Gamers of all ages are familiar with the work of Mark Cerny, thanks to an impressive record of designing not only games, but also the architecture of Sony’s PlayStation 4 and Vita. But, at the tender age of 17, Cerny joined the biggest stateside arcade producer of the time, Atari, and set about designing the work that would lay the base of his incredible career.
After the Berzerk-inspired Major Havoc, Cerny’s second game at Atari was Marble Madness. Working at Atari was programmer Bob Flanagan, also at the start of his career, and Flanagan was thrilled when he became involved with Cerny’s sophomore effort. “Mark’s design was inspirational, creative and unique,” he explains. “And I was very excited when he asked me to work with him, and then subsequently read his design.” Cerny’s design, clearly influenced by the Dutch artist Escher, took the themes of forced perspective, tessellation and geometry, and combined them with miniature golf and, in a key move for an arcade game, the video game racing genre.
The result was an exceptional and influential game that became a huge hit in the arcades. “My role was gameplay and base ROM programmer,” continues Flanagan. “Which meant most of the reusable code and self-test screens for the System 1 hardware.” Flanagan also contributed gameplay scripting, attract mode, sequencing, animation, collision detection and physics, the latter a key facet of Marble Madness’ game play. As he mentions above, the game utilises Atari’s System 1 in an upright cabinet, adapted to include a decoder for the trackball which the player uses to control their avatar.
In tune with its racing game origins, it’s apt that Marble Madness has no plot to speak of. The objective is simple and clear: controlling the little sphere, the player must race to the next stage, either solo or in competition with another human player, within a set time limit, and successful completion of this earns the player a shot at the next level. Each stage contains a set of increasingly-devious hills, yawning chasms and gullies, negotiation of which is essential, yet tricky given the marble’s momentum-infused movement. All of this is displayed in an isometric viewpoint, uncommon at the time, but helpful as far as Flanagan was concerned. “Actually the viewpoint made it way easier to do the math to translate the object positions into screen space, in order to make them look attached to the playfield. Also, the simplified math meant performance benefits too; perspective would have been more costly.”
Despite the minimalist design of Marble Madness, inevitable problems arose from Cerny’s template. “The moving wave at the bottom of the third playfield, and the moving patches at the end of the sixth were definite problems,” remembers Flanagan. “Mark spent a lot of time solving for functionality, graphics space, collision space and runtime performance. But we were able to get them in. And I’m sure there were other ideas from the original design, which didn’t make it, but probably more because of time than difficulty.”
Upon release in arcades, Marble Madness proved to be an instant hit. With the success of coin-ops notoriously difficult to judge during development, did Flanagan or others feel they could have a big seller on their hands? “By the time we showed it for the first time at AMOA [The Amusement and Music Operators Association], we were sure we had a great game,” he recalls. “Maybe not a hit, but definitely a great game.”
The key element, according to the coder, was the two player competitive racing mode. “There was definitely something there, and the first results of the focus group, and subsequent field tests, verified it.” The chance to competitively race against a friend, and deliberately bump them, sending them plummeting of the course into oblivion is a thrill that arcade fans clearly loved, akin to the joy of fairground dodgems.
One of the very few criticisms levelled at the game was its brevity; despite the steep learning curve, endeared by its control method and momentum, Marble Madness contains just six stages. “The number of levels was just based on the time it took Mark to create and tune them, and the production date for the game,” says Flanagan. “We just didn’t have time to add any more, and of course in the arcade, adding a few levels is not feasible as an add-on or update.”
Considering this complaint from fans, Flanagan’s sequel to the 1986 hit addresses the issue head-on. Marble Man: Marble Madness 2 (1991) contains 17 levels, multiple power-ups and adds a third player into the mix. “The original game has the basic that was known to be fun, with semi-realistic physics through patterns of living children’s toys such as the slinky, slime and whack-a-mole,” explains Flanagan. “But what it lacked was more, more of everything, so Marble Man was intended to add another player, more playfields and characters and also add new gameplay via the pinball levels.”
Unfortunately, the Marble Madness sequel saw little, if any, action within arcades, and Flanagan admits some responsibility for this. “Firstly, I made the design choice to target too young an audience with the Marble Man character, I should have kept it abstract like the original,” he laments. “And secondly, I made the mistake of switching the trackball to joysticks with an accelerator button, because by the time the game was to come out, more people had played the game that way in the home market, and didn’t even know what a trackball was.”
In fairness to its creator, Marble Man was not given a fair crack of the whip by Atari. No doubt obsessed by the success of beat-‘em-ups such as Street Fighter II, the arcade giant was soon focusing its energy elsewhere. Yet today, the original is still an inspired slice of game design, and great fun whether racing against a friend, or just trying to become one of those high rollers.