Inside Atari’s rise and fall

By the first few months of 1982, it had become more common to see electronics stores, toy stores, and discount variety stops selling 2600 games. This was before Electronics Boutique, Software Etc., and later, GameStop. Mostly you bought games at stores that sold other electronic products, like Sears or Consumer Distributors. Toys ’R’ Us was a big seller of 2600 games. To buy one, you had to get a piece of paper from the Atari aisle, bring it to the cashier, pay for it, and then wait at a pickup window behind the cash register lanes.

Everyone had a favorite store in their childhood; here’s a story about one of mine. A popular “destination” in south Brooklyn is Kings Plaza, a giant (for Brooklyn) two-story indoor mall with about 100 stores. My mother and grandmother were avid shoppers there. To get to the mall from our house, it was about a 10-minute car service ride. So once a week or thereabouts, we’d all go. The best part for me was when we went inside via its Avenue U entrance instead of on the Flatbush Avenue side. Don’t ask me what went into this decision each time; I assume it depended on the stores my mother wanted to go to. All I knew was the Avenue U side had this circular kiosk maybe 50 feet from the entrance. The name has faded from memory. I remember it was a kind of catch-all for things like magazines, camera film, and other random stuff.

But the most important things were the Atari cartridges. There used to be dozens of colorful Atari game boxes across the wall behind the counter. When we walked up to the cashier’s window, there was often a row of new Atari games across the top as well. Sometimes we left without a new cartridge, and sometimes I received one. But we always stopped and looked, and it was the highlight of my trip to the mall each time.

For whatever reason, I remember the guy behind the counter gave me a hard time one day. I bought one of Atari’s own cartridges—I no longer remember which, but I’m almost sure it was either Defender or Berzerk—that came with an issue of Atari Force, the DC comic book. I said I was excited to get it. The guy shot me a dirty look and said, “You’re buying a new Atari cartridge just for a comic book?” I was way too shy to argue with him, even though he was wrong and I wanted the cartridge. I don’t remember what my mother said, or if she even heard him. Being too shy to protest, I sheepishly took my game and we both walked away.

Mattel Stumbles, While Atari Face-Plants

Mattel began to run into trouble with its Intellivision once the company tried to branch out from sports games. Because Mattel couldn’t license properties from Atari, Nintendo, or Sega, it instead made its own translations of popular arcade games. Many looked better than what you’d find on the 2600, but ultimately played more slowly thanks to the Intellivision’s sluggish CPU. Perhaps the most successful was Astrosmash, a kind of hybrid of Asteroids and Space Invaders, where asteroids, space ships, and other objects fell from the sky and became progressively more difficult. Somewhat less successful were games like Space Armada (a Space Invaders knock off).

Mattel also added voice synthesis—something that was all the rage at the time—to the Intellivision courtesy of an add-on expansion module called Intellivoice. But only a few key games delivered voice capability: Space Spartans, Bomb Squad, B-17 Bomber (all three were launch titles), and later, Tron: Solar Sailer. The Intellivoice’s high cost, lack of a truly irresistible game, and overall poor sound quality meant this was one thing Atari didn’t have to find a way to answer with the 2600.

These events made it easier for Atari to further pull away from Mattel in the marketplace, and it did so—but not without a tremendous self-inflicted wound. A slew of new 2600 games arrived in the first part of 1982. Many important releases came in this period and those that followed, and we’ll get to those shortly. But there was one in particular that the entire story arc of the platform balanced on, and then fractured. It was more than a turning point; its repercussions reverberated throughout the then-new game industry, and to this day it sticks out as one of the key events that ultimately did in Atari.

Pac-Man (Atari, March 1982)

The single biggest image-shattering event for the 2600—and Atari itself—was the home release of its Pac-Man cartridge. I can still feel the crushing disappointment even now. So many of my friends and I looked forward to this release. We had talked about it all the time in elementary school. Pac-Man was simply the hottest thing around in the arcades, and we dreamed of playing it at home as much as we wanted. The two-year wait for Atari to release the 2600 cartridge seemed like forever. Retailers bought into the hype as well. Toy stores battled for inventory, JC Penney and Kmart bought in big along with Sears and advertised on TV, and even local drug stores started stocking the game. And yet, what we got…wasn’t right.

Just about everyone knows how Pac-Man is supposed to work, but just in case: You gobble up dots to gain points while avoiding four ghosts. Eat a power pellet, and you can turn the tables on the ghosts, chase them down, and eat them. Each time you do so, the “eyes” of the ghost fly back to the center of the screen and the ghost regenerates. Eat all the dots and power pellets on the screen, and you progress to the next one, which gets harder. Periodically, a piece of fruit appears at the center of the screen. You can eat it for bonus points, and the kind of fruit denotes the level you are on (cherry, strawberry, orange, and so on).

But that’s not the game Atari 2600 owners saw. After securing the rights to the game from Namco, Atari gave programmer Tod Frye just five weeks to complete the conversion. The company had learned from its earlier mistakes and promised Frye a royalty on every cartridge manufactured (not sold), which was an improvement. But this was another mistake. The royalty plus the rushed schedule meant Frye made money even if the game wasn’t up to snuff, and thus Frye had incentive to complete it regardless. Atari also required the game to fit into just 4KB like older 2600 cartridges, rather than the newer 8KB size that was becoming much more common by this point. That profit-driven limitation heavily influenced the way Frye approached the design of the game. To top it all off, Atari set itself up for a colossal failure by producing some 12 million cartridges, even though there were only 10 million 2600 consoles in circulation at the time. The company was confident that not only would every single existing 2600 owner buy the game, but that 2 million new customers would buy the console itself just for this cartridge.

We all know how it turned out. The instruction manual sets the tone for the differences from the arcade early on. The game is now set in “Mazeland.” You eat video wafers instead of dots. Every time you complete a board, you get an extra life. The manual says you also earn points from eating power pills, ghosts, and “vitamins.” Something is definitely amiss.

Pac-Man himself always looks to the right or left, even if he is going up or down. The video wafers are long and rectangular instead of small, square dots. Fruits don’t appear periodically at the center of the screen. Instead, you get the aforementioned vitamin, a clear placeholder for what would have been actual fruit had there been more time to get it right. The vitamin always looks the same and is always worth 100 points, instead of increasing as you clear levels. The rest of the scoring is much lower than it is in the arcade. Gobbling up all four ghosts totals just 300 points, and each video wafer is worth just 1 point.

The ghosts have tremendous amounts of flicker, and they all look and behave identically, instead of having different colors, distinct personalities, and eyes that pointed in the right direction. The flicker was there for a reason. Frye used it to draw the four ghosts in successive frames with a single sprite graphic register, and drew Pac-Man every frame using the other sprite graphic register. The 2600’s TIA chip synchronizes with an NTSC television picture 60 times per second, so you end up seeing a solid Pac-Man, maze, and video wafers (I can still barely type “video wafers” with a straight face), but the ghosts are each lit only one quarter of the time. A picture tube’s phosphorescent glow takes a little bit to fade, and your eye takes a little while to let go of a retained image as well, but the net result is that the flicker is still quite visible.

It gets worse. The janky, gritty sound effects are bizarre, and the theme song is reduced to four dissonant chords. (Oddly, these sounds resurfaced in some movies over the next 20 years and were a default “go-to” for sound designers working in post-production.) The horizontally stretched maze is nothing like the arcade, either, and the escape routes are at the top and bottom instead of the sides. The maze walls aren’t even blue; they’re orange, with a blue background, because it’s been reported Atari had a policy that only space games could have black backgrounds (!). At this point, don’t even ask about the lack of intermissions.

One of Frye’s own mistakes is that he made Pac-Man a two-player game. “Tod used a great deal of memory just tracking where each player had left off with eaten dots, power pellets, and score,” wrote Goldberg and Vendel in Atari Inc.: Business is Fun. Years later, when Frye looked at the code for the much more arcade-faithful 2600 Ms. Pac-Man, he saw the programmers were “able to use much more memory for graphics because it’s only a one player game.”

Interestingly, the game itself is still playable. Once you get past the initial huge letdown and just play it on its own merits, Pac-Man puts up a decent experience. It’s still “Pac-Man,” sort of, even if it delivers a rough approximation of the real thing as if it were seen and played through a straw. It’s worth playing today for nostalgia—after all, many of us played this cartridge to death anyway, because it was the one we had—and certainly as a historical curiosity for those who weren’t around for the golden age of arcades.

Many an Atari 2600 fan turned on the platform—and Atari in general—after the release of Pac-Man. Although the company still had plenty of excellent games and some of the best were yet to come, the betrayal was immediate and real and forever colored what much of the gaming public thought of Atari. The release of the Pac-Man cartridge didn’t curtail the 2600’s influence on the game industry by any means; we’ll visit many more innovations and developments as we go from here on out. But the 2600 conversion of Pac-Man gave the fledgling game industry its first template for how to botch a major title. It was the biggest release the Atari 2600 had and would ever see, and the company flubbed it about as hard as it could. It was New Coke before there was New Coke.

Grand Prix (Activision, March 1982)

The next few games we’ll discuss further illustrate the quality improvements upstart third-party developers delivered, in comparison with Atari, which had clearly become too comfortable in its lead position. First up is Activision’s Grand Prix, which in hindsight was a bit of an odd way to design a racer. It’s a side-scroller on rails that runs from left to right, and is what racing enthusiasts call a time trial. Although other computer-controlled cars are on the track, you’re racing against the clock, not them, and you don’t earn any points or increase your position on track for passing them.

Gameplay oddities aside, the oversized Formula One cars are wonderfully detailed, with brilliant use of color and animated spinning tires. The shaded color objects were the centerpiece of the design, as programmer David Crane said in a 1984 interview. “When I developed the capability for doing a large multicolored object on the [2600’s] screen, the capability fitted the pattern of the top view of a Grand Prix race car, so I made a racing game out of it.” Getting the opposing cars to appear and disappear properly as they entered and exited the screen also presented a problem, as the 2600’s lack of a frame buffer came into play again. The way TIA works, the 2600 would normally just make the car sprite begin to reappear on the opposite side of the screen as it disappeared from one side. To solve this issue, Crane ended up storing small “slices” of the car in ROM, and in real time the game drew whatever portions of the car were required to reach the edge of the screen. The effect is smooth and impossible to detect while playing.

The car accelerates over a fairly long period of time, and steps through simulated gears. Eventually it reaches a maximum speed and engine note, and you just travel along at that until you brake, crash into another car, or reach the finish line. As the manual points out, you don’t have to worry about cars coming back and passing you again, even if you crash. Once you pass them, they’re gone from the race.

The four game variations in Grand Prix are named after famous courses that resonate with racing fans (Watkins Glen, Brands Hatch, Le Mans, and Monaco). The courses bear no resemblance to the real ones; each game variation is simply longer and harder than the last. The tree-lined courses are just patterns of vehicles that appear on screen. Whenever you play a particular game variation, you see the same cars at the same times (unless you crash, which disrupts the pattern momentarily). The higher three variations include bridges, which you have to quickly steer onto or risk crashing. During gameplay, you get a warning in the form of a series of oil slicks that a bridge is coming up soon.

Although Atari’s Indy 500 set the bar early for home racing games on the 2600, Grand Prix demonstrated you could do one with a scrolling course and much better graphics. This game set the stage for more ambitious offerings the following year. And several decades later, people play games like this on their phones. We just call titles like Super Mario Run (a side-scroller) and Temple Run (3D-perspective) “endless runners,” as they have running characters instead of cars.

Activision soon became the template for other competing third-party 2600 developers. In 1981, Atari’s marketing vice president and a group of developers, including the programmers for Asteroids and Space Invaders on the console, started a company called Imagic. The company had a total of nine employees at the outset. Its name was derived from the words “imagination” and “magic”—two key components of every cartridge the company planned to release. Imagic games were known for their high quality, distinctive chrome boxes and labels, and trapezoidal cartridge edges. As with Activision, most Imagic games were solid efforts with an incredible amount of polish and were well worth purchasing.

Although Imagic technically became the second third-party developer for the 2600, the company’s first game didn’t arrive until March 1982. Another company, Games by Apollo, beat it to the punch by starting up in October 1981 and delivering its first (mediocre) game, Skeet Shoot, before the end of the year.

But when that first Imagic game did arrive, everyone noticed.

Demon Attack

At first glance, the visually striking Demon Attack looks kind of like a copy of the arcade game Phoenix, at least without the mothership screen (something it does gain in the Intellivision port). But the game comes into its own the more you play it. You’re stuck on the planet Krybor. Birdlike demons dart around and shoot clusters of lasers down toward you at the bottom of the screen. Your goal is to shoot the demons all out of the sky, wave after wave.

The playfield is mostly black, with a graded blue surface of the planet along the bottom of the screen. A pulsing, beating sound plays in the background. It increases in pitch the further you get into each level, only to pause and then start over with the next wave. The demons themselves are drawn beautifully, with finely detailed, colorful designs that are well animated and change from wave to wave. Every time you complete a wave, you get an extra life, to a maximum of six.

On later waves, the demons divide in two when shot, and are worth double the points. You can shoot the smaller demons, or just wait—eventually each one swoops down toward your laser cannon, back and forth until it reaches the bottom of the screen, at which point it disappears from the playfield. Shoot it while it’s diving and you get quadruple points. In the later stages, demons also shoot longer, faster clusters of lasers at your cannon.

The game is for one or two players, though there’s a cooperative mode that lets you take turns against the same waves of demons. There are also variations of the game that let you shoot faster lasers, as well as tracer shots that you can steer into the demons. After 84 waves, the game ends with a blank screen, though reportedly a later run of this cartridge eliminates that and lets you play indefinitely. If I were still nine years old, I could probably take a couple of days out of summer and see if this is true. I am no longer nine years old.

Demon Attack was one of Imagic’s first three games, along with Trick Shot and Star Voyager. Rob Fulop, originally of Atari fame and one of Imagic’s four founders, programmed Demon Attack. In November 1982, Atari sued Imagic because of Demon Attack’s similarity to Phoenix, the home rights of which Atari had purchased from Centuri. The case was eventually settled. Billboard magazine listed Demon Attack as one of the 10 best-selling games of 1982. It was also Imagic’s best-selling title, and Electronic Games magazine awarded it Game of the Year.

“The trick to the Demon Attack graphics was it was the first game to use my Scotch-taped/rubber-banded dedicated 2600 sprite animation authoring tool that ran on the Atari 800,” Fulop said in 1993. “The first time Michael Becker made a little test animation and we ran Bob Smith’s utility that successfully squirted his saved sprite data straight into the Demon Attack assembly code and it looked the same on the [2600] as it did on the 800 was HUGE! Before that day, all 2600 graphics ever seen were made using a #2 pencil, a sheet of graph paper, a lot of erasing, and a list of hex codes that were then retyped into the source assembly code, typically introducing a minimum of two pixel errors per eight-by-eight graphic stamp.”

Although you can draw a line from Space Invaders to just about any game like this, Demon Attack combines that with elements of Galaga and Phoenix, with a beautiful look and superb gameplay all its own.

Pitfall! (Activision, April 1982)

A watershed moment in video game history, David Crane’s Pitfall! was one of the best games released for the 2600. As Pitfall Harry, your goal is to race through the jungle and collect 32 treasures—money bags, silver bars, gold bars, and diamond rings, worth from 2,000 to 5,000 points each. Jump and grab vines, and you soar over lakes, quicksand, and alligators, complete with a Tarzan-style “yell.” You can stumble on a rolling log or fall into a hole, both of which just dock you some points. Each time you fall into quicksand or a tar pit, drown in a lake, burn in a fire, or get eaten by an alligator or scorpion, you lose a life. When that happens, you start the next one by dropping from the trees on the left side of the screen to keep playing.

Pushing the joystick left or right makes Pitfall Harry run. He picks up treasure automatically. Holding the stick in either direction while pressing the button makes him jump, either over an obstacle or onto a swinging vine (running into the vine without jumping also works). Push down while swinging to let go of the vine. You also can push up or down to climb ladders.

In an incredible feat of programming, the game contains 255 screens, with the 32 treasures scattered throughout them. The world loops around once you reach the last screen. Although Adventure pioneered the multiroom map on the 2600, Pitfall! was a considerably larger design. Crane fit the game into the same 4KB ROM as Adventure. But rather than storing all 255 screens as part of the ROM—which wouldn’t have fit—Crane’s solution was not to store the world in ROM at all. Instead, the world is generated by code, the same way each time. This is similar to games like Rogue, but even in that case, the game generates the world and then stores it during play. Pitfall! generates each screen via an algorithm, using a counter that increments in a pseudorandom sequence that is nonetheless consistent and can be run forwards or backwards. The 8 bits of each number in the counter sequence define the way the board looks. Bits 0 through 2 are object patterns, bits 3 through 5 are ground patterns, bits 6 and 7 cover the trees, and bit 7 also affects the underground pattern. This way, the world is generated the same way each and every single time. When you leave one screen, you always end up on the same next screen.

“The game was a jewel, a perfect world incised in a mere [4KB] of code,” Nick Montfort wrote in 2001 in Supercade: A Visual History of the Videogame Age, 1971-1984.

You get a total of three lives, and Crane points out in the manual that you need to use some of the underground passages (which skip three screens ahead instead of one) to complete the game on time. The inclusion of two on-screen levels—above ground and below ground, with ladders connecting them—makes the game an official platformer. And the game even gives you some say in where to go and what path you take to get there. Pitfall Harry is smoothly animated, and the vines deliver a genuine sensation of swinging even though the game is in 2D.

The game’s 20-minute timer, which approximates the 22-minute length of a standard half-hour television show, marked a milestone for console play. It was much longer than most arcade games and even cartridges like Adventure, which you could complete in a few minutes. The extra length allows for more in-depth play.

“Games in the early ’80s primarily used inanimate objects as main characters,” Crane said in a 2011 interview. “Rarely there would be a person, but even those weren’t fully articulated. I wanted to make a game character that could run, jump, climb, and otherwise interact with an on-screen world.” Crane spent the next couple of years tinkering with the idea before finally coming up with Pitfall!. “[After] only about 10 minutes I had a sketch of a man running on a path through the jungle collecting treasures. Then, after ‘only’ 1,000 hours of pixel drawing and programming, Pitfall Harry came to life.”

Crane said he had already gone beyond that 4KB ROM limit and back within it many times over hundreds of hours. Right before release, he was asked to add additional lives. “Now I had to add a display to show your number of lives remaining, and I had to bring in a new character when a new life was used.” The latter was easy, Crane said, because Pitfall Harry already knew how to fall and stop when he hit the ground. Crane just dropped him from behind the tree cover. “For the ‘Lives’ indicator I added vertical tally marks to the timer display. That probably only cost 24 bytes, and with another 20 hours of ‘scrunching’ the code I could fit that in.”

Pitfall! couldn’t have been timed more perfectly, as Raiders of the Lost Ark was the prior year’s biggest movie. The cartridge delivered the goods; it became the best-selling home video game of 1982 and it’s often credited as the game that kickstarted the platformer genre. Pitfall! held the top spot on Billboard’s chart for 64 consecutive weeks. “The fine graphic sense of the Activision design team greatly enriches the Pitfall! experience,” Electronic Games magazine wrote in January 1983, on bestowing the cartridge Best Adventure Videogame. “This is as richly complex a video game as you’ll find anywhere…Watching Harry swing across a quicksand pit on a slender vine while crocodiles snap their jaws frantically in a futile effort to tear off a little leg-of-hero snack is what video game adventures are all about.” Pitfall!’s influence is impossible to overstate. From Super Mario Bros. to Prince of Persia to Tomb Raider, it was the start of something huge.