Note: Opinion article, inspired by a discussion I had on the Gameological Society. Thanks to Cart Boy and NantenJex for edits.
It’s easy to treat Super Mario 64, the iconic 1996 platformer that introduced Nintendo’s mascot to the third dimension, as a canonized relic. It was not the first game to take place in a three dimensional space or to render objects through polygons, but it drove those conceits to extents and with a level of quality that seemed unreal at the time. It introduced the ability for the player or computer to operate the camera dynamically instead of from a fixed perspective. The Joystick allowed a precise degree of analog control, instead of relying on the standard Directional-Pad to interpret a 360° range of movement. Players searched large, contiguous worlds for “Power Stars” to unlock later levels, a nonlinear form of progression which became a standard for open world design. Whether directly or indirectly, it introduced and codified so many 3D action game tropes we take for granted it almost risks feeling generic today. But that influence came from it being radical, inventive, and experimental in ways we rarely associate with Mario.
Directors Shigeru Miyamoto, Yoshiaki Koizumi, and Takashi Tezuka had to overcome a number of issues endemic to this new kind of 3D game; three of the biggest were developing a unique camera system, ensuring Mario’s movement would be as fun in 3D as in 2D, and helping players acclimate to this overwhelming third dimension. The camera couldn’t be anything other than free and dynamic to work at all, and while the final product is deeply flawed it is surprisingly strong for such an early attempt. Anthropomorphizing it as a Lakitu cameraman filming Mario’s exploits was also a smart move, allowing the game to contextualize the concept and, arguably, helping integrate it into the adventure. As for Mario’s movement, the months Miyamoto spent perfecting his friction and weight in a “garden” he built just to test them paid off; the plumber moved with a level of control that was intuitive, deep, and satisfying.
More difficult to fully explain, though, is how its design helped players understand and master the space. It’s a complex array of graphical and stylistic choices, the momentum and weight of the title character, a deeply satisfying difficulty curve, and the design and architecture of the world itself. And for this article, I’d like to go through the latter, because in all the discussion surrounding Mario 64 I feel it may have less focus than it deserves. And it’s one that touches every other aspect of the design.
How it relates to difficulty, for instance, is important because the new territory Mario 64 was in demanded a challenge that was fair and encouraged exploration. And like most Mario games, the challenge never comes from combat but from the environment itself. This is the most placid start to any Mario platformer by far; there’s no time limit (a first for the series), no mini-boss fight, no dramatic opening, no music. You’re not even on a particularly strong impetus to go in the castle. It’s just you, Mario, and some lightly detailed polygonal ground. And the slow walk, lazy hills, trees, and the wooden bridge draw your eye. The opening space lends itself to experimentation and play; most first-time players probably stumbled onto the absurd “triple jump” by accident. There’s a reason why anecdotes about exploring Mario’s leaps for upwards of an hour before even going inside the castle abound; movement is satisfying on a purely kinetic level. Every main action – running, jumping, swimming, climbing – can be practiced before you even enter the castle and search for the missing stars. And beyond allowing so much experimentation, the topography is inviting and warm. Even when you enter the castle, the layout and camerawork of the first floor makes it easy to find your way to the first world, Bob-Omb Battlefield.
The level’s difficulty isn’t high, but it feels like it is: water bombs drop on Mario’s head, a humongous Chain Chomp guards the critical path, and all sorts of obstacles litter the way up the mountain. But the game includes safe zones spaced liberally between sections, and threats are survivable. If you choose to go after the next star on the list the game gives you, it can only be procured through a race to the same location as the first, only now with a time limit demanding a greater degree of skill (Whomp’s Fortress, the next world, does something similar). You can’t get every star in Battlefield without leaving to acquire the Wing Cap power, demanding you go to new areas – each one with more complex architecture than the last. The difficulty in the first four worlds is raised just a notch at a time, with two major spikes: the slippery and vertiginous Cool, Cool Mountain and the linear “Bowser in the Dark World” level, the latter of which serves as a climax before Mario can continue further into the castle. The other two Bowser levels are the same, traditional platforming sections that punctuate each act with linear, perilous routes.
But while the difficulty is paced incredibly well, what’s really commendable is how it manages that while also allowing a great – but not entirely free – degree of open endedness. After beating Bowser the first time, you can simply farm for stars in the first floor instead of going through the basement, and you can skip stars or even entire worlds, but eventually you will have to enter more dangerous areas. You’ll likely see them all, even if only Battlefield, the three Bowser levels, and Dire, Dire Docks (which initially blocks the second) are required to complete the story. And because the game has such a variety of challenges and losing all your lives isn’t punitive, you can simply try another area if the current one is too troublesome or unsatisfying.
Every world in Super Mario 64 has something special to say. New biomes come with basic ideas – moving underwater, slippery snow, fire that causes Mario to run uncontrollably – which are presented in largely safe spaces before being explored further, a design principle central to Mario. That “garden” area Miyamoto used to perfect the movement can be seen everywhere: those little parks in Bob-Omb Battlefield like the field with the wooden pegs, the ice sculpture in Snowman’s Land, the unique islands of Lethal Lava Land. This design choice compartmentalizes challenges, helps players create mental maps, and allows different avenues to explore. Even when it reuses ideas, like a second footrace with Koopa the Quick or fight with a Big Bully, it reframes it in a new context. It’s immaculate design, and for all the clunkiness of the camera and controls today the game was very much ahead of its time in the union of environmental and mechanics.
These more broadly experimental environments get fewer and further between as the game goes on, due to later levels more focused on exploring a single concept (like Tick Tock Clock, a vertically oriented world based around an unceasing system of gears) or the studio being pressed for time while making the later levels, but for first-time players they really are wonderful ways to learn the controls and safely goof around. And the later levels supplement this with entirely unique gimmicks to explore, like Wet-Dry World’s changing water level or the ability to switch between the two forms of Tiny-Huge Island. They’re proposals for ideas that could power entire games.
Also note its use of bottomless pits, a Mario staple. They’ve been cut down a bit here, and by necessity are used in a different way. In general you have to go around them, instead of over them, and many challenges don’t demand you interact with them at all. This is important, because a bottomless pit is far more terrifying in 3D than 2D. It’s harder to gauge distance and space without the cold clarity of lines telling you which part is safe and which isn’t, especially when you consider how little familiarity the vast majority of new players would have had with this kind of locomotion. While he would betray this somewhat with thin, precarious walkways in late game levels, Miyamoto understood that fundamentally, movement in 3D cannot be as easy or precise as in 2D. While our ability to create and traverse 3D digital spaces has gotten better by every metric, they still need to account for perception to at least a slightly greater degree than 2D. You can do precision platforming, just not pixel perfect precision.
But the biggest single environmental element played with in Mario 64 may be verticality. While the Mario side scrollers of course had vertically-oriented levels, they were still on a 2D axis and could be explored with ease. Navigating a three dimensional space is much more taxing on a number of levels, particularly in gauging and reacting to depth. A number of early 3D games compensated by creating largely flat plains with limited vertical movement, but that effectively removes all the potential a three dimensional world implies. Mario 64 felt a need to be a definitive text on the possibilities of 3D gaming, so looking at how it presented and explored verticality is illustrative.
Let’s go back to the castle grounds at the start. What’s important in this case is that there are almost no purely horizontal surfaces. Everything is on some kind of slant or slope, with subtle gradations that make it feel “realistic” (which is important for movement; you have to trust you can move about a space) and show off the engine. The game gives you a mostly flat pathway to the castle, but it’s actually out of view of the camera; it’s significantly easier to simply walk up the hill and realize elevation can be climbed. Once you accomplish that, you immediately start to wonder about the limits of Mario’s gait.
But once you go into the castle, things abruptly change. You’ll note the first star in Bob-Omb Battlefield is on top of a mountain; all demand you climb at least part of it. The five levels that make up the first floor are all based around elevation in distinct ways. Battlefield, Whomp’s Fortress, and Cool, Cool Mountain all involve ascending or descending dominating mountains, and Jolly Roger’s Bay trades climbing for underwater depth. Big Boo’s Haunt, which splits the first and second castle sections, plays with this as a multi-floored mansion. In the basement, we switch gears again to four mostly flatter levels. Hazy Maze Cave is a claustrophobic labyrinth, while Lethal Lava Land and Shifting Sand Land are horizontal obstacle courses based around new environmental hazards. And finally, the upper floors mix these two directions, with stages that mix or take horizontal and vertical direction to an extreme. Throughout all of this, you’re slowly forced to push yourself into less secure spaces and more demanding challenges.
And what’s shocking is that the team didn’t use extensive or particularly detailed planning before doing this; according to Miyamoto, they only made quick sketches for levels before developing each one. That they manage to express distance and space so well is incredible. And there are different kinds of verticality, from islands to interior spaces to areas built for wall-jumping to slopes of varying extremes. When you take that into account, along with the wide variety of challenges and mechanics and stuff with which the game is bursting at the seams, it’s clear EAD was intent on cramming every possible idea they had about what moving in a three-dimensional space entailed into the game. Some don’t work or work as well as they should, but the wide array of ideas and ability to jump from one to the next at a whim makes them all additive.
Other, smaller details further the game’s immersive and inviting design. The interactive title screen where you pull Mario’s stretchy face was an early way to explore 3D. Puzzles – some of which may have been initially built for The Legend of Zelda: Ocarina of Time, which was in development simultaneously – encourage environmental interaction. Odd secrets are everywhere, with entire worlds hidden away, for the same reason. Koji Kondo’s score is exquisite. It even plays with nostalgia in unexpected ways, like how Mario can only enter most worlds by jumping through a painting, a nod to Super Mario Bros.’ end of level flagpoles that have to be cleared with a jump. But those are mostly supplemental to what you do and the spaces in which you do them.
In Super Mario 64, movement, space, and challenge interact symbiotically; you cannot alter one without affecting the others. That’s excellent design for two reasons: everything is mutually supportive, and you can create entirely new scenarios simply through small alterations. Conceits like “you can fly” or “the ground is slippery” or “some ground is harmful” create possibilities of all sorts, telling a story not of a princess in a castle but about leaping into this incredible digital frontier. And I think that’s why it’s endured, because its ideas about three dimensional space were so strong. It’s why Nintendo has only had to refine Mario’s movement in later 3D games, not drastically alter it: all the basics were there from the start. The biggest and best changes in them are environmental – the micro-gravity of Mario Galaxy, for example – not the plumber, which allows them to innovate and experiment without relying on mechanical gimmicks like Super Mario Sunshine’s water gun (the upcoming Super Mario Odyssey does both, in a way that appears deeply satisfying). This design philosophy is a testament to the value of a game environment, and how it can be a driving force for innovation and artistic ambition, not just a canvas for but a collaborator with new, exciting ideas.