in 1889 or a rocket launch in a 1929 film, depending on what is being counted down to.

Bertie and Nimrod were advertisements with a game attached

Bertie the Brain stood four metres tall in the Engineering Building at the Canadian National Exhibition from 25 August to 9 September 1950, playing noughts and crosses against passers-by on a grid of nine illuminated squares. Josef Kates built it to show off the additron tube, a vacuum tube he had shrunk down, and once the fair closed the machine was dismantled and largely forgotten as a novelty.

Ferranti’s Nimrod was twelve feet wide, appeared at the Festival of Britain on 5 May 1951, and played only Nim. Players pressed buttons that corresponded to lights. Both machines existed to make a company look clever, and both spelled out their game in bulbs.

OXO, in 1952, is the first of the group to put its board on a cathode ray tube under stored-program control. Alexander Douglas wrote it for EDSAC at Cambridge, and the player picked a square with a rotary telephone dial.

The 1947 patent stuck its targets to the glass

US Patent 2,455,992 was filed on 25 January 1947 and granted on 14 December 1948 to Thomas T. Goldsmith Jr. and Estle Ray Mann. The patent text describes a missile game of pure analog circuitry: knobs move sliding contacts on resistors, the resistors bend the path of the electron beam, and a hit is signalled by defocusing the spot so it appears to explode.

The wording gives away the limit. Targets, the patent says, “such as pictures of airplanes, for example, are placed upon the face of the tube”. The machine never drew the thing you were shooting at. A person stuck it on the screen.

Twenty-five years later the first home console shipped with the same workaround. The Magnavox Odyssey could only put white shapes on a black screen, so Magnavox supplied plastic overlays that clung to the television by static electricity and provided the scenery the hardware could not.

Cost decisions like that one look ridiculous in hindsight and are almost always somebody solving a budget problem in the room they were standing in, the same way the connector everyone flips twice before it goes in got its shape.

Brookhaven hedges its own claim

The laboratory that built Tennis for Two will not call it the first. Its history page settles on “unquestionably a forerunner of the modern video game” and states flatly that “Higinbotham’s system did not create a video signal”. Brookhaven technician Robert V. Dvorak put the whole apparatus together in about three weeks. After the 1959 showing it was taken apart so the components could be used for something else.

Pong is wrong by a quarter century

The Strong National Museum of Play notes that most people asked this question answer Pong, then corrects them: it “was not even the first arcade video game created by Atari founder Nolan Bushnell”. Bushnell handed the design to Allan Alcorn as a secret warm-up exercise, and Atari announced the finished game on 29 November 1972. Bushnell and Dabney had already shipped Computer Space in late 1971, running on no microprocessor at all, just a state machine of 7400-series chips with the graphics held in diode arrays.

Pong’s genuine claim is money: Atari filled 2,500 orders by 1973 and had sold more than 8,000 cabinets by the end of 1974. Being the version most people saw first is a different achievement from being first, as Cher’s 1998 hit “Believe” showed when it made Auto-Tune famous after about a year of discreet use on other records. The obvious explanation for a familiar thing turns out wrong often enough to be a working rule, right down to what a fly is doing when it wrings its front legs together.

Spacewar! is the one that got out of the building

Steve Russell began coding Spacewar! on a DEC PDP-1 at MIT around December 1961, with Martin Graetz and Wayne Wiitanen on the concept, and the game was demonstrated at the MIT Science Open House at the end of April 1962. Copies travelled to PDP-1 installations at Stanford, the University of Minnesota and Bolt Beranek and Newman, which makes it, as far as anyone can document, the first video game available outside a single research institute.

Available is doing heavy lifting there. A PDP-1 cost US$120,000, so the audience was anyone who already owned a room-sized computer. Spacewar! then fed directly into the first two coin-operated arcade machines, Galaxy Game and Computer Space, both in 1971, and only Computer Space was ever sold. Early technical decisions keep outliving the reasons for them, which is also why the letters on your keyboard sit in that particular order.

Pick the rule and the answer falls out. A screen gives you 1947. A stored program gives you 1952. Real-time play gives you 1958. Distribution gives you 1962, and a paying customer gives you 1971. None of those five roads ends at Pong. Strictly FYI.