HobbyGameDev - Formerly GameDevLessons.com

There are a few things that I would like to say about bottom-up design, compared to top-down design. Before I can do so, I want to make sure that we’re on the same page as to how I use those terms – they mean somewhat different things to different fields, and I suspect it’s likely that some readers have never yet come across these handy terms for discussion.

Top-down design starts from what a mind wants, and shapes that into something a medium can do.

Bottom-up design starts from what a medium can do, and shapes that into something a mind wants.

For example, a top-down art process might begin by thinking, “I want to create an image of Napoleon, in a way that conveys great intensity – this concept is worth attempting and sharing.” The artist would then create the most complete or faithful realization of that idea possible, using whatever tools and format the artist is most proficient in (might be charcoal, oil paint, or 3D Studio Max). The goal is to realize an idea, and if it can be made less evident that this is done by oil paint applied to canvas, by making its focal point what it is showing instead of method by which it is shown, that is desirable in this case.

A bottom-up art process, by contrast, might begin by thinking, “The texture and shine of oil paint on canvas produce a very particular effect – I will find a subject that emphasizes the beauty of this effect.” The artist would then seek a technique and subject particularly well suited to make the most of what effect the artist’s tool and format convey. The resulting work is as much or more about the paint and canvas itself as it is about the subject matter; the grain and texture take part in the creative process, somewhat like a sculptor carving stone in a way that not only considers existing fracture lines, but perhaps even calls attention to them.

Naturally, there is a continuum, or smooth gradient, between these ideals. It is not as simple as being one or the other. By virtue of being represented through a medium, the work will necessarily account for constraints and affordances of how it is made; likewise, by being the work of a human being, it inevitably will come into contact with how the human mind conceives of reality. There is a significant difference, however, in whether the work starts mostly as a chosen idea or mostly through exploratory execution, and that difference is generally evident in the end result.

As a rough test: if a description of the project sounds normal to someone that doesn’t play videogames (or, in our ongoing analogy, study art), then it is probably top-down. The more eccentric, absurd, and arbitrary the description of what the player does (or what a painting is “about”), the further it probably is toward the bottom-up side of the spectrum.

For example, Bubble Bobble is a game where two dragons trap toys, witches, and drunks in bubbles to pop them then eat the fruits, vegetables, and candies that fly out, though if they take too long then the ghost monster chases them until they either die or kill the last enemy – that’s bottom-up, because the game mechanics clearly were not invented to support that haphazard concept, but the other way around. Likewise for Super Mario Bros, Joust, or Breakout. It’s not primarily about wish fulfillment, like being a superhero that fights crime or a soldier that wins the war for freedom – it’s first and foremost about how what’s going on is going on, the what being secondary and subservient to the how. The texture of the “paint” (software) and “canvas” (hardware) are central to what these videogames are, and celebrated instead of hidden.

The continuum extends farther off in both directions, too. Miles below my previous example of bottom-up, but in that same direction, an artist might think, “The texture and shine of oil paint produce a very particular effect – I will explore this effect,” thus abandoning subject altogether.

When the Wii came out, developers began asking, “What sort of interesting things can we do with this Wii Remote?” That was bottom-up design. When people began thinking in terms of designing games for iPhone that used multi-touch, gravity data, saved photos, or GPS, that was also bottom-up. The first racing games were made partly because racing is cool, but mostly because that was one of the few things videogame machines at the time could represent. Likewise for space-based shooting action – an all black background with no buildings or flora to collide with was the sort of thing primitive devices could represent.

That Pong is a crude representation of “tennis” instead of football or baseball was not a matter of its designer being a tennis player, nor was it being designed to satisfy tennis fans.

Racing the Beam by Ian Bogost and Nick Montfort provides a brief history of how the Atari 2600 hardware limitations influenced (and in many cases inspired) the games designed on the platform. Yars’ Revenge was co-designed by Howard Scott Warshaw and what-an-Atari-2600-can-do.

By contrast, when the PS3 came out, developers felt less constrained than ever in terms of computation, and began dreaming about how they might share new stories (Heavy Rain, Metal Gear Solid 4), fantasy characters and settings (Batman Arcane Asylum, Final Fantasy sequels), and appealing concept pitches (make your own physics game in Little Big Planet). That’s top-down.

When a modern videogame is about tennis, racing, or space combat, it is no longer because that’s one of the few physical events that a computer can process and display.

Even though many modern games now have the luxurious option of more faithfully representing top-down concepts, and even though bottom-up has largely taken place at the dawn of platforms utilizing innovative controls and graphics (vector to raster was as huge a shift as 2D to 3D), there is nothing ancient, unsuitable, or undesirable about bottom-up design. Part of why I opted to show paintings above is that the Napoleon is from 1800, Impression is from 1872, as Shimmering is from 1946 – that is, it took a lot of time, genius, and creativity for humankind to reach the sort of bottom-up, format-inspired thinking that the earliest videogames employed.

An inexperienced game designer, yet unfamiliar with the grain of the machine – what impressive things it can very easily do well, what somewhat less impressive things it can hardly handle, what the user’s relationship is like to the experience through the set of controls provided – most often thinks mainly in terms of top-down design. We’re all familiar enough with films and storytelling to think about what happens, where, and to whom in a story, but the typical level of procedural synthesis is that which only knows what it likes when it has tried it, and will thus “borrow” a mountain of assumptions about gameplay and interaction from another game. A certain sort of confidence and understanding of the technology available is necessary for bottom-up design to even be an option.

On this note: to build a clone or derivative of another game, either completely or in terms of core gameplay, is top-down. The elevator pitch for Goldeneye 007 on Nintendo 64 fits in this category. To paraphrase: “It’s like Virtua Cop – enemies react to where they are shot, and the gun can move freely about the screen – except not on rails, and it takes place in the James Bond universe with the player completing missions as James Bond.” Making a videogame version of Monopoly or soccer is top-down, of course, but so is making a game “like Super Mario Bros” or “like Tetris“, even if the original game’s roots were bottom-up design – no less than “resembling Napoleon” is rooted in the complete concept of Napoleon that existed before and outside of Jacques-Louis David’s painting.

An important thing to bear in mind about top-down design is that, despite the remarkable hardware advances of the past several decades, we are still a long, long way from being able to take a purely top-down approach to videogame design. Façade is the closest thing that we have to The Great Gatsby – or maybe The Sims or Second Life are no further, coming towards it from different angles.

Even though PS3 and modern computers give a lot flexibility to think in terms of, “Imagine a world, in the distant future, where…” there’s still a few qualifying sentences implied after that: “In this magical place, all characters will repeat the same things if spoken to multiple times, they’ll have a small set of awkward poses they can switch between to convey expression, and the player will speak by choosing statements from a list of 3-4 pre-written statements. There are only a few things that can be done to interact with the relatively small percentage of the world that can be interacted with, and the solutions to challenges will either be those that we explicitly planned, or possibly some slight variation by clumsily moving objects around in the game’s exaggerated physics…”

Top-down is unlikely to free us from limitations, because by its nature it echoes better than it speaks, it adopts the additional constraints that shaped previous gaming and story conventions much more naturally than it pushes the boundaries to test where they end.

Innovation – particularly the surprising, unintuitive innovations that no one could have seen coming – tend to come from bottom-up. It is no coincidence that the genrification of the industry, the attack of the clones, has coincided with the increase in hardware power inviting more top-down design, while those pockets of gaming that see the most aggressive innovation – indie games, mobile games, experimental games, and little web games – tend to be bottom-up (at least, those that don’t rely heavily on appeal to retro/nostalgia fandom).

The imagination – the source of top-down – copies but personalizes, mixes and matches what it has already experienced, or even when it thinks it is finally being clever and original, has been swept up by the zeitgeist of whatever recent films, best sellers, and news stories have everyone thinking about. There is, of course, a market for this; people like what’s recognizable, and people like getting what they’re looking for. But as the top-down market is predominantly derivative, from a creative standpoint it’s more like being a translator than like being a writer. For the craftsmen and artists among us, it is not our desired line of work to make variations on other people’s work.

With bottom-up the constraints, limitations, and natural texture or grain of a medium push into existence results that are not entirely of human origin. The medium becomes a co-creator, an inanimate but unyielding creative partner, arguing for or against ideas thrown its way by what it will and won’t agree to do.

A top-down development team strives to create as perfect a realization of their concept as their available platform technology and skill strengths can support. The hidden member of their team is the work that has been done before them, by bottom-up pioneers, and they benefit from the market’s feedback on past products.

A bottom-up development team strives to discover as perfect a concept for the platform technology and skill strengths that they have or can learn. The hidden member of their team is the technology itself, and though they don’t have the same benefit of borrowing from the past, everyone benefits when they help create the future.

(Originally posted as part of GameDevLessons.com Vol. 14)

Permalink: http://www.hobbygamedev.com/adv/bottom-up-vs-top-down-game-design/

Posted in:

Advanced, Vol. 14

Tagged with:

Experimental Gameplay, Game Art, Game Design, Popular Entry, Theory