Anti-Design / Backwards Game Design in GoldenEye 007

Jun 26, 2012

I’ve been studying GoldenEye recently–not the newer remake, but the old 1997 game from Nintendo 64. Rare put it together, and Martin Hollis was the producer. This was 15 years ago, back when I was in high school. Now 8 years ago, back in 2004, long after a lot of people forgot about the game, Martin Hollis did an interview about the game and its creation. In it he reveals a bunch of interesting facts about how it was originally inspired by Virtua Cop in certain ways–as he puts it when you hold down the aim button it kind of goes into Virtua Cop mode–or various other tidbits, but, if you’re interested in that I’ll link you to the full interview there’s really a lot of good content in there.

But there’s a few very specific gold nuggets in there that I want to pull to the surface and share today. Those have to do with the ideas of backwards game design, or anti-game design as he calls it in another place.

The excerpt I’m referring to begins as follows:

“One important factor was this. The level creators, or architects were working without much level design, by which I mean often they had no player start points or exits in mind. Certainly they didn’t think about enemy positions or object positions. Their job was simply to produce an interesting space. After the levels were made, Dave or sometimes Duncan would be faced with filling them with objectives, enemies, and stuff. The benefit of this sloppy unplanned approach was that many of the levels in the game have a realistic and non-linear feel. There are rooms with no direct relevance to the level. There are multiple routes across the level. This is an anti-game design approach, frankly. It is inefficient because much of the level is unnecessary to the gameplay. But it contributes to a greater sense of freedom, and also realism. And in turn this sense of freedom and realism contributed enormously to the success of the game.”

What interests me about this, to provide contrast to how game design typically occurs for level design in, say, a shooter or a lot of types of games, is you might first come up with a beat doc of what you expect the player to do in what order. He’s going to collect these documents, he’s going to get this key, he should blow up a helicopter, and then he’s going to escape the base. And then you would design the level around those objectives. You would provide a space in which he can complete the first objective, and then you would assume he would move on to the next and would complete some objective in the next segment of space you give him. You might give him another room with someone to battle, and he progresses past that. In each case in a lot of games, and we still see this a lot of the time, once you complete an area or complete the objective for that area you never go back.

Instead, what Martin Hollis had his level designers do for Goldeneye back in 1997, or I guess rather development in 1995 or 1996 or so, what he had them do was to just focus on making an interesting space… to make these spaces different from one another, to make them compelling to navigate, to make them all have a unique look and feel. They weren’t supposed to be worrying about where the player started, where the enemies were supposed to go, where the objectives would take place–what the objectives even were! Besides, perhaps: get out of this space, so they needed some sort of exit somewhere. But otherwise it was: put some rooms together, put some hallways together. Build a space. And then we’ll worry about how to populate it with enemies and objectives and items later.

This is, as he calls it, anti-game design. He’s beginning by focusing on making an interesting area then populating it, as opposed to the other way around. He uses these loaded words, saying that this approach is, “inefficient,” “unnecessary,” he calls it “sloppy,” “unplanned,” though of course he’s not saying these things negatively. I want to call attention to the fact that without using this approach they would not have gotten those same enormously successful results, the sense of realism and freedom that come from having spaces that are so interconnected and tied together, having so many optional rooms off to the side, having so many different hallways that don’t wind up being used. That was a product of the anti-game design approach. So if that’s inefficient, compared to a usual process that doesn’t produce as good of a result, then I would argue that it is in fact the most efficient way that will yield the result that’s desired. Something to keep in mind: is it sloppy? Is it unnecessary? Or is it the way that this sort of good work happens?

This approach actually extends beyond level design. As Martin Hollis mentions later in the interview,

“We milked the Bond universe in many ways. For example gadgets, I compiled a list of about 40 gadgets from various Bond films, most of which were modeled, and then Dave and Duncan tried to find levels where we could use them. This is backwards game design, but it worked very well. These models were the game design; there was very little written down on paper. And the models were researched and milked extensively. And, importantly, they all gelled together very well. These things helped to make the gameplay and the game style what it is.”

So it wasn’t just that they built levels without plans for what would happen in them, they modeled and textured 40 different gadgets before they really had specific plans for how to use these things, for which would be used with the others, for which would appear in which levels. They focused on what are some cool Bond gadgets, and then how can we shoehorn these in after they’ve already been made, and some of them didn’t make it. There’s, in the data for the game, a number of objects in the memory which never wind up used, not even through cheat codes, and have been pulled out partially but not entirely, because as development progressed they realized they didn’t necessarily need all of these. But they have them there to pull from.

He mentions that they all gelled together very well. It’s no coincidence of course that when you pick out 40 James Bond gadgets, they all gel together very well, because he’s pulling from a very specific source of intellectual property, that has evolved for a very long time, since the creator Ian Fleming came up with the character for James Bond. Since then every studio, every producer that has touched the property has contributed ideas to it based on what looks cool, what will be neat, what’s a great setting, what are some cool weapons or gadgets to introduce… and so that gave them a lot to pull from, in the same way that the Star Wars universe has a lot to pull from, the Indiana Jones universe has a lot to pull from, the Lord of the Rings has a whole bunch of dynamics about how that world works and how those pieces fit together. These gadgets all gelled because of course they’re mostly James Bond gadgets.

The gadgets were inspired by the James Bond films, but so were the spaces, obviously. As Hollis puts it,

“Karl constructed levels based on the film sets, which we visited several times. And Bea constructed characters based on the photos of people and costumes we had.”

So think about this–they’re really utilizing Hollywood film set design, Hollywood film costume design, in their game, by adopting those patterns. Although those aren’t just built off of skilled Hollywood craftsmen, although they certainly played a huge role in it. Those were scoped out real locations in Cuba, and in Russia, and so on. Likewise the costumes were based on, in many cases, real army outfits, and the weapons are all based on real guns. This keeps it grounded in a complex cultural, very realistic environment in which the guns and the costumes and the settings are based on something recognizable in the real world.

Anyhow, there are many ways that this can be applied. It doesn’t necessarily mean that you’re making your levels before you know what to do in them. It doesn’t necessarily mean that you’re making objects before you want to do something with them. It also doesn’t necessarily mean that you go back and write a full complex lore like James Bond’s universe before you start working on your game to ensure that everything will gel. There’s littler ways that this can happen, too.

David Jones, the guy who created the Grand Theft Auto series and later worked on Crackdown, also created the Lemmings game. This was a classic game where you had these little suicidal characters, or rather dumb characters, and you’re trying to real-time strategy them to block each other and build bridges and build stairs, to make their way to an exit. And so the story goes, that derived from sort of a bet David Jones had with a another developer about how small of an animated character he could make, that that small character became the Lemming. It’s a case where he didn’t start with the game’s idea, he didn’t start with the game’s story, or arc, or events, he simply started with a character sprite, and then found a way to build a game to use that.

We followed, of course to much less fame and success, a similar sort of approach with Zylatov Sisters. Before anything else about the game existed, I was modeling the little Zylatov Sisters, or drawing them, just to practice small pixelated animation. Once I had one of those drawn, and color shifted her and got different hair color from her, I decided we could make a co-op little game with guns. That turned into a game for us.

However it might apply to yours, I just want you to think about this week or this month backwards-game design, anti-game design, call it whatever you want. It’s the idea of starting by building something that’s neat or appealing or interesting or fascinating, or worthwhile in its own regard, as a model, as a space, as a character, as a sound effect, whatever it might be, and then work around making that fit. It’s about worrying after you produce it how to turn it into a game.

Anyway, that’s what I wanted to provide for the day.

Supplementary Material

GoldenEye 007 N64 Level Maps (Exported using Goldeneye Setup Editor 2.3.4 by SubDrag)

The 2004 interview with Martin Hollis (Producer of GoldenEye 007 for Nintendo 64)



Learn and practice team game development with Gamkedo Club.
Membership worldwide. Professional support. Proven process.




Subscribe by e-mail to receive weekly updates with Gamkedo.Community interviews and YouTube training videos for game developers!



8 Comments

  1. kloplop321 says:

    I do like the audio format, It makes it nice while I do my morning routine of going through all my RSS feeds before work, it also makes it easier to share to others that are also busy.

    Thanks for writing regularly.

  2. Smudge says:

    I haven’t been to this site before so I can’t comment on your podcast in comparison to your writing, but I enjoyed listening (your voice wasn’t annoying by any means) and this format is convenient, for me at least, while I’m working or somesuch.
    Cheers for the podcast at any rate, will have to have a proper look around the site when I get a chance.

  3. Jonny D says:

    I like reading these better (it’s more convenient in my current situation), but I appreciate that a podcast exists. As long as there’s a transcript, I’m happy.

  4. Matt Postema says:

    I really like the podcast style. It allows me to fill my boring commute with interesting thoughts.

  5. […] on positive responses to providing audio for the GoldenEye backwards design entry, I have also prepared this post in audio […]

  6. duck says:

    I like to read these articles on my mobile browser so if you can keep a transcript as well as audio please.

  7. […] to life. In some ways those images can sort of be your design document, in a way that as I said in the previous GoldenEye entry, they built in many cases models and textures for all kinds of gadgets for the levels before they […]

  8. Mr Blade says:

    Audio format for the win! I listened to this while I wrote C++ socket code. Ahhh it only burns my soul a little bit.

Leave a comment

Comments Form
  • Your email will not be used for any purpose other than to update you in replies to your comments. Your website address will be shown as a link from your name with your comment. Your profile photo is auto-magically provided by Gravatar.

All contents Copyright ©2017 Chris DeLeon.

Site production by Ryan Burrell.