Worldly Wednesdays: The Father of Video Games

This post is part of an ongoing series annotating my book They Create Worlds: The Story of the People and Companies That Shaped the Video Game Industry, Vol. I. It covers material found in the prologue on pages xviii-xx. It is not necessary to have read the book to comprehend and appreciate the post.

Choosing where exactly to start They Create Worlds was a challenge. My goal was to document most of the early experiments using a television and/or a digital computer to play a game, but starting at the chronological beginning of these efforts does not make for a compelling opening. Its great that in the late 1940s Alan Turing and Donald Michie wrote chess programs that were never implemented or that Thomas Goldsmith and Estle Ray Mann liked to pretend in the lab that a cathode ray beam might be the arc of a missile, but there is no throughline from either of these experiments to the $150 billion industry that exists today. These primordial works are examined in the book, of course, but they did not feel appropriate as a hook to draw the reader in. Clearly then, the book should not start at the beginning, but it still needed to focus on a beginning.

So I opened on a bus station in New York City on August 31, 1966, when Ralph Baer thought to himself it might be neat to control objects displayed on a television set rather than passively consuming network programming. Its nice to have a firm date like that to commence the narrative, though its not nearly so firm as one might think. Ralph Baer was a careful record keeper as befit the meticulous, detail-oriented personality that shines through in his various interviews and in his autobiographical examination of his work in the video game business. For this reason, we do know that he transformed his crazy bus station idea into a formal memo on September 1, 1966. In interviews, Baer usually stated he did so back at his office the day after his brainstorm, but the timeframe may not line up. After all, he was down from his native Nashua, New Hampshire, to meet a business client, and Google tells me that’s a good four hour trip in the modern day by car. A 1966 bus was probably taking it even slower than that. Did he really have a meeting with a client in New York City that afternoon and then immediately scurry back up to Nashua? Its not impossible, but maybe a tad improbable. Nevertheless, that’s his story, so we are sticking to it.

Whether this brainstorm happened on August 30 or August 31 though is really a minor matter of little consequence. A more important statement to analyze is the claim I make at the end of this little vignette: “But Baer was the first person to suggest creating an interactive entertainment experience by conveying game data to a display through use of a video signal, so even though he never used the term in any of his subsequent documentation or patents, he is nevertheless the progenitor of what we now call the video game.”

So there it is right? Extra, extra read all about it! Alexander Smith says Ralph Baer invented the video game! Baer himself would have certainly been pleased to see those words in print had he lived long enough to see this book published, as he always claimed the mantle “Father of Video Games” and defended that title against all comers. Repeatedly. And in detail. I don’t begrudge him any of that: the man was absolutely a key cog in the transformation of video games from backroom lab experiments to mass market entertainment, and he lived in the shadow of Nolan Bushnell much longer than he deserved. But did he really, truly invent the video game or have a valid claim to its paternity? Well, despite my glib pronouncement in the prologue of the book, the answer is a little more complicated.

Ralph Baer (L) and Bill Harrison demonstrate their video game prototype. Are they the proud parents of the video game?

Before ruling on Baer’s case, we must decide what the heck constitutes a video game anyway. Ralph Baer would tell us there is a simple technical definition we can go by: if you are playing a game on a screen and that game data was conveyed to said screen by a video signal, then you have a video game; otherwise you don’t.

So then what is a video signal? A video signal is a modulated electromagnetic wave that conveys image data, with the frequency of the wave determining the chrominance, or color, of the image and the amplitude defining the luma, or brightness. This signal provides instructions to the cathode ray tube (CRT) of a television, which focuses a stream of electrons on a single point on a phosphor-coated screen, causing a sustained glow at that point. A magnetic field generated by coils within the CRT allows this beam of electrons to sweep back and forth across the entire screen, one horizontal line at a time, to create a complete picture from this collection of individual dots according to the parameters of the incoming signal. This is the process Baer is describing when he tells us a video game must, by definition, use a video signal.

Right away, Baer’s definition presents a problem by excluding a set of early products that were widely defined by the public and within the industry as video games in their own time: coin-operated vector games like Atari’s Asteroids (1979). The graphics in these games are drawn by a vector generator that takes direct control of the CRT and instructs it to aim at a specific point on the screen and then move on a specific vector until a command is given to deflect the beam in a different direction. The CRT is still receiving and responding to a signal, but it’s not a video signal. There is no doubt, however, that even the most conservative modern definition of a video game would include Asteroids, so Baer’s simple straightforward technical definition must already be set aside.

According to Baer, this is not a video game.

But once we open up the definition, where do we stop? Well, we have to include the vector games clearly, so it logically follows that any time a player interacts with images drawn by a CRT, it counts as a video game. But why stop at a CRT? Modern video games played on a high-definition television or monitor hooked up to a PlayStation 5 or an IBM PC Compatible certainly must count too. While Baer hews to an old-school definition of a video signal that presumes an analog system, digital displays are also driven by a video signal, albeit in a slightly different way. The prime difference between the two is that a digital signal consists of a series of ones and zeroes that provide instructions to draw a complete bitmapped image all at once rather than the analog method in which the image is drawn one scanline at a time. These digital images are pulsed to the television at a specific, constant frequency that determines how many times a second the display will be updated with a new image, the so-called “frame rate” as measured in frames per second (FPS) that is the obsession of high-end graphics connoisseurs. Its still video, so it counts.

So now we know we have a video game whenever someone interacts with objects on a screen, right? Well, not exactly. One problem with merely focusing on the screen is that in the coin-op world, games with “screens” of one form or another existed as early as the 1920s through the use of film projectors. What do we do with driving simulators like Auto Test (1954), shooting games like Nintendo’s Wild Gunman (1974), or even the Nutting Associates Computer Quiz (1967), all of which use a film projector to display images with which the player interacts?

Is Computer Quiz a video game? It has a screen.

Furthermore, what do we do with old computer games that outputted data to a teletype or some other display that does not incorporate a screen? Baer would tell us these are “computer games” rather than “video games” and that these are overlapping, but not identical, categories. Practically speaking, this feels like a meaningless distinction. After all, if one plays Adventure (1977) on a teletype instead of a CRT terminal, is this truly a different experience considering the computer executes the same code and the game proceeds in an identical manner either way?

Fellow video game historian Ethan Johnson and I pondered this topic at length, and he came up with a critical discriminating element. He did a whole blog post about this, but the relevant portion is as follows:

“[R]ather than needing a certain sort of signal, a display for a video game must be arbitrary. This means the display as a whole has to have a direct relation to the program underlying it and is able to achieve a number of different states rather than just “on” or “off”. In the early tic-tac-toe games for instance, while an individual state of a square only has a boolean value, the board as a whole has hundreds of different possible outcomes which are ultimately not pre-determined. The individual state of a screen in Computer Quiz only has two possible variables: Light on or light off, and therefore can not be said to be using a display in the same way as video games.”

Well that does for Computer Quiz at least, but it does not necessarily answer the question for a more complex game like Wild Gunman, and it only gets us a little closer to solving the conundrum of games on early computer systems that lacked a CRT. Furthermore, by opening up our definition to include all computer games with an arbitrary display, we are forced to address how to treat analog computing devices like the Nimatron displayed in 1940 at the World’s Fair, or Claude Shannon’s chess-playing Caissac machine from 1949. These are unquestionably both computers that play games, but does that really make them video games too? Do we now extend the history of video games all the way back to 1912 and the Spanish El Ajedrecista chess-playing automaton? Clearly, we need to establish some other limiting criteria.

El Ajedrecista, the analog computer that could figure out how to mate a lone king with a king and a rook. Is this the beginning of video games?

The easiest way to distinguish these edge cases from video games is to distinguish between the internal components that generate the game elements. A game like Wild Gunman uses electro-mechanical components, that is wipers, switches, and contacts facilitate the completion of electrical circuits to create playfield action by powering relays, steppers, and other mechanical parts. All video games by the Baer definition, including his own Magnavox Odyssey and Atari’s Pong (1972), use electronic components instead, with streams of electrons directed through a series of logic gates determining what happens over the course of the game. This allows us to distinguish not only electro-mechanical coin-operated games with screens from video games, but also allows us to remove early electro-mechanical analog computing devices from the equation.

Now that we have defined two critical technological elements, we need to add a couple of conceptual components to complete our working definition of the term video game. First, we need to define the user’s place in this interplay between logic circuits and a display. The easiest way to do this is focus on the commonly accepted definition of “playing a game,” which requires active participants rather than passive viewers. For video games, this means the game needs to unfold through direct user interaction via a control scheme allowing the players to directly manipulate objects on the display. There is really no need to elaborate on this element any further: so long as this interaction is happening, the how of it is unimportant.

Finally, we need to define exactly what interactions between a user, some electronic logic, and a display constitute playing a game. If we don’t, then a word processor or a DVD menu is just as much a video game as Pong. The best we can do here is define a video game, which is generally understood to be a leisure activity, as a product intended to provide entertainment. This is the most subjective part of our definition because different people find different things entertaining and even a DVD menu could be turned into a game by a particularly bored individual. The best we can do is point to the intrinsic purpose of the product as determined by authorial intent. If the program was produced or marketed with the primary goal of entertaining a person, then its a game. If the entertainment value is secondary to serving some other function, then it’s not. This is not a perfect test. For example, a product primarily designed to educate might also be deliberately crafted to entertain to hold the user’s interest. More work needs to be done on this element of the definition to clarify gray areas, but I will leave that for others to work out.

Now we have a serviceable, though still imperfect, definition of a video game that eliminates many, though not all, of the edge cases: A video game is an entertainment product played on a device containing electronic logic circuits in which the players interact with objects rendered on an arbitrary display.

Sorry Nimatron, you are not a video game.

So now that we have identified the child, who is the father? There are a few ways to look at this. One is to employ our newly articulated definition and look for the first product that meets all these criteria. That might lead us to 1947 and the prototype cathode-ray tube amusement device (CRTAD) patented by Estle Ray Mann and Thomas Goldsmith. I personally feel CRTAD does not really hold together under our definition of a video game, but that is a discussion for another time and another annotation. Regardless, I feel comfortable ruling that these two engineers are not the fathers because they probably never built a finished product, certainly never displayed the system publicly, and did not influence any of the projects that came afterwards. By the same logic, we can also dismiss the dueling chess AIs created by Michie and Turing in 1948, which were complete computer programs on paper, but were never implemented on an actual computer.

So if the first conceived games do not make the cut, what about the first fully operational and publicly displayed product? That would lead us to Bertie the Brain, a custom tic-tac-toe computer built by Josef Kates and demoed in 1950. There is no doubt that this is the earliest known publicly played device that meets all our criteria for a video game, but is being first really all its cracked up to be? Bertie was displayed at a single Canadian trade show and received virtually no press. It may have been played by a decent number of people — the show draws over one million attendees every year — but it did not stick in the collective memory and was only rediscovered by scholars in the 2010s. Furthermore, it was solely intended to demonstrate the workings of a new type of vacuum tube and was not marketed as a new form of entertainment. Once again, I think our father — an appropriate term only because all our early pioneers in this field were men — needs to do more than bring a simulation into the world; he needs to understand he is creating something that could change the face of entertainment. Clearly, Kates wanted the attendees to be entertained while using his computer, but that is not quite the same thing.

Sorry Messieurs Goldsmith, Turing, Michie, and Kates. You are not the father.

So how about that master of physics and entertainment, “Wonderful Willie” Higinbotham? There is a solid case to be made that his tennis game, retroactively dubbed Tennis for Two (1958) by historians, marked the first time a video game was created solely to entertain the public. Therefore, he is our first real contender for the title “father of video games.” Once again though, I believe we need to exclude him because he did not start a wider movement. Our father is no good to us if his child failed to have children of its own.

So what about the first entertainment program that could be acquired by the general public? Right now, the earliest known game to fit that definition is a baseball simulation created by IBM employee John Burgeson in 1960-61 and briefly requestable as part of the program library for the IBM 1620 computer before being withdrawn from the catalog in 1963. There are a couple of problems here. First, this program only barely meets our definition of a video game because the only player interaction happens at the beginning when creating a team by selecting from a pool of players. More importantly, it appeared and vanished so quickly that it failed to have any sort of impact.

Then maybe its Steve Russell et al. and their Spacewar! (1962), which certainly achieved popularity across a select group of universities and research institutions and itself birthed the first commercial video game, Computer Space (1971)? Now I think we are getting closer. Baer would discount this game because it uses a point-plotting display, which functions in essentially the same manner as a vector monitor except that instead of drawing lines it draws each point individually. As Baer might say, “no video signal, no video game.” But we have already moved past that narrow definition. The main strike against the game is that it did remain confined to a small number of locations and was not commercialized. One could argue that since video games did not capture the imagination of the general public until commercial models were available that anyone could gain access to for a reasonable price, then our father needs to be someone that brought video games into the mass market. I find that argument flimsy, but it can be made.

Sorry Willie, you are not the father. Steve, we’ll get back to you…

So now we come at last to the final two contenders, Ralph Baer and Nolan Bushnell. Among the general public, I think the debate really comes down to just these two. The controversy over which of them birthed the video game has literally existed for as long as people have written about video game history, with Steven Bloom’s monograph Video Invaders debating this very topic as early as 1982. Both have strong claims to the title. Nolan Bushnell came to market first with Computer Space, but Ralph Baer started work on his system earlier and had largely completed it by 1968. Bushnell also debuted the first successful product, Pong, but the game only came about because he saw the table tennis game on the Odyssey.

Which person one prefers really depends on how you define the parameters. Is it first conceived that matters or first released? Is it enough to dream up a system, or does said system also need to capture the public’s imagination? Certainly Baer and Bushnell themselves expended most of their energy trying to prove who came up with the idea of commercializing video games first during a series of patent lawsuits in the 1970s. Baer, with that meticulous streak, was able to provide a plethora of verified documentation from 1966-72 elucidating every step along the way from initial spark to final product. Bushnell could only counter this by claiming he wrote a paper in college in the 1960s about playing games on a computer after he saw Spacewar!. When asked to submit the paper as evidence, he proved unable to do so. The courts rightly sided with Baer, but winning a patent suit is not quite the same thing as winning a paternity suit.

Clash of the Titans. Ralph Baer and Nolan Bushnell duke it out for the title Father of Video Games in this drawing by Howard Cruse found in the book Video Invaders by Steve Bloom.

So now that we have a video game definition and a list of the major contenders for our parental figure, is Ralph Baer the “progenitor of what we now call the video game”? Not really. I feel the video game really has two sets of parents, Russell and friends, who created the first video game to gain a significant following across multiple installations, and Bushnell and his partner Ted Dabney, who were inspired by the work of the Russell group to engineer the first commercial video game product. This leaves Baer the odd man out despite the pride of place I gave him in the book. Baer himself would have certainly not been pleased to see these words in print had he lived long enough to see this blog post published. That said, he really was the first person to follow through on the idea that manipulating objects on a standard television set could be fun; he was the first person to realize it was possible to create a hardware system to do so that was cheap enough it could be commercialized for home use, and he worked out how to interface this hardware system with a television set using an RF modulator and a video signal. These were the building blocks upon which the entire home video game industry was built, and that in itself is a monumental achievement. So while I am not entirely comfortable calling Baer the “father of video games,” I will gladly cede him the title “father of the video game console” and give him pride of place at the beginning of my three-volume history. Baer’s bus stop brainstorm may not have been the beginning, but there is no doubt it was quite a beginning.

They Create Worlds: The Story of the People and Companies That Shaped the Video Game Industry, Vol. I 1971-1982 is available in print or electronically direct from the publisher, CRC Press, as well as through Amazon and other major online retailers.

Advertisement

A Nutty Idea

In March 1971, Nolan Bushnell left Ampex to become the chief engineer of Nutting Associates and finally achieve his dream of building a commercial video game.  Over the next five or six months, he engineered a game so advanced, it would take at least three years for any other product to approach its combination of representational graphics, physics, and artificial intelligence.  Debuted at the MOA show that October in a sleek fiberglass cabinet, Computer Space conveyed a firm statement that the future was here and the coin-op industry would never be the same.  Unfortunately, the game also proved alien to distributors and intimidating to the general public.  While the success or lack thereof of Computer Space on the market continues to be a point of debate today, there is no doubt that it failed in its overall goal of igniting the video game revolution.  While it sold about as many units as a typical pinball machine, it failed to spark much excitement from the established manufacturers, and Bushnell was forced to return to the drawing board.

Nutting Associates

Bill Nutting, founder and president of Nutting Associates

William Gilbert Nutting was born on May 3, 1926, and grew up in River Forest, Illinois.  The son and grandson of executives of the Marshal Fields Department Store, Nutting graduated high school in the middle of World War II and entered Army Air Corps cadet training, where he began a life long love affair with flying.  According to a profile written in the March 1992 issue of Vintage Airplane, Nutting attended Colgate University for two years after the War and then transferred to Colorado University, where his childhood friend Claire Ullman also attended school.  In 1948, Bill and Claire were married, and two years later Bill graduated with a degree in business administration.  As stated in a profile in the February 17, 1968, edition of Cash Box, Bill and Claire then relocated to San Francisco, where Bill took a trainee position with the National Motor Bearing Company and then joined Rheem Manufacturing in 1951, where he moved through a variety of jobs from production line foreman to inventory control to purchasing and sales.  In 1956, he relocated to Rheem’s Chicago office for a sales and office management position, but in 1959 he returned to California to enter the retail business.  According to Goldberg and Vendel in Business is Fun, Nutting’s retail work had him following in his father’s and grandfather’s footsteps by taking a job at the San Francisco luxury department store Raphael Weill & Company — also known as the White House for its beaux-arts facade — where he started as a buyer in the gloves department.

According Cash Box, in 1962 Nutting invested in Edex Teaching Machines, a company established by Eugene Kleiner that same year after leaving Fairchild that built educational devices for the United States Military and other clients.  As revealed to me in an interview with Claire Nutting, the connection between Nutting and Kleiner was Claire’s father, a high-level executive at Revereware.  Claire’s father became friends with Kleiner and invested in Edex, and he invited Bill to invest as well.  One of Edex’s products was the Knowledge Computer, which according to a patent issued in 1965 was designed by engineer Thomas Nisbet and was a coin-operated multiple-choice question and answer machine that incorporated a film projector and buttons for selecting answers.  As explained in Cash Box, Nutting took over marketing of the product, and the October 24, 1964, issue of Billboard Magazine reports that Nutting and a man named Howard Starr attended the 1964 MOA show under the name Scientific Amusements — which judging by the October 17, 1964, edition of the magazine was a subsidiary of Edex — to display the Knowledge Computer, most likely to gauge the interest of coin-op distributors.  A small number of machines may have even been placed on location, as an article in the August 22, 1964, issue of Cash Box states the game can already be found in “bowling alleys, student unions, and transportation depots.”

Raytheon purchased Edex the next year.  According to Claire, Raytheon had no interest in the coin-op business and Bill had long been interested in starting his own company, so he purchased the rights to the machine so he could start marketing it himself.  According to the October 23, 1965, edition of Billboard Magazine Bill Nutting marketed the Knowledge Computer through an entity called Nutting Corporation and worked with distributor Advance Automatic Sales to place machines with 20-25 operators in the Bay area.  As explained in the February 17, 1968, issue of Cash Box, however, it soon became apparent that the game was too expensive and too hard to service to be a viable product.

 According to an article by Goldberg in issue 136 of Retro Gamer, in about 1966 Bill therefore contacted his brother Dave, then working as an industrial designer for the celebrated firm Brooks Stevens Design Associates, to ask for help in redesigning the Knowledge Computer. (Note:  Dave Nutting’s relatively short, but highly influential career in the coin-op industry necessitates a fuller biography than the scope of this post allows, so his background will be covered in more detail in a later post.)  According to Dave as relayed to Goldberg, the brothers agreed that Dave would design and manufacture the game in his home city of Milwaukee, while Bill would concentrate on selling the game from his home in California.  According to Dave as relayed to several authors including Goldberg and Donovan and also relayed in interviews with Dave’s partners Harold Montgomery and Gene Wagner conducted by Ethan Johnson, Claire did not like the idea of the brothers going into business together and insisted Bill call the deal off.  This may not, however, be entirely accurate.  According to a letter written by Bill to his son Craig that was shared with me by Claire Nutting, Dave had wanted to join Bill’s business and redesign the Knowledge Computer, but Bill ultimately turned him down for personal reasons due to a long-standing sibling rivalry.  Bill states that Claire was involved in the decision as a member of the board of the company, but that it was ultimately Bill’s decision alone.  Dave’s contention that Claire threatened to divorce Bill if he did not call off the arrangement likewise appears unfounded.

To redesign the Knowledge Computer, Bill established a new company called Nutting Associates in January 1966, and approached a company called Marketing Services for design help.  (Note: some sources claim the company was established in 1968, but the January 1966 date is confirmed by an employee handbook given to me by Claire Nutting that includes a brief history of the company.  He incorporated the company in February 1967 according to the company’s articles of incorporation).  That company assigned an industrial designer named Richard Ball to the project.  According to this author’s interview with Ball, he decided to place Nutting’s unit on test at the College of San Mateo and was amazed when he emptied the machine five days later and discovered it was filled to the brim with dimes.  Sensing a hit, Ball subsequently redesigned the game for easier manufacturing, most notably by building a new projector.  According to the company employee handbook, Nutting then spent a year placing units on location through a franchise system under the name Computer Quiz before moving to traditional coin-op distribution at the MOA show in 1967.  According to Ball, this first version of the game relied on copper relays, which created extreme service headaches.  Ball therefore approached a company called Applied Technology to design a circuit board that would accept plug-in relays.  The next summer, Ball worked with Applied Technology to redesign the game again to eliminate relays entirely. This third generation of Computer Quiz , released according to Cash Box in October 1968, may well have been the first fully solid-state arcade game ever created.

Computer Quiz from Nutting Associates, perhaps the first solid-state arcade game

Computer Quiz hit the industry at the perfect time.  Sega had already proven that operators would be willing to accept a larger, more sophisticated game far different from the traditional pinball table and had also established that the public would accept a machine set for quarter play.  Furthermore, the game launched during a period when the industry was still fighting for legitimacy due to the continuing stigma of association with organized crime, so a game with a perceived educational value proved a perfect lead-in product for locations that would not accept pin games or similar amusements.  As a result, in a time when the typical pinball game might sell 1,500 units and a hit would only sell around 3,000 both Goldberg and Donovan report that Nutting sold 4,400 units of Computer Quiz, while Nutting Industries sold another 3,600 units of Dave’s version, which he called I.Q. Computer.  Along with Periscope and Speedway, Computer Quiz played a critical role in the arcade game renaissance of the late 1960s that ultimately resulted in the birth of the video arcade game industry.

While Computer Quiz proved a sizable hit, Nutting Associates soon ran into difficulties.  According to former company general manager Ransom White as told to coin-op historian Keith Smith, the company was only marginally profitable, perhaps due to Nutting’s decision to bypass distributors and sell to operators directly in the early days, many of whom White felt were shady characters.  This view is supported by Ball, who came away from his time with Nutting believing that the entire coin-op industry was controlled by the mob.  Furthermore, according to Ball the company never had a particularly competent engineering department — after all, they had to rely on an intern to even build their first solid-state design — which made coming up with a follow up product tricky.  According to Ball, he warned Nutting in 1968 that Computer Quiz would soon run its course and created a marketing brief for a video product to replace it, but he and White were both fired soon after for objecting to Nutting buying an airplane using company funds.  (Note: No one else has come forward to corroborate Ball’s claim that he proposed some form of video game to Nutting in 1968, but if he did, this could explain why Nutting would be willing to experiment with the concept with Bushnell in 1971.)

With White, Ball, and sales executive Lance Hailstone departing, Nutting was forced to refresh his executive staff and hired Rod Geiman to serve as executive vice president and Dave Ralstin to serve as sales manager.  According to this author’s interview with Bushnell, before long Nutting himself was spending most of his time tinkering with his planes while Geiman ran Nutting Associates.  According to Billboard Magazine, in December 1968 Nutting moved from its 4,500 square foot facility to a much larger 18,500 square foot location as Geiman prepared to expand the company into new markets.  The next year, the company launched a sports trivia game called Sports World and a horrorscope machine called Astro Computer, but the company’s main business remained releasing updated questions for Computer Quiz, which could not sustain the company much longer.  Therefore, when Nolan Bushnell showed up in early 1971 with his own game idea and agreed to become the chief engineer the company desperately needed, Nutting lept at the chance to hire him.

Creating Computer Space

The memory circuit board for Computer Space.  Note the diodes laid out in the shape of the spaceships.

When Bushnell joined Nutting, he had a basic hardware system that could move dots around the screen and a vague idea to make a product like Spacewar!, but he did not yet have a fully formed game concept.  In fact, according to his 1976 deposition at one point he even toyed with moving to a simpler concept by just designing a piloting game in which the player controlled a spaceship and dodged asteroids.  In the end though, he decided to press on with a shooting game, which by the end of January 1971 had acquired the working title Cosmic Combat.  The limitations of the hardware necessitated some changes from Spacewar!, however.  Gone was the sun and its gravity, the hyperspace function, and two-player combat.  Instead, Bushnell crafted a game in which the player controls a rocket ship and attempts to shoot down two flying saucers that also shoot back at him.  Both the player and the saucers score a point each time they destroy one another.  The game lasts for ninety seconds.  If the player has more points than the saucers at the end of the round, he gets another ninety seconds of play; otherwise the game ends. (Note: In the Ultimate History of Video Games, Kent claims that Computer Space played just like Spacewar! complete with gravity and hyperspace, but just observing footage of the game demonstrates this is clearly false.)

While there is a great deal of disagreement between Bushnell and Dabney over who did what when building the motion control prototype, there is much less controversy over the creation of the game itself.  As described by Goldberg and Vendel, Bushnell’s deal with Nutting was that he would join the company as chief engineer and work on Nutting products during business hours, while working on his video game after hours and on weekends.  Bushnell insisted on this arrangement so that Nutting would not acquire shop rights to his video game technology, which remained the property of Syzygy Engineering.  When the game was completed, Nutting would manufacture and sell the game and give Syzygy a five percent royalty on each cabinet sold.  Dabney, meanwhile, remained at Ampex until the summer, when Bushnell’s progress convinced him they would be able to make a go of it, and he resigned to join Nutting as well.

Once work on the game shifted to Nutting, the majority of the engineering was completed by Bushnell.  According to Benj Edwards’s article on the development of the game, Bushnell spent his days hunched over a drafting table just outside his office door at Nutting plotting out the circuitry that would tell the spot generator where to place dots on the screen and how they should interact with the player’s controls.  He also created the graphics for the game, rendering the player’s ship and the enemy saucers as a series of dots and creating a series of routines that allowed them to rotate smoothly.  In one of his more clever feats of engineering, Bushnell used mirroring techniques so that he would only have to store four different ship positions in memory rather than the sixteen needed to cover every possible facing.  He also chose to lay out the diodes used for the graphical memory — mask ROM being far too expensive at the time — in the shape of the ships themselves, which allowed operators to easily figure out which diode needed to be replaced in case of malfunction.  Finally, Bushnell crafted the AI of the hardware-controlled opponents by dividing the playfield into quadrants and giving the saucers the ability to detect which quadrant the rocket ship was currently in so they would fire in that direction.

In one of the rare cases of agreement between Bushnell and Dabney on engineering matters, Dabney concurs that Bushnell completed all the work outlined above, though he does claim that Bushnell came to him for advice on how to implement several features.  However, Dabney also claims in his Retro Gamer Roundup interview that Bushnell had help in designing the circuits from Steve Bristow, a young engineer interning at Ampex who later became a key Atari employee.  In both an interview in Retro Gamer issue 75 and in a Computer Space retrospective in issue 93, Bristow has confirmed that he did help build the motion control and memory boards used in the game, but in Edwards’s article he concurs with Bushnell that he was not involved in any of the design or layout, just the construction.

According to Edwards, Dabney concentrated on the mechanical and analog engineering required to turn the game into a finished product, building a power supply, developing a working coin mechanism, implementing the controls, and constructing a wooden cabinet in which to house the game.  According to both Edwards and Retro Gamer 93, he also developed the sound for the game by taking a voltage regulating diode that generated pink noise and attaching an amplifier and integrator that allowed for changes in volume.  According to his Retro Gaming Roundup interview, Dabney also came up with the idea of inverting the video every time the player cleared a round to provide a sense of progression.  The duo identified the reverse-color screen as “hyperspace.”  According to Goldberg and Vendel, Bill Nutting provided the final name for the game, Computer Space — a variation on the name of the company’s first hit, Computer Quiz — while Edwards claims it was Ralstin.

According to Goldberg and Vendel, in August 1971 the game was far enough along that Bushnell and Dabney decided to do a location test, an important step in the arcade industry in which a prototype game is placed on location and the coin-drop is measured to see if the game is shaping up to be a hit.  (Note: Both Donovan and Retro Gamer place the test in November 1971, but this is clearly far too late: Nutting was already taking orders for the game at the MOA show in October, and location tests always occur before the game is made available to distributors.  These sources have apparently confused the location test date with the most commonly claimed general release date.)  Dave Ralstin ran a coin route on the side to generate extra income, so Nutting chose to place the game in a bar on the route frequented by students of Stanford University called the Dutch Goose.  Packing the prototype unit in Dabney’s Datsun pickup truck, Bushnell and Dabney brought the game to the bar and watched as players flocked to the machine.  It looked like the duo had a hit, but a second test at a pizza place did not go nearly so well.  Like Spacewar!, Computer Space used a multi-button control scheme and realistically depicted the physics of movement in a zero-g environment — in which an object continues to move in the same direction until a force is exerted in the opposite direction.  The Stanford engineering students at the Goose, some of whom were probably Spacewar! veterans, caught on right away.  The working class patrons at the pizza place did not.

Galaxy Game

Bill Pitts (l) and Hugh Tuck working on the first version of Galaxy Game

Not long before putting Computer Space out on test, Bushnell learned that he was not the only person working on an arcade adaptation of Spacewar!, though after meeting his competition and observing their work, he decided they were not a threat.  According to a testimonial he wrote for the Stanford alumni magazine, Bill Pitts was a Palo Alto native interested in chemistry and physics who matriculated to Stanford University in the Fall of 1964.  Pitts quickly signed up for an electrical engineering class and received his formal introduction to computers when an EE professor discovered his interests and helped him push back his “History of Western Civilization” requirement so he could take the brand new “Introduction to Computer Science” course being offered by George Forsythe.  According to an interview conducted by coin-op historian Keith Smith for his book All in Color for a Quarter, Pitts ultimately graduated with a degree in statistics, Stanford not having a computer science degree at the time, but he had the opportunity to take high-level courses in computer programming because the statistics department allowed students to take graduate courses as part of their undergraduate degree in an effort to recruit more students to the major.

According to Donovan and Smith, Pitts also enjoyed the unusual hobbies of breaking into university buildings and exploring the maze of steam tunnels beneath the campus.  As relayed to Donovan, in 1966 Pitts was en route to a bar when he passed a driveway leading off into the hills about five miles from the center of campus.  From the sign at the foot of the drive, he could tell it led to a university building, and it was one that he had never broken into before.  He returned several hours later with the intent of breaking and entering only to discover all the doors were unlocked.  Entering the building, Pitts observed brightly lit rooms and a PDP-6 computer.  He had unwittingly just discovered the Stanford Artificial Intelligence Laboratory.  Already enamored with computers, Pitts convinced Lester Earnest, the man in charge of overseeing SAIL, to allow him to log computer time when no one else was using the system.  Soon, classes were completely forgotten as Pitts began engaging in all night coding sessions.

According to Smith, Pitts had already been exposed to Spacewar! on a PDP-1 in Polya Hall before he discovered SAIL, but hanging around the AI lab gave him the opportunity to play it more frequently and share the game with others.  As explained by Donovan and the John Markoff book What the Doormouse Said, one person with whom Pitts played the game was his high school buddy Hugh Tuck, whom he would take up to the lab whenever he was back in town from California Polytechnic State University in San Luis Obispo.  According to Donovan and Markoff, Tuck remarked one night that if someone could stick a coin slot on the computer, that person could probably become rich. (Note: Donovan claims this exchange took place in 1966, while Markoff claims 1969.  In his alumni magazine testimonial, Pitts claims the exchange took place three years before he started working at Lockheed, which would be 1968 if Markoff’s date for his employment is correct.  1968 seems more reasonable than 1969 if only because Pitts graduated in 1968 and may not have been hanging out in the AI lab anymore.  Either way, Donovan has probably placed the event too early.) There was no way to package a PDP-6 and a display cheaply enough to create a commercial game, however, so the duo never tried to take the idea any further.  In 1971, however, Pitts took a job at Lockheed as a PDP-10 programmer, but had to wait until the computer actually arrived before he could do any work.  During this idle time, Pitts noticed that the year before, DEC had released a new computer called the PDP-11.  Designed as a successor to the PDP-8 and initiated after the PDP-X debacle and the defection of Ed De Castro to start Data General, the 16-bit PDP-11 occupied roughly the same niche as the Nova and retailed for just $12,000.  Pitts thought back to Tuck’s idea to recreate Spacewar! as an arcade game and decided that with a PDP-11 it just might work.

From Smith, the hardware Pitts used to create the game consisted of a PDP-11/20 with 8K of memory, a Hewlett Packard 1300A Electrostatic Display, and a point-plotting display interface designed by a man named Ted Panofsky. A coin box was provided by jukebox company Rowe International, while Pitts purchased used B-52 joysticks from San Carlos military surplus store J&H Outlet for use as controllers, commercial joysticks being virtually nonexistent at that time.  According to Donovan, Pitts assembled this hardware system and programmed the game, while Tuck, a mechanical engineer, designed the cabinet, which was actually engineered by a firm in Palo Alto.  Tuck, who came from a wealthy family, also provided the bulk of the $20,000 required to build the machine.  According to Donovan, the funding came from Tuck’s parents, but in Pitts’s Stanford testimonial, he says that he, Tuck, and Tuck’s brothers and sisters provided the cash.  As the game neared completion, Pitts and Tuck established a company called Computer Recreations, Inc. in June 1971 in anticipation of selling it.  Originally called Spacewar! just like the original, Pitts changed the name at the last minute to Galaxy Game due to the profound antiwar sentiment currently pervading area college campuses.

The second version of Galaxy Game, completed in 1972

According to a business plan written by Pitts and Tuck in early 1972, they installed the first version of Galaxy Game in the Tresidder Union on the Stanford University campus in a music listening room on the second floor in November 1971.  The massive walnut cabinet included seats for the player to encourage extended playing sessions and housed the monitor and controls.  The PDP-11, meanwhile, resided in the attic and was connected to the cabinet by a 100-foot cable. Unlike Computer Space, Galaxy Game was a faithful recreation of Spacewar! complete with hyperspace and the central sun and even allowed the player to choose whether or not to use the sun, whether to use faster or slower ships and faster or slower torpedoes, and whether or not to allow ships to wrap around to the other side of the screen.  Play was set at a dime per game or three games for a quarter.  According to Pitts, the game was an instant hit as long lines of eager players waited an hour or more for a chance on the machine.  In an email correspondence with historian Marty Goldberg, Pitts said they even briefly installed a second monitor hanging above the cabinet so people in line and other passers by could observe the unfolding action.

As Pitts told Smith, the initial version of the game was never meant to be a commercial product due to its cost: its only purpose was to gauge public interest in a Spacewar! arcade game.  With the prototype proving a hit, Pitts and Tuck progressed to version two.  This version would be placed in a fancier blue fiberglass cabinet, while the computer would drive four monitors instead of just one in order to make the whole system cost effective.  These monitors could each run separate games, or they could be linked to allow more than two players to play in the same game.  Building the new system ran the cost of the entire project to $60,000.  In June 1972, this version was installed in a coffee house in the Tresidder Union, though it had to be cut down to only two monitors to fit into the space allotted by the University.  (Note: The June 1972 date reported by Smith comes from a summary of the game’s creation posted by Pitts on a Stanford website in 1997.  At the 2013 California Extreme Show, Pitts gives a September 1972 date instead, but as this talk came another sixteen years after the fact, I find this claim less reliable.  My guess is he was confusing the installation of the first version, which Pitts has erroneously claimed happened in September 1971, with the installation of version two.)  Once version two went live at Stanford, Pitts carted version one to other locations around town, but it never did as well as the installation on campus.

According to Pitts’s 1997 testimonial, version two remained in operation until May 1979, when it had to be retired because the display processor had become unreliable.  According to Pitts, the game remained popular right up until the end, with clusters of a dozen or more students gathered around the game on any given Friday or Saturday night during the school year.  He had long since given up on turning Galaxy Game into a commercial product by then, however, as even running multiple games off a single PDP-11 resulted in a price tag that was simply too high.  Nevertheless, Pitts claimed to Smith that by the time Galaxy Game had been retired in 1979, he had recouped the $60,000 cost of developing the game.

Selling Computer Space

Computer Space, which shipped in a variety of colors, in its final form

Roughly a month before Galaxy Game debuted at the Tresidder Union, the Nutting Associates team flew to Chicago to attend the MOA show, held that year from October 15-17 at the Conrad Hilton Hotel.  The original prototype debuted at the Dutch Goose had been housed in a simple wooden cabinet built by Dabney, but by the MOA, the game now sported a futuristic-looking fiberglass cabinet, which according to Edwards was designed by Bushnell using modelling clay and then built by a seamless swimming pool manufacturer named John Hebbler located by Dabney.  The controls were placed on a lighted panel jutting out from the cabinet and consisted of four buttons — two for left and right rotation, one for thrust, and one for firing.  According to Goldberg and Vendel, after initial tests demonstrated the buttons might be too complicated a control scheme Bushnell and Dabney had hoped to use a metal joystick-like device for movement, but it proved too fragile and broke the first night that version went out on test.  The panel bore the Nutting Associates name and logo, but in a nod to the game’s creators, it also included the phrase “Syzygy engineered.”

According to Edwards, Nutting brought four cabinets to the MOA show in an attempt to make it appear the game was already in production, though in truth these were the only four copies of the game in existence.  Each game was housed in a different color cabinet — yellow, red, white, and blue.  Disaster nearly struck when they discovered the monitors had all broken loose from their cabinets during shipping, but Bushnell and Dabney were able to repair three of the units.  The fourth was left open as a display of the internal components of the system, a clever ploy to mask the accident from distributors.  How well the game performed at the show is a matter of some debate.  In his 1976 deposition, Bushnell claims no orders were taken at the show.  When speaking to Edwards in 2011, however, he claimed that distributors felt it would be worthwhile to try the game out, and they came away with a “good order book.”  Goldberg and Vendel claim that Nutting took a handful of orders.  All sources agree, however, that no one quite knew what to make of the game.  In Steve Bloom’s Video Invaders, Bill Nutting claims the game “blew the industry’s mind,” while Goldberg and Vendel emphasize that distributors were skeptical of the game’s reliability and play value while also fearing that hoodlums would steal the TV right out of the cabinet.  The game must have generated at least some interest, however, as Goldberg and Vendel report Nutting took Bushnell and Dabney for a flight in one of his planes to celebrate after they returned to California and ordered a respectable production run.

According to Smith, Nutting displayed the game one more time at the International Association of Amusement Parks and Attractions show from November 9-12 and then began shipping the game before the end of the year.  According to most sources, including Edwards, the game shipped in November, though in his 1974 deposition, Bushnell remembers the game shipping in December, or even in early January 1972.  Kent claims in his book that a man named Keith Feinstein acquired documents that prove the game was shipping before the end of 1971, but he does not elaborate.  The November 27, 1971, edition of Cash Box magazine contains an ad for the game claiming it is already available from distributors, strongly implying it had already been released by that point, but an article in the December 4 issue states the game is “being readied for U.S. Distribution,” which implies it is still forthcoming.

The number of units built also varies depending on who tells the tale.  Bushnell stated in Retro Gamer that they produced around 2,200 units, but Nutting claimed to Bloom in Video Invaders that he built just 1,500 units.  As this interview came closer to the events in question and the source was the person actually in charge of the company, I tend to believe the lower figure.  Either way, this was an ambitious initial manufacturing run in a time when 2,000 units constituted a decent hit.

According to Nutting in Video Invaders, Computer Space faced a mixed reception, with some of the more more progressive distributors interested in the game, but most feeling the game held little merit.  This claim is backed up by individual anecdotes.  For example, Smith reports that Portale Automatic Sales in Los Angeles became the game’s biggest champion and largest distributor, while Donovan reports that Empire Distributing in Chicago, one of the largest distributors in the country, felt video games were a passing fad and had no interest in the game.  According to Goldberg and Vendel, Dave Ralstin overcame some initial skepticism by giving away the first five machines off the assembly line to five of the largest distributors in the country.  The ploy must have worked, for Goldberg and Vendel report that by spring 1972 the game had sold 1,000 units.  Both Bushnell and Dabney have claimed in separate interviews, however, that with the game starting to become successful and Ralstin earning a nice commission on each unit, Nutting decided to dismiss him and sell the game himself.  According to Goldberg and Vendel, sales dropped off quickly after that.  This story may not be strictly true, however.  While sourcing on Nutting Associates is difficult to come by, the May 13, 1972, issue of Cash Box identifies Ralstin as still being the sales manager at Nutting.  While its still certainly possible that Ralstin was dismissed because Nutting felt his commission was too large and his services were not needed, I imagine that Computer Space sales were probably largely complete by that point anyway and that his dismissal therefore had little effect on the fortunes of the game.  On the other hand, Bob Portale praises sales of Computer Space in the same issue, so its possible that the game was still selling, at least in some parts of the country.

Final sales figures for Computer Space are not known.  Bushnell stated to Retro Gamer that he believed they sold 2,200 units, which seems high, but in his 1976 deposition, he placed the number at between 1,300 and 1,500 units.  Goldberg and Vendel claim 1,500 units as well, while Edwards claims between 500 and 1,000 units, which is almost certainly too low.  Kent claims Nutting built 1,500 units but failed to sell them all.  Considering Nutting’s claim of a 1,500 production run and some difficulty enticing distributors to buy the game, Bushnell’s 1976 estimate of 1,300 to 1,500 units appears to be the most accurate.

So was Computer Space a success?  That is a difficult question to answer.  As Donovan points out, sales of 1,500 units were nothing to be ashamed of in the early 1970s, and Bushnell has claimed in several interviews that the game grossed $3 million, though this figure is probably a little high.  According to Edwards, Dabney remembers being disappointed by the game’s performance, yet Bushnell remembers it as a modest success, though one he felt could have been bigger.  In his memoir Lucky That Way, former Activision producer Brad Fregger recalls being hired by his friend Rod Geiman to collect coins along Nutting’s game route and being impressed by how many quarters Computer Space took in.  Nutting must not have been too disappointed with the game’s performance either, as he ultimately negotiated with Bushnell to create a two-player version of the game, and a flop does not get a sequel.

On the other hand, the game only sold modestly well and failed to excite most distributors.  Nutting claims to Bloom that he had to “force” some companies to take the game, while Kent states he could not sell the whole production run.  Also, while 1,300 to 1,500 units was not a bad showing for the period, it paled in comparison to the biggest hits of the day like Speedway and Nutting’s own Computer Quiz.  According to Bushnell as told to authors from Kent to Edwards, working-class bar patrons, one of the coin-op industry’s most important demographics in those days, were unable to grasp the complex controls and realistic physics and soon tired of the game.  Furthermore, in an industry where a hit product was guaranteed to be knocked off a dozen times over, only one other company, a small Burbank firm called For Play Manufacturing, ever created a Computer Space clone, while the big manufacturers in Chicago could not be bothered to adopt video technology at all.  Therefore, while Computer Space did not really perform poorly, it failed to interest the industry or the public in pursuing video games further.

Despite the mixed reception for Computer Space, Nolan Bushnell remained convinced that the future of popular entertainment lay with the video game.  He therefore resolved to move forward with new products, either in partnership with Nutting or through an agreement with one of the big Chicago firms.  Bushnell realized, however, that he would need a more accessible game for his next project.  While Bushnell mulled his next step, Nutting learned in May 1972 that yet another group was attempting to create a video game, so once again Bushnell decided to scout out the competition.  Unlike Pitts and Tuck, however, this group was not looking to create an arcade cabinet; they were preparing to bring video games into the home.