The First Video Game

Here’s a question for you: What was the first ever video game? It seems like a simple enough question – but
it’s trickier than you think. The trouble is, you’ll get a different answer
depending on who you ask – and none of them seem definitive. So, let’s take a critical look at the origin
of video games – starting with a brief history of the popular consensus. For a long time, the most common answer to
the question was Pong: Atari’s 1972 arcade game. Pong was a phenomenon, the first game to break
through into popular culture – and for that reason, quite a few magazines from the 80s
and early 90s cite it as the first ever game: such as Sinclair User in 1986; a 1987 issue of Your Computer; the first issue of ST/Amiga Format in 1988; and a 1990 issue of Amiga World. Some do at least add a qualifier, such as
Computer Gaming World’s claim in 1982 that Pong was the first computer game ‘of renown’: Similarly as ‘the first video game for the
masses’ in a 1987 issue of Ahoy. So there was at least some awareness of an
earlier history, but not much in the way of in-depth research. I suppose it was a different era – Google
wasn’t a thing, and even a trip to a local library wouldn’t be much use as video games
weren’t exactly well documented. So with Pong being so prominent, it was generally
accepted as an origin point – and while it does mark the popular emergence of games,
there lies a long and obscure history behind it. The funny thing about Pong is that it isn’t
even Atari’s first game, never mind the first game ever. Before Atari were Atari, they were called
Syzygy Engineering – and along with manufacturer Nutting Associates, they released a game called
Computer Space in 1971. Computer Space was never as popular as Pong
– selling maybe 1,500 units as opposed to the tens of thousands of Pong arcade machines,
and the millions of subsequent clones. But one of the co-founders of Atari, Nolan
Bushnell, was certainly no stranger to publicity – and in interviews during the early 1980s
he was quick to point out that Pong wasn’t his first game. Some of the gaming press picked up on this,
reproducing his words verbatim: Big K in 1985 cites Computer Space as ‘the
first marketed video game’, which might be correct; And a 1993 Amiga Format Special is proud to
reveal that Pong isn’t the first video game after all – instead granting that accolade
to Computer Space. So for a long time, Nolan Bushnell was seen
as the ‘father of video games’ – but here’s the thing: Computer Space wasn’t exactly original:
it was based on an earlier game that Bushnell had played during his time at Stanford University
– again, something that he freely admitted. That game was called Spacewar!, and it first
emerged in the computing labs of MIT by 1962. Unlike Computer Space, it wasn’t an arcade
machine – it was never intended for commercial release. It simply wasn’t a viable option – it ran
on an expensive minicomputer, the PDP-1, which even in a basic configuration sold for $120,000
– in the 1960s! So it was a purely academic exercise, but
one that was surprisingly popular on campus – at least amongst those with access to the
computing lab. The July 1971 issue of ‘Analog: Science Fiction,
Science Fact’ describes it as ‘the first true space-age game’, and they aren’t wrong – this
was just two years after the moon landings. This is one of very few accounts that detail
an interactive game before video games had a meaningful impact on popular culture – it’s
interesting to see how they treat Spacewar as a unique thing rather than as part of something
else. Spacewar! isn’t seen as a video game: Spacewar!
is just Spacewar! Over its life, the game spread from institution
to institution, and became reasonably well-known within the right community. But it wasn’t
until the early 1980s that the game started to reach wider public attention. In 1981, one of the game’s co-creators – Martin
Graetz – wrote an article for Creative Computing: a fairly comprehensive 8-page history of its
development and operation. However, it was a 1984 book that did the most
to bring Spacewar to wider public attention: Steven Levy’s ‘Hackers: Heroes of the Computer
Revolution’ detailed some of the exploits of MIT engineers over the last couple of decades,
with a whole chapter dedicated to Spacewar! So, as time went on and more people learned
about the game, it gained some recognition as ‘the first computer game’: Byte magazine
described it as such in 1990. This reign would be short-lived, however – as
another, earlier, game came to light soon after: A game from 1958 called Tennis for Two. Tennis for Two was the creation of a man named
William Higinbotham, an American nuclear physicist. While working at the Brookhaven National Laboratory,
he created a game designed to show off the lab’s instrumentation at their annual exposition. Tennis for Two was shown to the public in
1958, and once again in 1959 – before being dismantled and largely forgotten about. Until 1982, at least – a mention of the game
appears in an issue of Creative Computing: an article titled ‘Who Really Invented The
Video Game?’ It details the hardware used – an analogue
computer with a 5-inch oscilloscope display – and goes on to describe the gameplay. It was a tennis game, not dissimilar to Pong,
with a ball following a realistic trajectory as each player hits it back and forth. It was quite a convincing simulation, but
considering it was using hardware designed to calculate the trajectory of ballistic missiles,
that might not be too surprising. Despite Higinbotham’s subsequent radio interview
with NPR in ’83, the revelation of a video game that dated to as early as 1958 passed
most people by – most were still content in thinking that Pong was the origin. In 1993, author Ira Flatow sought to correct
this as part of his book: ‘They All Laughed’ – a collection of invention stories from throughout
history. The final chapter is all about Tennis for
Two. Flatow had worked at NPR, so it’s likely he’s familiar with Tennis for Two through
Higinbotham’s interview in 1983. Sadly, William Higinbotham died in 1994 – but
his obituaries detailed his achievements, including his game: And subsequently, Tennis
for Two became well-known – the new answer to the question of ‘what was the first video
game?’ Perfectly satisfactory. Looks like Pong, dates
to 1958. There can’t be much earlier, can there? Today, if you look at Wikipedia’s article
on the ‘Early history of Video Games’ you’ll see mention of a 1947 invention: The ‘Cathode-ray tube amusement device’. Sounds like a delightfully archaic term for
video games, and we can verify its existence in the patent record: number 2,455,992; submitted
on the 25th of January 1947; and granted on the 14th December 1948. Its impact must have been limited – it’s only
recently come to prominence, roughly as long as the existence of Wikipedia. This re-emergence can be traced to Jed Margolin’s
2001 article: ‘The Secret Life of Vector Generators’. Margolin was an Atari engineer familiar with
the golden era of arcade games – and crucially, is the named inventor on a number of patents
relevant to early video game hardware. This 1947 patent first came to light in a
lawsuit: Magnavox vs. Bally, circa 1976. The ‘amusement device’ was cited as prior art
in Bally’s defense. So someone working on Bally’s behalf must
have done the necessary legwork to uncover this – after all, there was a strong incentive:
Magnavox held a key patent on video games that earned them $100 million dollars over
its lifetime. And according to the man behind that patent,
Ralph Baer – video games were his invention. The man had an interesting career: he fled
Nazi Germany in 1938; was drafted into World War 2 in 1943; and, courtesy of the G.I. Bill,
attained the first ever BSc in Television Engineering in 1949. By 1961 he ended up at Sanders Associates
– a defense contractor that specialised in electronic systems. In September 1966, Baer laid out his ideas
for an interactive television game – and work began on a series of prototypes. These prototypes culminated in the ‘Brown
Box’ design in 1968 – so named for its wood-grain effect vinyl covering. This device connected to any standard television
and could play a variety of games – including a tennis game that has a startling similarity
to Pong, seen here in a 1969 demonstration. This was the system that would eventually
be licensed and sold as the Magnavox Odyssey in 1972: the very first home video game system. Everything at Sanders was remarkably well-documented
– perhaps because of their work for the military – but this means there’s a large body of dated
documentation to accompany Baer’s invention. Additionally, unlike Tennis for Two at Brookhaven
National Laboratory, and SpaceWar! at MIT – Sanders was a commercial entity: they existed
for profit, and were at liberty to patent their innovations. And patent they did, with Baer’s work submitted
in 1968, amended while pending in 1971 – and granted in 1973: US patent No. 3,728,480. A second was filed in ’69, granted in ’72
and reissued in ’75 – Reissue. 28,507. Collectively, the ‘480 and ‘507 patents gave
Sanders (and licensee Magnavox) a very strong legal right to fundamental concepts behind
early video games. A right which they exploited: Magnavox filed
lawsuits against Atari, Bally Midway, Chicago Coin, Mattel, Activision – and won all of
them, garnering substantial licensing fees. Nintendo pre-emptively sued Magnavox for the
US launch of the NES in 1985: in an effort to overturn these patents. They lost as well. Patents are strong, and
Magnavox had precedence. Despite the legal victories, Baer was regarded
as something of an ‘also-ran’ in the early video game stakes, but he gained quite some
recognition following the 2004 award of the National Medal of Technology. His time spent forgotten is evident in the
frustrated tone of his 2005 book: ‘Videogames: In The Beginning’ He lays his claim to the industry’s genesis,
and the disdain for Nolan Bushnell – and Atari’s later success – is particularly apparent. He’s also very dismissive of any precursors:
he writes off Spacewar! as impractical, and Tennis For Two as akin to ‘fiddling with an
oscilloscope’. But it’s hard to blame him. He was definitely
a pioneer. Ralph Baer died in 2014 at the age of 92,
his accomplishments in video games eulogised. Whether or not he invented video games, he
can lay legitimate claim to the first ‘home video game’ – but I feel there’s more to video
games than just consoles. To complicate things further: while Tennis
for Two and the 1947 patent steal the limelight, anything between these milestones falls through
the cracks. Take OXO for instance – a 1952 game developed
for the EDSAC, an early digital computer. It’s a simple game of noughts and crosses
(or tic tac toe if you’d prefer). Its display uses a cathode ray tube – in fact,
the computer has three of them used to indicate program and system status. The middle tube is normally used to display
part of memory, in a 35 by 16 dot matrix – but rather than store gamestate in binary, OXO
exploited the visual memory display to create a game board: filling up with Xs and Os as
the game is played. Then there’s NIMROD, part of the 1951 Festival
of Britain. A dedicated machine designed to play the game
of Nim ‘faster than thought’, played on a bank of lights: the player and AI taking turns
to remove them, attempting to leave the last light for their opponent. Bertie The Brain was another computing demonstration
in the summer of 1950 at the Canadian National Exhibition. Similar to the later OXO, Bertie was a computer
AI designed to play a game of tic-tac-toe on a display illuminated by an array of light
bulbs. So we have a number of candidates for ‘the
first ever game’, and we’ve painted a picture of our understanding over the last few decades. Paradoxically, the further you go back, the
more recent the consensus of the ‘first ever video game’ – and everything seems to converge
on Pong. As video gaming developed, so too did the
appreciation of its history – and as more scrutiny was given to the limits of popular
memory, the origin was pushed farther back – first to Spacewar!, then Tennis for Two. The principal catalyst for this interest,
at least at first, were the Magnavox lawsuits – with a very strong financial incentive for
someone to overturn these key patents. But as we shed more light on the past, another
paradox emerges: The more we uncover, the less certain we become. So, we have a problem. There’s no shortage of candidates for the
title of ‘first video game’ – but some of them feel like they might not qualify. To land at a satisfactory answer to our original
question we must first answer another: What is a video game? I told you it was tricky!: But a definite
conclusion, needs a concrete definition. Defining the video game is hard, but determining
one less so. The duck test is applicable here, i.e. if it looks like a video game and plays
like a video game, then we can probably call it a video game. So, Pong? Video game. Asteroids? Video game. Space Invaders? Video game. However, with the earliest examples of games
we start to approach the edge of the definition – so we need to determine a hard delineation
between ‘video game’ and ‘not video game’. The way I see it, there are two essential
elements: the first is a video display of some kind, and the second is interaction for the purpose
of entertainment: i.e. a game. We’ll start with the ‘video’ part. The word comes from the Latin ‘video’, meaning
‘I see’ – think of it as a counterpart to audio, ‘I hear’. However, it’s modern meaning is slightly more
specific – it has the implication of an electronic signal driving a visual display, distinct
from film which is projected. This normally means television, but can also
include computer monitors, oscilloscopes and liquid crystal arrays: any display capable
of receiving a signal and producing an image. For this reason, I think we can exclude games
that use indicator lamps: these tend to be electronically integrated into the logic circuit,
rather than driven by a separate signal. Now, you could argue that any display is just
an array of lights, and that there need not be a minimum array size – and I agree that
we don’t need to stipulate a minimum resolution: but on the other hand, a single light bulb
doesn’t constitute a video display. The key difference lies in the generation
of a video signal. Video signals come in a variety of forms,
but they are generally low voltage, high frequency and transmissible between two different electronic
circuits. They often have a certain timing, either a
particular synchronisation frequency or a separate clock – accompanied with some form
of image signal. I don’t think we need to stipulate any particular
technical format, but for our definition of ‘video game’ we are going to need at least
some kind of video signal. This means that neither Bertie The Brain nor
Nimrod qualify: their banks of light bulbs don’t meet our video signal requirement. Next, consider the Nintendo Game Boy. This particular breed of handheld LCD game
presents another boundary. Compare the earlier Game & Watch: I would
consider this just an electronic game, but the Game Boy a bona fide handheld video game. So what’s the distinction? I think both are
borderline, but there is a key difference in the way the LCD displays are driven. The Game & Watch is essentially running on
the same type of display as a calculator – an active liquid crystal display with a quad-multiplexed
LCD driver. The only difference is that the LCD segments represent gameplay elements instead
of numerical digits. The Game Boy, however, has a considerably
more complex screen: At 160 pixels across, and 144 pixels tall, there are over 20,000
segments to control – each with 4 levels of brightness. Because of this, the signals sent to the display
resemble a conventional video signal: There are horizontal and vertical sync lines, along
with a 2-bit digital data channel. It’s not NTSC-compliant, but it could definitely
be described as a video signal – and in fact, hardware such as the Super Game Boy can adapt
them for use on a standard television. If there was any doubt to the distinction,
the language used on their respective patents is another clear indicator: the Game & Watch
is described as an ‘Electronic Toy Having a Game Function’, whereas the Game Boy is
a ‘Compact Hand-held Video Game System’. So we can draw a line between the two, and
we’ve established that a cathode ray tube is a nice-to-have, but not an absolute necessity. However, even when dealing with CRT displays
there are some that make the argument that not everything qualifies: Ralph Baer, for
instance, is quite specific with his definition. Baer states that to be considered a true video
game, the device on which it is played must generate a raster signal capable of being
displayed on a domestic television set. Fair enough I suppose – and most games do
fit this definition – but it’s a little too strict for my taste. Technically it would exclude many early computer
games that targeted monitors rather than televisions – but most crucially, it excludes every single
vector game. While they’re extinct today, vector games
were an important part of the golden arcade era: including Asteroids, Star Castle, the
Vectrex console, and the 1983 Star Wars arcade game. There are dozens of examples, and I can’t
think of a good reason to discard them. Their graphics aren’t painted by a raster
signal but they are still painted by a signal – a two-axis XY signal that directly controls
the electron beam’s deflection. My gut says Asteroids is a video game, and
so our definition of ‘video game’ has to include vector displays – but this has significant
meaning when it comes to answering our question. By keeping our definition broad enough to
include vector video, we have to also include some of the earliest candidates: Spacewar! had vector graphics: Tennis for
Two had vector graphics; the 1947 patent would have used vector graphics. So we can’t yet exclude any of them. But now we have a reasonably good sense of
what the term ‘video’ means, we can move onto the other essential aspect: the ‘game’. We don’t need to be too strict with this definition:
when we say ‘game’ what we really mean is ‘interactive entertainment’. But what our definition needs to do is separate
video games from something like television itself. For instance, watching television
is definitely entertainment – and to some extent it is interactive: changing channels,
or simply powering the device on qualifies as interaction. The critical difference lies with the generation
of the video signal: in television, these are broadcast remotely – and the signal itself
cannot be changed, only switched between. With a video game, the signal is generated
locally: and crucially, the player’s input has a direct influence on the signal generated.
The game responds to the player through the medium of video. There are some recent exceptions to local
signal generation: services such as OnLive or Google Stadia stream video over the internet
– but in these cases the player still controls the signal, albeit through an additional network
layer. So, a video game must give some degree of
control over the video signal itself to be considered sufficiently ‘interactive’. Next, consider a DVD menu: technically, the
video signal is altered in response to user input – options are highlighted, and the menus
are navigable in a non-linear way. All of the elements necessary for a video
game are present – in fact, some DVDs even feature ‘games’ – and these arguably qualify
as video games. That said, however, a typical DVD menu definitely
isn’t a video game. So what’s the difference? Put it this way: How do we separate non-entertainment
software from games? Consider a spreadsheet used for managing resources
– versus a resource management simulator. Their interfaces might be very similar, and
the mode of operation nearly identical – the key difference is intent. A spreadsheet for resource management is created
to manage resources – and there are some people who find that entertaining, but entertainment
is not the primary purpose. The simulator, on the other hand, can be primarily
intended for entertainment. Some simulators aren’t – they might be for training purposes,
such as a military flight simulator – but if it’s a commercial product that you can
play on a home computer, then there’s a good chance you can consider it a video game. The same applies to DVD menus – while they
are interactive, and do serve to deliver entertainment – their interaction is not intended to be
intrinsically entertaining. So intent is very important: not only must
interaction be present, but the principle intended purpose of that interaction must
be some form of entertainment. So we’ve covered the broad strokes, but there
are some additional points of contention that might shape our final definition. For instance: does a video game need to be
implemented to be considered? Is a concept enough? The 1947 CRT Amusement Device is only known
to exist as a patent. Some sources claim prototypes were built, but there’s no evidence of this. The inventors, Thomas T. Goldsmith Jr. & Estle
Ray Mann, did work for an established manufacturer of television equipment – DuMont Laboratories
– so it’s not unreasonable to assume that they had the means to manufacture such a device. But given the absence of any marketed product,
and the lack of any further development – it seems the project wasn’t deemed viable. Now, it seems fairly logical to insist that
the ‘first ever game’ actually existed, but in a sea of prototypes and experiments it’s
less clear cut than I’d like. I think prototypes should still count – but
our definition should only include games with evidence of a practical implementation. So is that enough to discount the 1947 device?
I think it depends on how strict you are with the definition of ‘evidence’. A patent might be enough – they describe the
invention in detail, and are subject to scrutiny before being granted. But where the 1947 device falls flat, however,
is with the gameplay itself: while it does generate a video signal that is directly controlled
by the player, the game relies on other elements – physical targets on the face of the display. Without these, you’re left with a triggered
signal generator with an oscilloscope display. A neat novelty, but I do think this puts the
1947 device in the realm of ‘electromechanical games’ rather than that of true video games. Consider a pinball machine. Modern examples
can be quite advanced, even incorporating games played on a video display – but I don’t
think anyone would make the argument that a modern pinball table is a video game. The core gameplay is the mechanical ball and
flippers – the video display is accessory to that. Conversely, it’s possible for a video game
to have accessory elements: think of an arcade game’s coinbox, marquee lights et cetera;
an arcade cabinet has more than just a video display going on, but that’s not a reason
for exclusion – as the primary game can be played solely through the video display: Whereas pinball relies on the mechanical elements
for complete function. So, the 1947 device might be the first use
of a CRT for the purpose of interactive entertainment, but its use of mechanical elements as a core
part of its game means it isn’t a true video game: it’s an electromechanical game with
a video element. And this means it doesn’t qualify under our
definition. Our game has to be playable solely through
the use of the video display(s). Accessory indicators or electronics aren’t reason for
exclusion, but – in terms of visual output – the game has to be playable by the video
signal alone. And so, that leaves us with a fairly robust
definition – one that consists of five points. Firstly, a video game must exist: It must involve some kind of video signal; This signal should be in some way interactive; and the principal intended purpose of this
interactivity should be entertainment. Finally, the game should be centred on and
fully playable through the use of the video display(s). If a candidate meets these five requirements,
then I think we can consider it a video game. So we have our definition. The only thing
left to do is apply it. So, the candidates that remain are Pong: Computer
Space; Baer’s Brown Box; Spacewar!; Tennis for Two; and OXO. All of these are documented to exist, so there’s
no question of practical implementation for any of them – least of all Pong. In terms of video signal, Pong, Computer Space
and the Brown Box generate a raster signal suitable for an NTSC television – whereas
SpaceWar!, Tennis for Two, and OXO generate an X-Y signal for display on an oscilloscope
or vector monitor. OXO’s display is a little contentious – while
it did use a cathode ray tube and it was driven by a signal, the display wasn’t really intended
to be a graphical one. In fact, the cathode ray tube wasn’t chosen
because of its display capability at all – in the early days, CRTs were used for storage. One implementation was called the Williams-Kilburn
Tube, and it relied on a curious side effect of energised phosphors – once lit by the scanning
electron beam, a charge would persist for a fraction of a second. This secondary emission could be sensed, then
the phosphor re-charged on the next scan: meaning you could store an array of a couple
of thousand bits indefinitely. This was presented on the face of the display
as a grid of illuminated dots – the electron beam scanning rows in a fairly similar way
to a raster display. However, a sensor sheet on the face of the
tube did cover up most of the visible pattern, making this arrangement less than ideal for
use as a display – but quite often a second tube would be connected in parallel without
the obstruction. EDSAC, the machine that ran OXO, didn’t use
a Williams tube for storage – it used mercury delay line memory instead – but the contents
of this memory were displayed on a diagnostic CRT in the same manner. So while it was never really intended to be
used for graphics, the fact is it was – and so it still qualifies. Back to the candidates: in terms of interactivity
I think most here count – again, the only possible exception is OXO. Because of the limitations of the display,
OXO is considerably less dynamic than the other examples – with the display only updating
when a move is made. That said, the display does change in response
to player input. So it is interactive, if not exciting. Intent is the fourth criterion: and it’s fairly
clear that Pong, Computer Space and Baer’s Brown Box were all intended for entertainment
– that was their ultimate commercial purpose. Spacewar! was definitely a diversionary pursuit,
and is frequently described as a game. Tennis for Two was an exhibit intended for
public display – and it may have had an ulterior motive to kindle an interest in science – but
the prime appeal was in the playing. Now, OXO. OXO was the product of study at Cambridge
University – a practical proof as part of a thesis. If we take it as part of a PhD, the work is
purely academic – but I think it’s fairer to judge it on its own merits. As a piece of software, it stands alone as
what could easily be described as a ‘computer game’, so while entertainment isn’t the purpose
of implementation – it was the purpose of playing. So the final test: can the game be played
with no other visual output than the video display? Pong and Computer Space qualify without question.
They are undoubtedly video games. Few would argue otherwise. You could say that the Brown Box’s use of
screen overlays for some of its games disqualifies it on this point – and there is some merit
to this in cases where the plastic overlay is absolutely necessary to play. But for most of the games it’s not – just
an accessory – and the basic Tennis game doesn’t have an overlay at all, so the Brown Box gets
a pass. It’s a video game. Spacewar! ran on the PDP-1, which was a modular
system with a range of potential visual outputs. The CRT monitor was central to the game, but
it wasn’t the only output that SpaceWar used. In some versions, game scores were printed. However, these aspects are secondary to the
actual gameplay – so Spacewar! is a video game. Tennis For Two is simpler: its only output
was the oscilloscope display, so it is a bona fide video game. What about OXO? Like Spacewar! it ran on a full computer system,
with its CRT display, line printer – and banks of valves and das blinkenlights. The question is: is it possible to play a
full game with only the video display? The answer is… yes. With the caveat that
you have to know what you’re doing. If you memorise which number corresponds to
each grid position (or at least make use of a printed reference) then it’s entirely possible
to play without any other output. It’s clumsy and archaic – your moves are input
on a rotary telephone-style dial – but poor controls are no reason to discount it. There
are plenty of video games with terrible controls. So OXO is, without doubt, a computer game
played on something which can be described as a video display. Under any reasonable definition,
I think that qualifies as a video game. It might even be the first. However, to be sure it might be worth further
investigation: a closer examination of the environment that spawned OXO may turn up some
other potentials. What we think of as video games really comprise
multiple strands: Arcade machines; Television games; Electronic games; And more recently,
Virtual Reality; And Mobile Games, which have rapidly become a significant part of the video
games industry. When we tease out these strands, the reason
why it’s so difficult to pin down the ‘first’ video game becomes apparent. Each of these strands has its own history:
its own progression; and its own origin. If we look at them individually, the picture
becomes much clearer: The first arcade video game? It’s either Computer
Space or Galaxy Game, both from the end of 1971. The first television game? Ralph Baer’s Brown
Box. The first electronic handheld video game?
It could be the Game Boy, unless you count the Atari Lynx prototypes from 1987. The first virtual reality video game? Possibly
played via the VPL Research EyePhone in 1987, or the Virtuality from 1991 – although research
in head-mounted displays goes all the way back to the 1950s. And the first mobile game? Tetris, on the
1994 Hagenuk MT-2000. Nokia’s Snake gets an honorable question. So when we split video gaming into its components,
the question becomes much easier to answer – and the oldest strand becomes obvious. It’s computing. The first video game was almost
certainly an early computer game like OXO – it’s just a question of determining the
earliest example that qualifies. The good news is, computers weren’t exactly
common in the 1940s and 50s – and fewer still had video displays. In the brief window between the invention
of the Williams tube circa 1946 and OXO in 1952, there are seven potential machines. Three of these potential computers were located
at the University of Manchester, where the Williams tube was developed. The first was the Manchester Baby, operational
in 1948 – a prototype, which was replaced by the full-scale Manchester Mark 1 in 1949. This in turn was developed into and replaced
by a commercial model: the Ferranti Mark 1, operational in early 1951. There were two other British machines: The
Pilot ACE at the National Physical Laboratory in Teddington, London – operational in mid-1950; And EDSAC at Cambridge University, which was
first operational in 1949. At this time, the Americans had two machines
with a video display: the Whirlwind I, built by MIT for the US Navy – partially operational
in 1949, and reaching full capacity in 1951; And the Standards Western Automatic Computer,
or SWAC – located in Los Angeles. It’s impossible to know every program that
ran on these early machines – but most of them are fairly well documented. Access was exclusive, and programs were filed
in a job queue. There was normally a paper trail, and there was an incentive for institutions
to highlight their machine’s accomplishments. Unfortunately, this transparency and a need
to justify the computer’s existence meant that games were a very low priority. Starting with the SWAC, a commercial machine,
we can see most applications were either pure mathematics, or scientific in nature: Such as searching for prime numbers; developing
climatological models; the X-ray analysis of vitamin B12. Processing time was split between the Institute
for Numerical Analysis, and other government agencies. It’s very unlikely there were any games made
for the SWAC. Next, the Whirlwind 1 was part academic, part
military. It was intended to be used as a flight simulator – not a graphical one, but
as something that could simulate aerodynamics and control flight instrumentation. In addition, it was used for industrial process
control, air traffic control – and to demonstrate the potential of computerised air defence. However, half of the computing time available
was allotted to MIT for engineering and scientific calculations – and here there is evidence
for slightly less serious use. Whirwind’s capabilities were quite remarkable
for the time, with a generously sized CRT display capable of realtime graphics. It was
demonstrated live on American television in 1951, in a Christmas episode of ‘See It Now’.
It even played a rendition of Jingle Bells! Later, in a 1953 film, ‘Making Electrons Count’,
we can see Whirlwind demonstrating a bouncing ball program – it’s the realtime calculation
of three differential equations, but in practice the CRT traces a convincing trajectory of
a ball. This demo dates to as early as 1949. But it
doesn’t count as a video game due to a lack of interactivity. However, some sources claim that there was
a version circa late 1950 to 1951 that did have a degree of interactivity: with the player
able to adjust the ball trajectory in order to make it drop through the hole. Unfortunately, there’s no evidence of this:
the claims are anecdotal, undated – and might be misremembered. So we must presume that
the Whirlwind’s Bouncing Ball was a non-interactive demo, and not a game. Next, the EDSAC at Cambridge – which we already
know played host to OXO, but was far more frequently used for scientific purpose. Atomic wave functions: Astronomical equations
of motion; Crystallography and the structure of myoglobin; and economic modelling. So much more than a simple game of noughts
and crosses, but OXO is what we’re interested in. Although I should mention – OXO is not the
original name of A S Douglas’ program – he referred to it as ‘a game of noughts and crosses’
in his 1954 thesis. (The OXO appellation seems to be a product of a preservation effort by
the EDSAC Replica Project). While Douglas didn’t give an exact date for
the program’s creation, he did specify that it ran ‘circa 1952.’ So for the purpose of primordiality we must
assume the worst: the earliest date that we can be sure of is the end of 1952. If we wanted
to be really cynical, we could take the date that Douglas submitted his thesis – March
1953. In either case, this means that OXO or noughts
& crosses remains the prime candidate – but was there anything else done with the EDSAC
that could take the crown? It turns out there is a whisper of another
game, attested in the thesis of Stanley Gill – which was submitted in November 1952. That would put it ahead of OXO. It was a game of sheep and gates: the screen
vertically divided by a fence, with two gates. A procession of sheep approach these gates,
and you must open the correct one by interrupting a light beam in EDSAC’s tape reader. Primitive, naturally – and sadly, there are
no photos. But it qualifies on the same merits as OXO – and does so with an earlier verifiable
date. But let’s not draw our conclusions yet. More
machines remain. The Pilot ACE at the National Physical Laboratory
in Teddington, London boasted strong floating point performance, ideal for the heavy scientific
computations it was used for. It was a prototype machine that wasn’t particularly
easy to program – but it was still fast, calculating integrals and prime numbers with aplomb. It’s here that I should introduce a man named
Christopher Strachey, a student of mathematics and physics at Cambridge, and by 1949 a schoolmaster
at Harrow. In early 1951, he was introduced to Mike Woodger
at the NPL – and given access to the Pilot ACE machine. His work here was inspired by an June 1950
article in Penguin Science News: ‘A Theory of Chess and Noughts and Crosses’, written
by Donald Davies – the ACE project lead at NPL. The article was a mathematical breakdown of
the decision making and rules that governed games like chess – part of the newly emerging
field of game theory, led by computer scientists like John von Neumann. This was the dawn of artificial intelligence,
and there was a great interest in teaching these so-called ‘electronic brains’ to think:
to teach them how to play games. Strachey was a teacher, and he decided to
undertake this implementation: not noughts and crosses, as that was too simple; Nor chess,
for that was far too complex; Instead, he chose the intermediate game of draughts (or
checkers, if you prefer). With only two types of piece (men and kings),
and only half the squares traversable, it would be feasible given the confines of early
computer capability. Of course, programs for these machines were
planned on paper before being compiled on punched tape and executed, which meant that
Strachey didn’t have the luxury of debuggers, break points, or anything like that – he would
have to write and check his code manually. He finished a preliminary draughts program
by May 1951, but it wasn’t until the 30th July that Strachey had the chance to run his
program for the first time. It didn’t work. However, by this time there were more capable
machines out there, and the computing project at the University of Manchester had made considerable
progress. They had a full-scale production machine delivered in February 1951 – the culmination
of five years of development. The first of the Manchester computers was
the ‘Baby’. This was designed as a testbed, in much the same way as the Pilot ACE – so
its practical use was limited. There were some exercises in pure mathematics:
highest proper factors, long division and ‘other arithmetical facilities’ – but it wasn’t
long before the machine was rebuilt into a more expanded form. The Manchester Mark 1 was a full scale computer
which, while still a prototype, started to see more serious use: Including the calculation of Mersenne primes;
Investigation of the Riemann hypothesis; Symbolic logic; Ray tracing; and Laguerre functions. Even so, it was a short-lived computer, quickly
replaced by its commercial version: the Ferranti Mark 1. Eventually, there would be 9 such
machines made (including a later revision) – but the first Mark 1 went to the University
of Manchester. Here, it served the University’s computing
needs until its replacement in 1959. Notable uses include calculations for the
Armaments Research Establishment – the British Nuclear weapons program – and the development
of ‘Mark 1 Autocode’: a higher-level means of programming the computer that was widely
used. The Ferranti machine also saw some work relating
to games: Dietrich Prinz developed the first ever chess-playing program in November 1951. It wasn’t a video game – it didn’t use the
Williams tube display for any graphical purpose. But it was only a matter of time. Back to Christopher Strachey and his implementation
of draughts: Strachey first learned of the Ferranti Mark 1 in the spring of 1951, and
quickly took the opportunity to write to Alan Turing, describing his interest. In return, Strachey received a copy of Turing’s
infamously opaque Programming Handbook for the Manchester computer, and an invite to
see the machine for himself. He first visited Manchester in July, and over
the next year, between careful study of the handbook and occasional visit – he familiarised
himself with the machine: Ran some programs; And forged a reputation as a formidable programmer. But as much as he had mastered the technical
side, there was something of an artistic streak that ran through his work. He made the computer play music: compose love
letters; And, in the summer of 1952 he successfully ran his revised draughts program – and played
a video game. His implementation worked – the computer was
able to ‘play a complete game of Draughts at a reasonable speed’. Generously, he even took pictures of the video
display and published his work. So, what does this mean? It means the first video game is Christopher
Strachey’s Draughts, first implemented circa 1951 but confirmed working (complete with
evidence) by no later than July, 1952. Also, video games are a British invention. In any case – we have an answer! There are, of course, a couple of caveats:
Firstly, to be clear – Strachey’s draughts is the first video game for which we have
evidence. It remains possible that there was undocumented
work prior – but we can only draw conclusions from the evidence we have. Secondly – our assertion only holds true as
long as you subscribe to the definition we laid out. A looser take might include the
1947 device – or maybe something even earlier. That said, I think our definition is robust. One thing is clear: the 1947 device has had
zero lasting impact on video games, long forgotten until it was dredged up during the Magnavox
lawsuits. Christopher Strachey, on the other hand, was
a bit of a pioneer. His 1952 presentation at the University of Toronto inspired Arthur
Samuel’s later work on machine learning. Strachey developed the CPL programming language, an
important ancestor of C. His influence is broad, and it permeates computer
science. In regards to the first video game, I was
equal parts pleased and dismayed to discover that I wasn’t the first to arrive at this
conclusion. There are two others. One is Jack Copeland: Professor of Philosophy
at the University of Canterbury, Christchurch, New Zealand and an authority on the life and
work of Alan Turing. As a product of his detailed research into
early computing history, and Turing’s correspondence with Strachey – he describes the operation
of Strachey’s Draughts program and calls it ‘the world’s first computer game’. (It’s not the first computer game, but we
can assume he means the first computer game with a video display). The second to draw this conclusion is Alvy
Ray Smith, co-founder of Lucasfilms’ Computer Division and Pixar. His 2015 article, ‘The Dawn of Digital Light’
is a detailed examination of the origins of computer generated digital images that makes
some rather bold statements: The first digital picture; The first computer
animations; The first video game. He ranks Strachey first: Gill’s Sheep and
Gates second; and Douglas’ Noughts and Crosses third. I’m inclined to agree. In broader coverage, Strachey’s draughts is
sometimes included in the early history of video games, but is seldom upheld as the champion.
It’s a difficult call to make, given the lack of a universal definition of ‘video game’
– and the rather ragged margin at its furthest extent. But I feel as though we have our answer. The mid 20th century was a real frontier for
electronics: by 1950, transistors were in their infancy – and integrated circuits were
a decade away. The only place video games could have started prior to the age of Pong
was in the computer lab. Here, there was high technology with no concern
for ‘commercial viability’ – and vast budgets were approved, ostensibly for more serious
computations – but the games slipped through the cracks. But here’s the thing: nobody called these
early computer experiments ‘video games’, at least not at the time. It’s a retroactive label, given only once
video games had pierced popular culture, and the term ‘video game’ dates to later than
you might think. So, one final question: When did video games become ‘video games’? A search on Google Book’s Ngram viewer gives
us an idea of the timeframe: the term emerged at some point between 1970 and 1975, coinciding
with Pong’s early success. This is borne out with a search through the
popular gaming press, scarce as it was in the 1970s – the February 1976 issue of BYTE
Magazine uses the term ‘Pong style video game’ in a familiar way. A broader search of literature on Google Books
doesn’t turn up anything prior to 1975 – however, Google Newspapers does have a 1974 feature
on ‘Space Age Pinball Machines’ that details the rise of ‘the new coin-operated video games’,
which means the term must have started to gain traction by this point. There’s a 1973 article in Business Week that
must have been written just after Pong had found early commercial success: it describes
‘A red-hot market for video games’, in what might be the earliest instance of the term
in a major publication. There are also a couple of false positives
that crop up – a 1974 issue of ‘New Scientist’ uses the term to mean ‘video business’, as
in ‘the video game’: it was around this time that the first video playback devices were
hitting the domestic market. Earlier still, there are some surprising mentions
of ‘video game’ in the 1950s – but it turns out this is an older term for televised sports
games, back when such a thing wasn’t universal. It’s increasingly clear that Pong is close
to the origin, so let’s skip straight to the source – here’s the promotional flyer for
the 1972 Atari arcade game. ‘The Newest 2 Player Video Skill Game’: ‘From
Atari Corporation, Syzygy Engineered – The Team That Pioneered Video Technology’ So in 1972 Pong was described as a ‘Video
Skill Game’, which is almost – but not quite – what we’re looking for. There is a definite
emphasis on the term ‘video’. Atari’s next game: ‘Space Race’, released
July 1973, does use the term ‘video game’ within its flyer text; And later that year in September the flyer
for Pong Doubles featured it prominently, at the very top: ‘Atari’s New Video Game’. So 1973 was the year that the term ‘video
game’ shot to prominence and was fixed within the arcade lexicon – but where’s the first
use? In researching this I found a particularly
good post from The Golden Age Arcade Historian – titled “The etymology of “video game” It confirms our suspicion – Atari and Pong
are very much at the centre of the term, tracking the first use in the trade press to March
1973. Interestingly, there is also a letter from
Nolan Bushnell to John Britz at Bally dated July 10th, 1972 – in which Nolan uses the
term. This might be the first documented use. It
seems ‘video game’ was the term used by Atari to describe their new product to potential
purchasers – venue owners, rather than players. So, two things: the term ‘video game’ dates
to around 1972, reaching wider popularity by 73-74; and it seems it was devised for the benefit
of the arcade amusement industry, as a marketing term to distinguish this new class of machine. So while Nolan Bushnell can’t take credit
for many original ideas – it seems he is the man that named the industry, and the one that
kickstarted its commercial success. There’s an argument to be made that the first
video game was the game first described as such, and it seems the answer to that is Pong. Pong was the catalyst, and Atari’s games are
likely the main reason we use the label ‘video game’ to cover all kinds of interactive entertainment. If Pong had failed and it was the 8-bit microcomputers
that found the earliest success, then ‘computer games’ might have been the definitive term
– in fact, in certain regions at certain times this has been true. But the Americans settled on ‘video game’
– and their influence means most of the world has adopted the term as the standard. It’s far divorced from its original meaning
– a term meant to distinguish early arcade games from their electromechanical counterparts
– but it seems it’s stuck, and has since been applied to all sorts of things: Computer games; Arcade amusements; Television games; Handheld electronic games; Mobile games; And virtual reality. What we know as video games don’t have a single
origin – they’re a constellation. So if anybody asks you, ‘What was the first
ever video game?’ You could tell them it was Strachey’s drafts,
1952 – but it might be easier to say ‘it was probably Pong, or something’. Because, well, it’s trickier than they think. Thank you very much for watching, and until
next time – farewell.