Postmortem: Ion Storm's Deus Ex
December 6, 2000 Page 2 of 5
What Went Right
1. A clear high-level vision. It's pretty self-evident that you can't achieve goals if you're not clear about what they are. We knew with a high degree of confidence what kind of game we wanted to make. This was possible for two reasons. First, Deus Ex is a natural outgrowth of work done by and in some cases with the late, lamented Looking Glass Technologies. We were inspired as well by games made at Valve, Origin, and a host of other places. Many of the things we wanted to do were a reaction to things they (or we) didn't do, didn't do well or couldn't do at all in earlier games. We weren't building from scratch, but rather building on a foundation already laid for us.
Second, and on a personal level, Deus Ex is a game I've been thinking about since right around the time Underworld 2 shipped. I've tried to get a game like this started several times (as Troubleshooter at Origin; in some respects, as Junction Point, for Looking Glass). Those games didn't happen for a variety of reasons:
- I didn't or couldn't sell the concepts to the money people.
- Then current technology wasn't up to the job.
- I didn't have a team that wanted to make this kind of game or the resources to build one.
- Publishers weren't interested in a first-person, cross-genre game.
Still, I never stopped thinking about these games and, despite their failure to reach production, they laid much of the conceptual groundwork for Deus Ex. The lesson here is that if there's a game you really want to make, don't give up on it. Someone will be foolish enough to give you the money eventually.
As an interesting (I hope!) historical footnote, I include here for the first time publicly, complete with typos and misspellings, the very first proposal I ever submitted for the old Troubleshooter concept back at Origin. Note the budget and projected release date and, oh, those system requirements! Note also the similarities (and differences) between Troubleshooter and what eventually became Deus Ex. [Troubleshooter Proposal]
1.0 High Concept:
It's Underworld-style, first-person action. But this is no fantasy. It's today. The real world.
No monsters. No magic. All action.
Everyone knows the movies: Die Hard, Passenger 57, The Last Action Hero, Under Siege, Dirty Harry...
Everyone knows the stars: Arnold Schwarzenneger, Steven Seagal, Bruce Willis, Wesley Snipes, Clint Eastwood...
Everyone knows the weapons: .44 Magnum, Ingram Mac-10, Atchisson assault shotgun, Browning High-Power, mini-Uzi...
Everyone knows the situations: It's you against the world, you against terrorists, psychos, the dregs of society. They're armed with high tech weapons and they've taken hostages.
You know what to do....
The question is, are you good enough?
2.0 Why this is an ORIGIN product:
It's Hollywood-inspired, big-budget, non-stop action. It's significant new technology. It's a logical extension of our existing first-person line -- we have fantasy covered with Underworld, we WILL have science fiction covered with Bounty Hunter, we have the real world covered with... well, we don't. Troubleshooter is bigger than life, but it's clearly rooted in the real world. I can't believe no one's done this before -- we have to jump on it before someone else does!
3.0 Product Overview:
You're an ex-cop turned "security specialist." That just means you get all the dirty jobs no one else has the guts to do. When the government or the police or business can't handle a problem, they call on you. Bomb threats? You get to check 'em out. Hijackers threaten to take over a plane? You end up on board. Some radical group takes a millionaire's daughter hostage? You get the call to go in and get her out.
You scope out the situation, checking maps and photos, walking around the site, probing for the best way in, the way that will put the fewest innocent people at risk. You try to talk a madman into surrendering before he blows himself and his hostages to kingdom come. You crawl through air ducts and sewers hoping you don't attract the attention of the bad guys with all the guns. You shoot it out with terrorists wielding enough firepower to take on a third world army.
Troubleshooter is a mission-oriented action simulation with no huge plot -- just get in and get out of each mission. Maybe 10-30 minutes of action per scenario. None of this 100 hours to finish the game and get your reward stuff. Like a flight sim, but it's just you, on the ground, with a gun.
I originally envisioned this as all new technology, but I could probably leach off of Bounty Hunter, once that project gets going. In game play, I see it being like Underworld in the richness of its world simulation, but like Wolfenstein in its emphasis on action over roleplaying and inventory manipulation. Ideally, I'd like to incorporate a head-to-head modem/network option, allowing one player to be the bad guy and the other to be the troubleshooter.
4.0 Technical Overview:
IBM PC 486, 4 Megs RAM, 320 x 200 VGA, full sound board support. Mouse, joystick and keyboard supported.
Traditional ORIGIN buyers. I also hope the basis in reality and the short duration mission structure make Troubleshooter appeal to overworked older folks (the ones who have the money to buy machines capable of playing our games...) who just want to work off some frustration and then get back to their real lives.
Planned ship in Q4 (March '95).
High. There are all sorts of technological unknowns, things I want to do that haven't been done before. All in all, this is probably the toughest project on my wish list, but it might be the most satisfying... We might be able to minimize the risk by leaching off of Bounty Hunter.
Looking for concept approval so we can Go For Script.
Several years passed. Lots of games somewhat like Troubleshooter came and went. Game budgets went up dramatically -- $500,000 indeed! I left Origin and go to work for Looking Glass. Troubleshooter stayed on my mind.
In the fall of 1997, before Ion Storm entered the Deus Ex picture, I drafted a manifesto -- a description of an ideal game -- and also a set of "Rules of Role-Playing." Much of that material ended up in an article published in Game Developer ("Remodeling RPGs for the New Millennium". Here (for more of those historical reasons mentioned above) is the original draft of my Rules of Role-Playing, circa 1997:
The Rules of Role-Playing
- Always show the goal. Players should see their next goal (or encounter an intriguing mystery) before they can achieve (or explain) it.
- Problems not puzzles. It's an obstacle course, not a jigsaw puzzle. Game situations should make logical sense and solutions should never depend on reading the designer's mind. And there should always be more than one way to get past a game obstacle. Always.
- No forced failure. Failure isn't fun. Getting knocked unconscious and waking up in a strange place or finding yourself standing over dead bodies while holding a smoking gun can be cool story elements, but situations the player has no chance to react to are bad. Used sparingly, to drive a story forward, O.K. Don't overuse!
- It's the people, stupid. Role-playing is about interacting with other people in a variety of ways (not just combat… not just conversation…).
- Players do; NPCs watch. It's no fun to watch an NPC do something cool. If it's a cool thing, let the player do it. If it's a boring or mundane thing, don't even let the player think about it -- let an NPC do it.
- Have you patted your player on the back today? Constant rewards will drive players onward. Make sure you reward players regularly. And make sure the rewards get more impressive as the game goes on.
- Players get smarter so games get harder. Make sure game difficulty escalates as players become more accustomed to your interface and more familiar with your world. Make sure you reward the player by making him or her more powerful as the game goes on.
- Think 3D. A 3D map cannot be laid out on graph paper. It has to take into account things over the player's head and under the player's feet. If there's no need to look up and down -- constantly -- make a 2D game!
- Are You Connected? Maps in a 3D game world must feature massive interconnectivity. Tunnels that go direct from Point A to Point B are bad; loops (horizontal and vertical) and areas with multiple entrance and exit points are good.
A year or so later, Deus Ex lead designer Harvey Smith clarified and extended the original rules as follows:
DEUS EX Rules Amendments & Addenda
Drafted by Harvey Smith (and endorsed enthusiastically by me) in 1998
- Problems will have multiple solutions. Locations will be reachable
in several ways. All missions, locations, and problems will be specifically
- Skills (and skill levels)
- Augmentations (and augmentation levels)
will rely on a variety of "tools," rather than just one:
- Character capabilities (skills/augmentations)
- Resource management
- Character interaction
will require more thought than "What's the biggest gun in my
- A more relevant question might be, "How do I deal with this situation involving a few intelligent, dangerous enemies?"
should contribute to gameplay -- Whenever possible, show players a
goal or destination before they can get there. This encourages players
to find the route:
- The route should include cool stuff players want or should force players through an area they wants to avoid. (The latter is something we don't want to do too often.)
- Make sure there's more than one way to get to all destinations.
- Dead ends should be avoided unless tactically significant.
overall mood and tone will be clear and consistent:
- Release (through combat and/or reaching a predetermined goal or NPC conversation)
The details of Deus Ex -- plot, character, game system design -- all changed radically since the days of Troubleshooter and manifestos and rules and rules addenda, but conceptually the game still follows most of the rules and meets the ideals outlined in the Game Developer article. With these conceptual tools in mind, the Deus Ex team was able to assess design decisions and game system specifications, in light of what we wanted players to experience during the game and in light of our ultimate design goals.
So What Were Our Goals (In the Beginning)?
How did we intend to move from abstract ideas to game design specifics? We had to take our thinking to a deeper level. We had to start thinking about what we wanted players to be doing and thinking about as they played the game, rather than what we would be thinking about as we developed it.
This led to some critical concepts, outlined here:
- Who are you? We wanted players thinking about who they wanted to be in our game world. Character differentiation was to be paramount. Even more important, though, was the desire to ensure that those character differences would be more than cosmetic -- they had to be expressible, minute-to-minute, in a deeply simulated game world, resulting in a unique gameplay experience for each player. This proved to be a far more important goal than any of us expected it to be.
- How do you behave? We wanted players thinking about how they wanted to interact with our game world. We knew we'd have to wean players from traditional puzzle/solution thinking and show them that Deus Ex was a game of problems (not puzzles!), all solvable in a variety of ways. This seemed critical to making character differentiation meaningful. The idea was to create a believable world and then offer game systems that encouraged players to explore that world in whatever way or ways they chose. The game would tune itself (however slightly) to the player's play style rather than forcing the developers' desired play style on players. We were tired of games that kept us on rails, offering the illusion of freedom and interactivity but without the reality, and we hoped players were as tired as we were of guessing what developers had in mind. We're a long way from being able to create a game in which players are truly free to do whatever they want -- believe me, there's plenty of illusion in Deus Ex -- but we knew we wanted to start taking at least some steps on the road to player control.
- Are you willing to pay the price? "Choice" and "consequence" were the two most frequently uttered words during our two to three years of development. What good is player control if all choices lead to the same result? Without real, predictable consequences, choice is irrelevant. (Which is probably why so many games seem so trivial -- they are trivial!)
- Are you there? We wanted players to feel like they were actually there, in the real world. We wanted this not only because we thought it would be cool (though we did think that!) but because we thought it was critical to making the above concepts actually work. If players were going to solve problems, they needed to be able to make some informed guesses about how those problems worked. If they were going to deal with consequences, they needed to be able to predict what those consequences might be. In other words, they needed to be able to apply some real-world common sense. If you fire a gun in the room you're sitting in as you read this, you can pretty much tell what's going to happen. (If you're in a public place, all hell is going to break loose; if you're at home alone, your neighbors might complain about the noise but that's about it…) Fire a gun in a game set in a fantastic alternate dimension and there's no telling. How can you possibly make a plan and execute it if you can't apply simple, real-world logic to the most straightforward situation imaginable? We wanted that real-world common-sense stuff. Firing a gun in Deus Ex should result in the real-world response. (Speaking personally, I was also just sick and tired of goofy fantasy settings and alien invasions.) Everything in the game would be real, based on something real or based on something someone, somewhere believed to be real. (We can show you the research behind most everything in the game, no matter how outlandish it may seem.) We looked for two kinds of real world spaces: Those that were naturally great game spaces (highly interconnected, multi-level stuff) and places people would just enjoy poking around in ways they couldn't in the real world (such as the White House). There was never a time when we thought realism should get in the way of fun -- anytime that happened, reality lost.
To recap: Know what your gameplay goals are and what kind of experience you want players to have before you spend ten seconds thinking about anything specific. Nice talk, but what did clear goals, manifestos, and commandments buy us?
2. We didn't skimp on preproduction. We spent the first six months of I (before we licensed a game engine), with a team of about six, just thinking about how we could turn our high-level goals into a game. We hammered on the setting and decided to move the game into the near future to buy ourselves some room to play around -- the real world, as we quickly discovered, was very limiting. Ultimately, we settled on a conspiracy-oriented background.
Here's what we had when we started: the very first design proposal (again, as is) for Shooter, our ironic working title for a game we never intended to be "just" a first-person shooter. [Shooter Proposal]
Real-world spaces, such as the Statue of Liberty in New York City, can be compelling game spaces, but offer unique challenges to game developers.
First of all, ignore the projected ship date of Christmas 1998. That was never possible, not for an instant. I don't know what I was thinking. Anyway, other than that (ahem) little misstep, the original Shooter doc does a pretty good job of describing the game that eventually became Deus Ex. Details changed. System specs definitely changed, but overall I don't think anyone can say we didn't deliver the game we said we would.
But how did we get from Shooter to Deus Ex? What were our first steps?
Roleplaying in a World of Secrets, Lies and Conspiracies
Target is Christmas 1998.
Modern Day plus about 50 years (to allow us to fudge reality, where necessary).
Half-Life (Sierra), Fallout (Interplay), The Dark Project (LookingGlass) Goldeneye (N64).
Several unique, real-world locations in the U.S., Europe and Far East.
1st-person 3D with external camera views available as player option.
The world of the 2050's is a dangerous and chaotic place. Terrorists operate openly, espousing a hundred beliefs and killing thousands of innocents to call attention to their causes. The world's economies are close to ruin on a scale not seen for over 100 years. The media openly encourage the worst in mankind. Governments seem powerless to deal with the situation.
In hidden meeting places around the world, a cabal of men and women confer. These are the leaders of societies so ancient and yet so secret, most people refuse to believe they exist. For hundreds, perhaps thousands, of years, they have controlled the world's finances, information flow, weapon production, medicine, religions… Now, as the world descends into chaos, they meet to determine if the time has come to emerge from the shadows, to take overt control of a world they have run from behind the scenes for so long. But, unable to reach agreement, they begin instead an internecine war that brings civilization even closer to the brink of destruction.
At the same time, in several major cities around the globe, another group of men and women -- the deadliest the world has ever known -- lie in wait. Enhanced senses, increased intelligence, biomechanically supercharged muscles and fantastic weaponry make these augmented agents all but unstoppable. And each is implanted with a tiny, high-tech brainwashing device which, activated, can turn its host into a helpless buffoon or a remorseless, conscienceless, killing machine -- whatever the controller wishes.
You are one of these agents.
Recruited from among the elite of the world's intelligence agencies, and endowed with nearly superhuman abilities, you await a signal from the most ambitious, most manipulative and most dangerous member of the shadowy cabal, a man known only as "Adam." Horrified at the chaos he sees all around him, he has concocted a plan to end the secret war and restore order to the world -- at his signal, you and your fellow augmented agents, will emerge to cripple or destroy each of the ancient and secret societies, in turn, clearing the way for a savior (Adam, of course) to take his rightful place as leader of all the societies. Once in control of the societies, he will offer mankind a simple choice: Obey and live in an orderly world free of pain and suffering, a world of Adam's creation; refuse and die.
Adam wants only to save the world from itself -- it's up to you to stop him.
The question is, can you assemble enough clues to figure all this out? Can you find and recruit the allies you'll need to survive, including the other augmented agents? Can you free yourself and the others from the grip of Adam's brainwashing device and put a stop to the secret society cabalists and the master manipulator, Adam. himself?
Each mission leads the PC deeper into a morass of suspicion, false motives, paranoia and conspiracy involving the highest levels of government, the media and the military-industrial complex. The missions stand alone but, taken together, they add up to a big story, with the player at the heart of earth-shattering yet believable events.
Set in a world very much like our own (if the conspiracy buffs are right), Shooter combines the best of The Manchurian Candidate, Robocop and Colossus: The Forbin Project in a world inspired by The X-Files and Men in Black.
- Exciting, first-person, 3D roleplaying in a world that teeters on the brink of madness.
- True six degrees of freedom 3D engine, polygonal figures, killer AI and an innovative conversation system put you right in the thick of things.
- In-depth world simulation allows players to solve problems in a variety of ways.
- More than 30 core missions, and plenty of optional adventures provide 40+ hours of gameplay.
- Plugs into two popular fantasies -- the millennial madness that's gripping the world, exemplified by The X-Files and Men in Black and a general fascination with conspiracy theories and the desire to play with high-tech espionage toys.
- Emphasis on character development ensures that every player character will be unique. This, combined with our deep world simulation, ensures that each player's experience of the story is different, without having to resort to brute force branching tree structures.
- Non-combat interaction with dozens of unique non-player characters. A simple, elegant conversation system results in non-player characters you really care about. Engage them in conversation, seek information from them, recruit them to your cause, decide for yourself who you can trust…
- Clear goals, constant rewards, varied interactions with people and places as well as varied mission types (including Sabotage, Infiltration, Extraction, Rescue, Intelligence-Gathering, Thievery, Reconnaissance, Assassination, and all-out Combat) keep players coming back for more.
- Goals can be accomplished through stealth, careful planning, undercover work or conversation, through the use of unbelievably high tech equipment or brute force combat tactics.
- No weird "game spaces" -- every map recreates either a real place or a place with an instantly recognizable real-world function. We hope to recreate places like Camp David, the Kremlin (and the tunnels beneath it), the bone-strewn catacombs beneath Paris, Hong Kong's junk-filled harbor, the London underground and more.
- Situations that evolve over time, influenced by player action (or inaction). The state of the world changes to reflect the impact of player choices. Game events and the passage of time turn familiar locations into strange and terrifying places; trusted friends become the most bitter of enemies.
- Engine: 3D 6-degrees of freedom
- Characters: Real time 3D models.
- Creatures: Real time 3D models.
- Surroundings: Real time 3D models, sprites.
- Structures: Real time 3D.
- Vehicles: Real time 3D models.
- Interface: 2D graphics.
- Operating System: Windows 95
- Processor: Pentium 133
- Memory: 16 Megs
- Format: 4X CD-ROM Drive, 2 CDs
- Hard Drive Space: 50/100/200 Megs
- Video: DirectX 5; 8- and 16-bit color (hardware only); All DirectX-supported resolutions; 3DFX, etc.
- Audio: DirectX 5
- Multiplayer Support: DirectX 5 (DirectConnect, Modem, Network, Internet)
Page 2 of 5