21
Hey Stella

On April 15, 1974, the Magnavox Corporation filed a lawsuit in the Federal District Court for the Northern District of Illinois, Eastern Division for patent infringement related to its Odyssey home video game system. Named in the suit were four companies that had released coin-operated ball-and-paddle video games – Atari, Bally, Allied Leisure, and Chicago Dynamic Industries – alongside one of the country’s largest coin-op distributors, Bally subsidiary Empire.¹ A second suit filed in August targeted Seeburg Industries, its coin-operated games subsidiary, Williams Electronics, and another large Chicago-based distributor, World Wide.² In 1975, a third lawsuit targeted Sears over the Atari-made Pong consoles it marketed for home use.³

As a veteran engineer at a major defense contractor, Odyssey creator Ralph Baer accumulated broad experience patenting new inventions and thoroughly documented the creation of the Odyssey from his first proposal in 1966. Baer subsequently filed for a series of patents between 1969 and 1974. The most important of these, issued in 1973 as US3728480, described in detail Baer’s plans for a “television gaming and training apparatus.”⁴ A second patent filed by co-worker Bill Rusch and issued in 1975 as USRE28507 described a game in which a machine-controlled dot collides with a player-controlled dot and changes its vector.⁵ These two patents together constituted a legal claim for having invented the concept of the video game generally as well as Pong-style ball-and-paddle games in particular.

¹ Complaint for Patent Infringement, Magnavox Co. v. Bally Manufacturing Corp., No. 74-1030 (N.D. Ill), April 15, 1974. This filing was preceded by a filing in New York by Bally subsidiary Midway attempting to have the patents invalidated, which was superseded by the April 15 case. Complaint for Declaratory Relief, Midway Mfg. Co. v. Magnavox Co., No. 74-1657 (S.D. NY), April 12, 1974.

² Complaint for Patent Infringement, Magnavox Co. v. Seeburg Industries, No. 74-2510 (N.D. Ill), August 30, 1974.

³ “Legal Battle Royal,” Weekly Television Digest with Consumer Electronics, December 1, 1975, 9.

⁴ Ralph Baer, “Television gaming and training apparatus,” U.S. Patent 3,728,480A filed March 22, 1971, and issued April 17, 1973.

⁵ William Rusch, “Television Gaming Apparatus,” U.S. Patent RE28507E, filed April 25, 1974, and issued August 5, 1975.

As Sanders Associates derived its primary video game income via licensing its technology to other companies, management was keen that anyone marketing a product similar to the Odyssey pay a licensing fee and/or a royalty for the privilege. When the Pong boom started in the arcades, however, no company did so. Under Sanders’ agreement with Magnavox, it was the television company’s duty to bring any lawsuits: hence the spate of patent infringement suits wending their way through the Federal Court system by 1975.

One-by-one, the defendants in these three patent lawsuits, which were ultimately consolidated into a single case, began to fall away. Allied Leisure argued improper venue to remove itself, which bought the Florida company a short reprieve before Magnavox sued it in its home state in 1976, resulting in a settlement in early 1978.⁶ Bally decided to settle in May 1976,⁷ while Chicago Dynamic Industries went bankrupt before the case could be decided, though it technically remained a party to the suit. At Atari, Nolan Bushnell initially desired to fight the suit, incensed that Magnavox claimed that what he considered his superior arcade technology owed anything to Ralph Baer’s more primitive home system, and launched a countersuit in California in 1975 seeking to invalidate the Magnavox patents that was consolidated into the Chicago case as well.⁸ Atari ultimately decided continuing the fight in court could be financially ruinous,⁹ so on June 8, 1976, the company settled.¹⁰

Seeburg and what remained of Chicago Dynamic continued to fight Magnavox to the bitter end and lost in court in early 1977. The presiding judge in the case, John Grady, declared Ralph Baer’s ‘480 patent the “pioneering patent” in video game technology and set a precedent that any company that created a game using a video signal in which a machine-controlled symbol collided with a player-controlled symbol and changed vector would run afoul of the ‘507 patent.¹¹ These judgments insured that Sanders would collect millions of dollars in licensing fees from companies all over the world before the patent finally expired in the early 1990s.

⁶ “Tandy & Allied Leisure,” Weekly Television Digest with Consumer Electronics, February 6, 1978, 9.

⁷ “RCA into Games,” Weekly Television Digest with Consumer Electronics, May 24, 1976, 12.

⁸ Complaint for Declaratory Judgement, Atari, Inc. v. Magnavox Co., No. 75-1442 (N.D. Cal.), July 11, 1975.

⁹ Kent, 2001, 46–47.

¹⁰ Non-Exclusive Cross License for Video Games, agreement between Atari and Magnavox, June 8, 1976.

¹¹ “Case: The Magnavox Co., et al. V. Chicago Dynamic Industries, et al.,” Patent Arcade, June 2010, http://patentarcade.com/2010/06/case-magnavox-co-et-al-v-chicago.html.

Atari’s settlement agreement with Magnavox called upon the video game company to pay $1.5 million in two installments to license Ralph Baer’s patents in perpetuity. The company also agreed to provide detailed information on any technology it developed for any product manufactured or available for sale by June 1, 1977.¹² The latter term may have led Atari to delay its entry into the emerging programmable console market, but at the June CES in 1977 the company unveiled the product that would propel it into the next generation of video games, the Video Computer System, or VCS.¹³

***

As Home Pong inched toward completion in summer 1975, Steve Mayer and Ron Milner at Cyan Engineering discussed how best to continue Atari’s move into the home market.¹⁴ They discerned that if Home Pong proved successful, the next logical product would be Tank, at the time one of the most popular coin-op games on the market. Doing that game would require a more sophisticated IC than the one used for Pong, and the thought of designing a new microchip every time Atari wanted to produce a new game was not palatable at all.¹⁵

Mayer saw another way forward. Inspired by Hewlett-Packard’s pocket calculator line, in which the primary difference between business and scientific variants was a different ROM chip inside a standard form factor, Mayer envisioned creating a base unit containing a microprocessor into which the company could plug in ROM chips containing individual games. This would provide Atari with a universal hardware standard and allow the company to design each new dedicated console game in software rather than through LSI circuits. After developing this basic idea, Mayer came to the same realization as Wallace Kirschner at Alpex that if they were putting individual games on interchangeable ROM chips, they might as well just release the base unit and let consumers switch out games themselves.¹⁶

¹² Agreement between Atari and Magnavox, 1976. Most companies paid royalties rather than a flat fee. Coleco’s license, which was probably typical, called for $100,000 against a 5.5% royalty on net sales of the first 100,000 units followed by a sliding scale down to 3% of net sales on units above 350,000. “Magnavox Game Suit,” Weekly Television Digest with Consumer Electronics, December 13, 1976, 11.

¹³ Al Alcorn has stated that Atari kept the system a secret until the June CES so it would not have to share info on it with Magnavox, as the show was held right after the deal expired. Other former Atari executives are not so sure. Furthermore, Magnavox’s patent attorney, Thomas Briody, testified that Atari did share information on the system pursuant to the agreement. Alcorn, oral history 2008, and Thomas Briody, testimony, Magnavox Company v. Activision, Inc., No. 82-5270 (N.D. Cal), 1985, 99-100. It is possible Atari disclosed general information on the VCS while it was under development to ensure it did not violate the settlement, but did not have to disclose specific technical details since it was not being offered for sale to distributors and retailers until after June 1.

¹⁴ Marty Goldberg, “Inside the Atari 2600,” Retro Gamer, no. 103, 2012, 27.

¹⁵ Hans Reutter, “PRGE 2017 – Ron Milner – Portland Retro Gaming Expo,” filmed October 21, 2017, YouTube video, posted November 2, 2017, https://www.youtube.com/watch?v=CbaAgAAY77U.

¹⁶ Mayer, 2018.

As Mayer and Milner began sketching out the design of their new system, with Milner focusing on hardware and Mayer on software, they realized keeping the retail price around their target of $200 would limit the budget for the microprocessor at its heart to perhaps no more than $30.¹⁷ This put the whole project in jeopardy, as in their survey of the emerging microprocessor industry they were unable to find any that sold for less than $100. The solution arrived in the form of a letter mailed to Cyan advertising a new processor debuting publicly at the upcoming Western Electronics Show and Convention (Wescon) created by a team led by Chuck Peddle.¹⁸

A graduate of the University of Maine, Chuck Peddle entered the computer industry in 1961 at General Electric, where he worked on time-sharing and electronic cash register designs for nearly a decade until GE exited the computer industry in 1970. After briefly working at two startups he co-founded that focused on cash registers and word processing, respectively, Peddle took a job at Motorola in 1973 to help finish the company’s first microprocessor, the 8-bit 6800. Afterwards, Peddle was tasked with evangelizing the 6800 to potential customers. While the new chip generated enthusiasm, there was disappointment in some circles due to the chip’s $300 price tag. While this was a reasonable price for companies looking to replace the processing power of an entire computer, some engineers hoped to use the chip as a microcontroller and were disappointed that doing so would not be cost effective. Peddle realized a potential market existed for cheaper microprocessors for use in computer peripherals, appliances, and similar devices, but when he proposed that Motorola create a low-cost version of the 6800 to meet this demand, he was rebuffed. Peddle, therefore, decided to pursue this chip himself.¹⁹

In order to proceed on his own, Peddle required an existing semiconductor company that could develop and manufacture his design. After several companies turned him down, Peddle reconnected with a former co-worker at GE named John Paivinen, co-founder and president of a firm called MOS Technology. Established in Valley Forge, PA, in 1969 as a subsidiary of Allen-Bradley to bring the industrial manufacturing company into the semiconductor industry, MOS had become the largest independent supplier of chips to the calculator industry by the mid-1970s. Paivinen was happy to take on Peddle’s new project, so Peddle and seven members of his design team left Motorola in August 1974 and moved east. After just under a year of work, they completed their low-cost microprocessor in June 1975 and laid it out in two versions,²⁰ the 6501, which had the same pin layout as the 6800 for socket compatibility and cost $20, and the 6502, which featured a different layout and cost $25. Knowing the industry would likely greet the announcement of such cheap CPUs with skepticism, Peddle decided to debut them at the Wescon show in September 1975.

¹⁷ Ibid.

¹⁸ Goldberg, 2012, 29.

¹⁹ Brian Bagnall, Commodore: A Company on the Edge (Winnipeg: Variant Press, 2010), 2–9.

²⁰ Ibid, 10–19.

Shortly before the show, Peddle began running ads in industry trades announcing that the new 6502 processor would be available for purchase on the floor of the Wescon show for $25.²¹ When the Wescon organizers learned of the ad, they responded by banning sales on the show floor to forestall a flea market atmosphere. Unperturbed, Peddle set up shop in a hotel room and stationed a small booth on the floor of the show to point people to where they could buy the chip.²² Mayer and Milner were two of his many customers that day, and after evaluating the chip they called Peddle the next day to express interest in using it in their game system. After a brief negotiation, they concluded a deal with Peddle to use the MOS Technology 6507 microprocessor, a slightly more limited version of the 6502 that sold at a far cheaper price, and its supporting chips at a cost of $12 a set.²³ With a CPU now in place, Mayer and Milner began assembling a prototype of their new system.

As Mayer and Milner turned to their video game project in earnest, they were guided by one simple principle: include as much functionality as possible in the hardware, but never at the expense of a competitive price point. This philosophy led them to move as many capabilities out of the hardware and into software as they could. The 6507 was synchronized to the television scan rate and created the display only one or two lines at a time in lieu of a frame buffer; the less important background images were rendered at a much lower resolution than moving objects, and vertical synchronization was moved out of hardware entirely, all in an effort to cut down on video memory.²⁴ Indeed, with DRAM still incredibly expensive in 1975, the duo initially decided to include only 64 bytes of system memory, though after realizing the cost savings from using a 6507 rather than a 6502 they doubled the memory to 128 bytes. This was still a paltry amount of memory, but they could simply not afford to include more.²⁵

²¹ The 6501 was never released because a lawsuit by Motorola forced MOS to remove it from the market.

²² Ibid, 32–35.

²³ Ibid, 35–36.

²⁴ “Design Case History: The Atari Video Computer System,” IEEE Spectrum, March 1983, 45.

²⁵ Mayer, 2018.

Mayer and Milner assumed a limited lifecycle for their system and only built in the functionality required to play all of Atari’s major arcade hits to that point: Pong, Pong Doubles, Quadrapong, Gran Trak 10, Tank, and Jet Fighter.²⁶ Therefore, it could only generate five sprites, two 8-bit sprites to represent players and three 1-bit sprites to represent two missiles and a ball.²⁷ These sprites could be one of 128 colors and placed anywhere on what was essentially a 160×192 playfield.²⁸ They could be superimposed against a background image up to 40 bits wide that could alternately be mirrored to show the same 20-bit wide image on each half of the screen. These graphical tricks insured that simple games like Pong or Tank could be housed within 2K of ROM, so the duo designed the system to address up to 4K of ROM memory. Addressing more ROM was deemed too expensive to remain within the desired price point for the system and its attendant cartridges.

By the beginning of December 1975, Mayer and Milner had completed a rough prototype of the system and created a simple adaptation of Tank to play on it.²⁹ That month, they hired a third team member named Joe Decuir. A Berkeley grad with a master’s degree in biomedical engineering, Decuir had been interested in computers since high school and purchased his own 6502 processor at the same Wescon as Mayer and Milner. Decuir helped debug the Tank game and assisted in the development of a gate-level prototype of the hardware incorporating a breadboard full of integrated circuits to mirror as closely as possible the custom graphics and sound chip Atari would have to design for the final system.³⁰ He also provided the name for the prototype of the chip, which the team dubbed Stella after the brand of bicycle Decuir rode to work every day. Before long, the codename Stella was applied to the entire console project.³¹

²⁶ Ibid.

²⁷ The term “sprite” as referring to a predefined block of pixels had not yet been coined in 1975. Atari called its objects “player-missile graphics.”

²⁸ Due to the way the VCS drew the screen without a bitmap, there were not technically any pixels being turned on and off, but there were essentially 160×192 different spots on the screen where a sprite could be placed.

²⁹ Goldberg, 2012, 30.

³⁰ Scott Stilphen, “DP Interviews Joe Decuir,” Digital Press, 2005, http://www.digitpress.com/library/interviews/interview_joe_decuir.html.

³¹ Goldberg, 2012, 31–32.

With Stella’s basic features established by February 1976, Joe Decuir was pulled out of Cyan and set up shop at a new Atari facility on Division Street in Sunnyvale separate from the rest of the company’s microelectronics department to turn the prototype into a viable product away from prying eyes.³² The most important task was completing the custom graphics and sound chip needed for the system, which was beyond the expertise of anyone working at Cyan or Atari. It fell to R&D VP Al Alcorn and microelectronics director Bob Brown to find someone who could complete the chip, but it was their Home Pong partner Harold Lee who presented the solution, telling them that only Jay Miner of Synertek could execute the design.³³

Born in May 1932 in Prescott, Arizona, Jay Glenn Miner graduated from high school in Southern California and then joined the U.S. Coast Guard. He received his first formal electronics training through a six-month course at the Coast Guard Electronics Technician School in Groton, Connecticut, and then joined the North Atlantic Weather Patrol to service radar stations and radio installations on remote islands. After three years of this work, he enrolled at the University of California at Berkeley with a major in electrical engineering and graduated in 1958 with a focus on the design of generators and servomotors. In 1964, Miner was hired by Fairchild spin-off General Micro Electronics and helped design the first MOS calculator chip. In 1973, he left the company and co-founded Synertek, where he served as the principal chip designer.³⁴ By that time, he was already considered one of the best designers in the field.

Synertek served as the primary source of LSI chips for Atari’s dedicated console systems,³⁵ so setting up a meeting proved simple. Convincing the company to give up its best chip designer proved harder. Alcorn, however, had an ace up his sleeve. Atari president Joe Keenan had insisted that the only way he would allow his engineers to use the 6507 in their video game system rather than a Motorola 6800 was if they set up a second source for the chips, a standard practice in the electronics industry to reduce dependence on a single company in the volatile semiconductor market. Atari worked out a deal for Synertek to serve as that second source. Alcorn reminded Synertek that if Atari’s video game system proved successful, it would be receiving orders for vast quantities of chips over the next few years, and he got his man.³⁶ Miner worked with Joe Decuir and a mathematician Alcorn met after giving a talk named Larry Wagner to reduce Stella’s breadboard full of parts to a chip labeled the Television Interface Adapter (TIA). Meanwhile, Atari began drawing up plans for putting Stella into production via its new consumer electronics division.

³² Golderg and Vendel, 2012, chap. 5.

³³ Kent, 2001, 99.

³⁴ Brian Bagnall, Commodore: The Amiga Years (Winnipeg: Variant Press, 2017), 3.

³⁵ Although, the original Home Pong LSI was developed by AMI, the chip company got cold feet about entering mass production and only produced a limited run. Synertek took over production and became Atari's main chip supplier.

³⁶ Alcorn, oral history, 2008.

***

With the success of Home Pong in 1975, Nolan Bushnell and Joe Keenan decided at the beginning of 1976 to split Atari into two divisions. Marketing VP Gene Lipkin became the general manager of the new coin-operated games division, while a former employee of National Semiconductor’s Novus division named Sheldon Ritter took control of a new consumer electronics division.³⁷ These entities would exercise complete control over product development, sales, marketing, and manufacturing in their respective spheres. At Atari Consumer, Malcolm Kuhn and Michael Shea were hired as directors of sales and marketing, respectively. To fill out his product development team, Ritter hired a group of former Novus employees led by his former co-worker John Ellis as VP of Engineering. Former Novus general manager Gene Landrum was also retained as a consultant and created a market analysis and product planning strategy document to help define the Stella system – playing a similar role as he had for Fairchild on their Channel F console – that he delivered in May 1976.³⁸

Based on Landrum’s report, Ellis began drawing up the final production specs for Stella, while remaining mindful of the potential costs. Landrum recommended the system ship with several built-in games like the Channel F, but the required ROM chips were deemed too expensive. The report also mirrored the Channel F in calling for sound generated by speakers located within the console itself, but engineers Niles Strohl and Wade Tuma, who designed the RF modulator for the system, realized they could easily run the sound through the modulator to the television speakers instead. Making this modification not only provided better sound, but also reduced the audio part cost from $10 to a mere 25 cents.³⁹ Solving the cartridge insertion problem proved trickier until Atari hired an industrial engineer named Doug Hardy who had worked with Ron Smith at Fairchild on the cartridge mechanism for the Channel F. Although Hardy used a different system to avoid running afoul of Fairchild’s patents, his solution was informed by the work done by the Fairchild team.⁴⁰

³⁷ “Trade Personals,” Weekly Television Digest with Consumer Electronics, February 2, 1976, 12.

³⁸ Goldberg and Vendel, 2012, chap. 5.

³⁹ Ibid.

⁴⁰ Edwards, 2015.

The case for the console was created by industrial designer Fred Thompson, who joined Atari in 1975 to refine the case design for Home Pong. Thompson referenced the aesthetics of component stereo equipment and developed a form factor that could be cradled in a lap because he figured the insertion and removal of cartridges, the need to set several toggle switches to select game modes, and the short length of the cords on the controllers would require the players to stay close to the console. The primary controller, a joystick and button set into a square base, was designed by Kevin McKinsey for a Tank console that Atari developed as a backup in case the VCS did not work out.⁴¹ The system also shipped with a pair of paddle controllers that could be plugged into a single controller port.

In late 1976, microelectronics director Bob Brown began assembling a software team to create games for the console. Larry Wagner was tapped to lead the new group, and he also put the finishing touches on the Tank game that Joe Decuir had implemented as a test program during hardware development, which now went by the name Combat. Decuir, meanwhile, implemented Pong on the system under the name Video Olympics.⁴² Because both games were relatively simple, Wagner and Decuir enhanced them via the addition of multiple game modes to increase replay value. Combat included parameters like bouncing bullets, invisible tanks, and a bank shot variant in which a tank could only be destroyed by a bullet that bounced off a wall first. It also included several variations on the coin-op game Jet Fighter. Video Olympics shipped not only with Pong, Super Pong, and Pong Doubles ports, but also with Quadrapong, soccer, hockey, volleyball, and basketball variants.

In August 1976, Brown and Wagner hired their first dedicated Stella programmer, Larry Kaplan. Kaplan had initially matriculated to UC Berkeley in spring 1968 with plans to major in biochemistry, but because he arrived in the middle of the school year, he was unable to start his core courses until the next fall. His academic advisor was a member of the faculty of the brand new computer science department, so he encouraged Kaplan to take a programming class as an elective that spring as part of a larger strategy to boost enrollment in computer courses so that the university would not shut down the nascent department. That summer, Kaplan took a work-study job with the U.S. Forestry Service as a statistician to help pay for his schooling and worked with both calculators and computers. After taking several more programming courses over the next two years, he dropped out of school in 1970. After getting married and being drafted, he returned to Berkeley as a computer science major to secure a deferment and graduated in 1974.⁴³

⁴¹ Tim Lapetino, Art of Atari (Runnenmede, NJ: Dynamite Entertainment, 2016), 296–300.

⁴² Curt Vendel, “Interview with Joe Decuir,” Atari Museum, accessed June 29, 2019, http://www.atarimuseum.com/articles/joedecuir.html.

⁴³ Larry Kaplan, interview with the author, September 3, 2018.

After graduation, Kaplan took a job with Control Systems Industries to work on a computerized power grid for the Missouri River for the Bureau of Reclamation in Missoula, Montana.⁴⁴ Already familiar with microprocessors because he had taken a course programming the Intel 4004 after Berkeley alum Ted Hoff donated several chips to the university,⁴⁵ Kaplan became interested in the emerging home computer market and bought an Altair 8800 to fool around with. In 1976, Kaplan saw an Atari ad in the San Jose Mercury News seeking programmers, and he beat out roughly 100 applicants for the job of Atari’s first consumer programmer largely on the strength of his interest in, and familiarity with, home computers.⁴⁶

For his first game, Kaplan chose to convert the 1975 Atari coin-op game Anti-Aircraft, in which two players shoot at airplanes flying overhead to score the most points. Creating a target-rich environment proved challenging because Stella was designed to display only five objects on the screen at one time. Fortunately, Jay Miner included a register called “H-move” within the TIA architecture that allowed Stella to calculate the horizontal motion of objects between frames. Kaplan discovered that since the system had to be aware of the horizontal position of objects on each line of the display during the rendering of a frame due to the way vertical draw was synchronized to the CRT, the H-move could be executed more than once per frame to update the position of objects line by line in horizontal bands. The upside of this quirk was that sprites could be reused as the screen was being drawn, and multiple objects could thus be rendered using the same character so long as they appeared in different horizontal positions.⁴⁷ Kaplan took full advantage of this capability to not only faithfully recreate Anti-Aircraft, but also to include naval variants that mimicked hit arcade games like Sea Wolf and shooting gallery variants that allowed the players to move their gun batteries along the bottom of the screen while replacing ships and planes as targets with clowns and animals. In light of these many variants, Atari called the game Air-Sea Battle.

⁴⁴ Scott Stilphen, “DP Interviews Larry Kaplan,” Digital Press, 2006.

⁴⁵ Kaplan, 2018.

⁴⁶ Stilphen, 2006.

⁴⁷ “Design Case History,” 1983, 46.

In January 1977, Atari hired a second programmer named Bob Whitehead. A native of the Santa Clara Valley, Whitehead majored in computer science and mathematics at San Jose State University and then took a job with defense contractor GTE Sylvania, where his boss was future Atari microelectronics director Bob Brown. Brown remembered Whitehead as a quick learner with microprocessor experience and personally called to offer him a job at Atari.⁴⁸ Whitehead adapted the first-person space shooter Starship 1 as his first project under the name Star Ship. Like the other early programmers, Whitehead incorporated several variants into the game, including a take on the early Atari arcade game Space Race and a primitive version of the Lunar Lander computer game.

In February 1977, Atari hired the final two programmers of the original VCS group, Alan Miller and Gary Palmer. In addition to creating games, both men played a crucial role in improving the development environment for creating new games. At the time, the programmers were required to input all their code using a teletype that could only display one line at a time connected to a time-sharing computer in Oakland. They debugged the software using a simple modified KIM-1 board adapted by an electrical engineer named Ed Riddle. Miller had experience working with PDP-11 minicomputers, and he successfully lobbied jointly with Kaplan to convince Atari to acquire PDP development systems. Palmer, meanwhile, designed proper debugger stations to replace the KIM boards. Palmer subsequently developed an educational cartridge called Basic Math, while Miller created a clone of the arcade game Blockade under the name Surround.⁴⁹

Three more cartridges rounded out the Stella launch lineup. Original debugger designer Ed Riddle decided to try his hand at programming and developed a Gran Trak-style driving game called Indy 500 based on a design started by Wagner.⁵⁰ With its overhead view and twisty curves, the game was unsuited to either the joystick or paddle controllers slated to ship with the system, so Atari packaged a rotary controller with the game that looked similar to the system’s paddle controllers but could be rotated a full 360°.⁵¹ Kaplan and Whitehead, meanwhile, each completed a second game, a Wheels clone called Street Racer and a Blackjack game, respectively.

⁴⁸ Bob Whitehead, interview with the author, February 2, 2016.

⁴⁹ Al Backiel, “DP Interviews Alan Miller,” Digital Press, 2003.

⁵⁰ Scott Stilphen, “Ed Riddle Interview,” Atari Compendium, 2017.

⁵¹ Indy 500 was the only game to ever use this controller.

***

As work progressed on Stella in 1976, Nolan Bushnell and Joe Keenan realized they did not possess the financial resources to complete the hardware and software and build inventory for the 1977 holiday season. While the company was now profitable, most of its reserves were tied up in inventory for the 1976 Christmas season, leaving little cash on hand. Bushnell hoped to take the company public to raise additional funds, but he became spooked when a soft market led his underwriter to believe they would not get nearly as large a return as hoped in an initial public offering.⁵² Don Valentine, who in addition to being the principal venture investor in Atari also sat on the company board, pushed Bushnell to sell the company instead, promising they could achieve a similar payout to what they could expect on the open market. The board drew up a list of 40 or so companies active across a wide range of technological and entertainment fields and split them up among its members to put out feelers for a sale. For Valentine, there was only one choice that made sense based on his prior history of investments with the company: Warner Communications.⁵³

Warner mastermind Steven Jay Ross was born in 1927 in Brooklyn and grew up in impoverished circumstances after his father was ruined in the Great Depression. In 1953, he married Carol Rosenthal, whose father owned a large funeral home called Riverside Memorial Chapel. Two years later, Ross went to work for Mr. Rosenthal as a funeral director. Ever the businessman, when Ross realized that the funeral home limousines were idle in the evening, he received permission from his father-in-law to start a rental business on the side.⁵⁴ In 1959, Mr. Rosenthal started a joint venture called Abbey Rent a Car that leased parking garage space from a company called the Kinney Parking Company. Ross helped nurture this business, which grew so successful that it began supplanting the garage business, making its owners unhappy. Ross provided the solution: merge Riverside and Kinney into a single firm that he would take public. The organizations merged as the Kinney Service Company in 1960, which went public in 1962 with Ross as president.⁵⁵

The 1960s were an era of conglomerating, and the skillful dealmaker Ross began expanding his company. In 1966, he capped off his first series of acquisitions with a $25 million purchase of National Cleaning Contractors, instantly doubling the size of the firm that was now called Kinney National Service. The next year, his friend and former employee Kenneth Rosen introduced him to Ted Ashley, owner of the Ashley Famous Talent Agency. By now, Ross was looking to expand out of Kinney’s core businesses and had identified the leisure and entertainment industries as potential growth fields. Ashley initially had no desire to associate with someone in the funeral, parking, and janitorial business, but Ross won him over in a half-hour meeting that stretched to three hours and purchased the company. After consummating the deal, Ross was promoted to co-CEO of Kinney National Service.⁵⁶

⁵² Bushnell, 2015.

⁵³ Keenan, 2018.

⁵⁴ Connie Bruck, Master of the Game: How Steve Ross Rode the Light Fantastic from Undertaker to Creator of the Largest Media Conglomerate in the World (New York: Simon & Schuster, 1994), kindle, chap. 1.

⁵⁵ Ibid.

⁵⁶ Ibid, chap. 2.

The purchase of the Ashley Famous Talent Agency focused Ross’s attention on entertainment, and he made overtures to acquire Warner-Seven Arts, the parent company of Warner Records and the Warner Brothers movie studio. Initially rebuffed because Warner already had a deal in place with another company, Ross acquired the company in 1969 for $64 million after the original deal failed to receive government approval due to anti-trust problems.⁵⁷ With the era of the conglomerate nearing its end, and the Kinney parking business facing a price fixing scandal, Ross spun off all the non-entertainment assets of the business into a new firm called National Kinney Corp in 1972. Ross assumed the roles of chairman, CEO, and president at Kinney National, which he renamed Warner Communications.⁵⁸

While putting together the Warner-Seven Arts deal, Ross relied heavily on entertainment analyst Emmanuel “Manny” Gerard. After graduating from Brown University and attending Harvard Business School, Gerard cut his teeth as an analyst with Wood Struthers & Co. in the late 1950s. Although he was rotated between a number of industries, he gravitated toward entertainment and founded his own boutique firm focused on that market with Alan Roth in 1961. After his consulting work on the Warner deal, Ross asked him to join Warner in 1974, where his mission was to seek out acquisitions in the entertainment business.⁵⁹ In 1976, Gerard became one of four members of a new “Office of the President” established by Ross to run Warner Communications and serve as the interface between the parent company and its subsidiaries, which Ross encouraged to run almost entirely independent of central oversight.⁶⁰

⁵⁷ Ibid.

⁵⁸ Ibid, chap. 3.

⁵⁹ “The Return of Manny Gerard,” Barron’s National Business and Financial Weekly, May 14, 1990.

⁶⁰ The other members were Ken Rosen, who helped facilitate the Ashley Famous sale, David Horowitz, a lawyer who had been the general counsel for Columbia Pictures, and Jay Emmett, who had run a company called the Licensing Corporation of America acquired by Kinney and later became a close friend and confidant to Ross. Bruck, 1994, chap. 3.

In early 1976, Gerard received a telephone call from Gordon Crawford of Capital Research on behalf of Don Valentine.⁶¹ Crawford simply asked if Gerard would be interested in “a technology-based entertainment company, growing rapidly.” Gerard replied in the affirmative. After flying out to see the Atari facilities in person, Gerard penned an analysis for Warner in which he boldly proclaimed that he had “seen the future, and its name is Stella.” Convinced the console would become a massive hit, he encouraged Warner to buy the company.⁶²

Negotiations began in earnest in spring 1976 but were complicated by several factors. Atari CFO Bill White grew severely ill during negotiations and was forced to limit his involvement, while an attempt by National Semiconductor to sell its moribund Novus consumer division to Atari was misread by Warner as a move by National to buy Atari. In July, an article appeared in a local newspaper describing Nolan Bushnell’s recreational activities that included a picture of him in a hot tub with a lady friend. Nolan’s ex-wife, Paula, whom he had divorced in 1974, was so enraged by the article that she threatened to rescind their divorce agreement. Doing so would force Atari to restore her 25% stake in the company, so Warner worked with Nolan and Paula to forge a new agreement in which she agreed to surrender her ownership stake or any future profits from the company. After surmounting these obstacles, Warner purchased Atari in October 1976 for $28 million. Nolan Bushnell and Joe Keenan remained in their positions as chairman and president, respectively, and were promised near-complete autonomy, though they now reported to Manny Gerard. Warner began investing money into Atari so it could bring Stella to market.⁶³

***

Atari began shipping Stella to retailers as the Video Computer System (VCS) in July 1977 at a suggested retail price of $190.⁶⁴ Combat came packaged with the system, while the other eight launch titles were available for $20.00 each, save for Indy 500, which cost $40.00 due to the included rotary controllers. Sears also marketed the console under its own Telegames brand as the Video Arcade and packaged it with Air-Sea Battle under the name Target Fun.⁶⁵

⁶¹ Capital Research was the company that funded Valentine’s Sequoia Capital at startup.

⁶² Manny Gerard, interview with the author, September 5, 2017.

⁶³ Goldberg and Vendel, 2012, chap. 5.

⁶⁴ “Games – Déjà vu Sets In,” Weekly Television Digest with Consumer Electronics, October 17, 1977, 10.

⁶⁵ Over the first two years of the VCS life cycle, Sears marketed all Atari games under alternate names.

The console sold 340,000 units,⁶⁶ but the launch was beset with problems that limited overall profitability. Manufacturing difficulties led to defects and delays, blown delivery schedules, and an unacceptably high return rate.⁶⁷ Storms on the East Coast delayed overland shipments even further,⁶⁸ while the chip shortages that had plagued the console industry since the start of the Pong boom in 1976 further hampered Atari’s ability to manufacture cartridges in sufficient quantities to meet demand.⁶⁹ Atari attempted to compensate by continuing to run the assembly lines practically right up to Christmas, but retailers were not interested in ordering product past the beginning of December. Atari got stuck with unsold inventory sitting in warehouses that would not be viable until the next Christmas season, as high ticket items like video game consoles remained a seasonal product.⁷⁰ Despite a 35% increase in revenue and the highest sales in the industry by dollar volume, the consumer division lost money.⁷¹

Overall, the video game market proved disappointing in 1977. While dedicated console sales increased to 5.5 million units worth $230 million,⁷² they fell below estimates of 7–10 million.⁷³ Dollar volume also suffered due to deep discounting during the holidays intended to clear out stock so as not to repeat the situation at the beginning of the year when retailers proved unable to sell remaining inventory after the previous Christmas.⁷⁴ Programmable sales topped out at around 450,000 units,⁷⁵ which also lagged analyst expectations. Both Fairchild and Atari attributed the shortfall to chip shortages. Worse, prices eroded rapidly during the holidays because retailers were terrified that they would not be able to sell excess stock come January 1 because of their experience with dedicated consoles the year before.⁷⁶ Atari led the way in the programmable market with its 340,000 VCS systems, while Fairchild, with its conservative approach to the market, sold around 100,000 units.⁷⁷ RCA did so poorly with its antiquated Studio II – even after dropping the retail price to $90 in early December⁷⁸ – that it shut down production entirely in February 1978 and sold its remaining stock to Radio Shack. The company had been working on a full-color Studio III and had even begun design work on a Studio IV,⁷⁹ but these projects were cancelled as RCA left the video game business behind.

⁶⁶ Warner Communications, Annual Report, 1977.

⁶⁷ Tristan Donovan, “The Replay Interviews: Ray Kassar,” Gamasutra, April 29, 2011, https://www.gamasutra.com/view/feature/134733/the_replay_interviews_ray_kassar.php.

⁶⁸ Bushnell, 2009.

⁶⁹ Warner Communications, 1977.

⁷⁰ Malcolm Kuhn, interview with the author, July 9, 2013.

⁷¹ Warner Communications, 1977.

⁷² Frost & Sullivan, 1983, 24.

⁷³ “Games and Computers,” The Emporia Gazette, December 21, 1977.

⁷⁴ “How the Two Big TV Toys are Faring,” Broadcasting, 1978, 80.

⁷⁵ Frost & Sullivan, 1979.

⁷⁶ “Games – Retail Confusion,” Weekly Television Digest with Consumer Electronics, December 12, 1977, 10.

⁷⁷ Frost & Sullivan, 1979.

⁷⁸ “Games – Retail Confusion,” 10.

⁷⁹ G.B. Herzog, “Fred Systems,” internal progress report on Studio III and Studio IV, c. 1977.

In the dedicated console market, cheaper models did well, while advanced systems in the $70–$100 price range faltered.⁸⁰ Coleco remained the market leader with sales of 1.75 million units, but an East Coast dockworkers strike left the company unable to ship its entire inventory in time for the Christmas buying season, and it barely eked out a $1.7 million profit for the year. Forced to liquidate leftover stock in early 1978, the company took a $22.3 million loss and exited the video game industry⁸¹ In the process, it scrapped plans to release a new cartridge-based system called the Telstar Game Computer.⁸² APF took second place by selling around 1 million units of its latest TV Fun models. Atari came in third with around 510,000 units sold,⁸³ and Magnavox rounded out the major companies by moving 400,000 units of its 1977 models.⁸⁴

As the dedicated console video market fell apart on the low end of the video game market, the microcomputer industry looked poised to take a giant step forward and settle in as a new alternative to video games on the high end of the market. At its inception, the industry targeted hobbyists and electronics enthusiasts who desired to build their own computers. Now, in 1977 several companies were preparing to release the first fully assembled microcomputers.

⁸⁰ “Changing Games Market,” Weekly Television Digest with Consumer Electronics, January 16, 1978, 9.

⁸¹ Michael Rosenbaum, “Gamesmanship: Coleco Industries’ Electronic Toys are Doing Just Fine,” Barron’s National Business and Financial Weekly, March 16, 1981.

⁸² Despite the name, the Telstar Game Computer was not programmable. Like the Telstar Arcade, there was no CPU and each cartridge contained its own LSI.

⁸³ Frost & Sullivan, 1976, 8.

⁸⁴ Baer, 2005, 133.