In defense of GTA, part 2: The real problem in video games

In part 1, I talked about the development of video games to date. I explained how we ended up with games with complex multi-path plots, and games with worlds modeled in true 3D. However, game developers started to hit problems when they began trying to build 3D games with complex plots…

The complexity problem

The issue of managing game complexity had been discovered by text adventure programmers back in the 80s. If you give the player a single tool (a gun) and a single set of adversaries (invading aliens), the number of possible interactions you have to program responses for is very limited. The player needs to be able to fire the gun, maybe reload the gun, perhaps pick up additional ammunition or other types of gun. The gun is either successfully targeted at an enemy when fired, in which case some damage is done to the enemy, or it isn’t.

Now consider a world in which the player has a gun, a flashlight, a screwdriver, and some health-giving combat rations; and suppose the enemy base has locked doors and guard dogs. What if the player tries to club an enemy to death with the flashlight? What if he tries to distract a guard dog with the food rations? What if he tries to open a locked door by unscrewing the hinges? Suddenly the number of possible actions the player can take increases significantly. In mathematical terms, there is a combinatorial explosion, as each object can potentially interact with every other object, and the game designer needs to decide what will happen for each combination. Even if the answer is “nothing relevant happens”, that’s still a design decision that must be made. Furthermore, too many “nothing happens” or “you can’t do that” responses will destroy the player’s suspension of disbelief, or even become outright annoying.

So as the complexity of the new realistically-drawn 3D worlds increased, the problem of making those worlds behave realistically increased much faster.

In addition, text adventure programmers discovered that providing players with genuine choices led to increased complexity. What if the knight sacrifices his henchman to escape the dragon? Who will perform the actions the henchman would have performed in the game? What about the fact that there’s still a dragon roaming around, shouldn’t that impact the plot?

Most satisfyingly complex plot-driven video games found a convenient solution to these problems: they put many completely arbitrary restrictions on what the player could do, in order to ensure that the plot didn’t “break”.

To see how these restrictions are enforced, it suffices to look at the Final Fantasy series of games, which became incredibly successful after Final Fantasy VII introduced full 3D graphics.

In Final Fantasy, if you encounter someone who has to remain alive in order for the plot to work, then you simply cannot attack that person. It doesn’t matter how annoying they are, or how many weapons you have–they are invulnerable. No explanation is given in the context of the game; they just are.

Furthermore, while the worlds of Final Fantasy appear large and open, they are full of invisible walls. If you are meant to wander through a forest, and there’s something to the north that you’re not meant to discover until your return journey, the game developers will think nothing of placing a temporary invisible wall there to prevent you taking that path too soon. So while the Final Fantasy games are almost universally acclaimed for their rich plots and character development, as well as their state-of-the-art graphics, there’s no denying that they lack realism and immersion.

Nintendo’s acclaimed 3D Legend of Zelda games are more immersive, as they mostly use clever world design rather than invisible walls to limit the player’s roaming. However, they take a surreal approach to preventing unwanted conflict: if a creature or person in the game is friendly, then you can swing your sword at them as much as you like, and it will simply pass straight through them or bounce off of them harmlessly–because that is what the plot demands.

Which brings me to Grand Theft Auto III.

The GTA revolution

Grand Theft Auto III (henceforth GTA3) was the 3D sequel to a moderately successful franchise of 2D games. The earlier games had presented the player with a top-down view of city streets, and allowed him to drive vehicles around, committing crimes and evading law enforcement. While there were tasks to perform to advance the game towards “winning”, players were given fairly free reign to decide where to go and how much mayhem to cause.

The revolutionary aspect of GTA3 was that it took this idea of player freedom even further, modeling an entire city in 3D, complete with parking lots, outdoor cafes, car showrooms, gas stations, apartment buildings, warehouses, airports, and all the other architectural features found in cities across the USA. These detailed virtual worlds were populated with hundreds of people–emergency services crews, police, shoppers, drug dealers, businessmen, construction workers, bus drivers–and, of course, criminals. An attempt was made to give the non-player characters their own personalities and agendas, and to model the physics of the world somewhat accurately. The game launched the genre known as the sandbox game, where you have no mandatory goals or tasks, and can do what pleases you rather than what will advance the plot.

Early on in GTA3, I was driving beneath the elevated railway lines in one of the seedier parts of town. As I cruised towards the Italian district of the city, I suddenly saw a piece of unexpected drama playing out on a nearby sidewalk. There was a woman, who from her dress was presumably a prostitute. She was being punched by a man who I assumed was either a john, or her pimp. I slammed on the brakes, jumped out of the car, ran over–and hit the guy with a baseball bat I was carrying. He stopped attacking the woman, who ran off, and began attacking me instead. I ended up beating him into unconsciousness. Unsure of what to do next, I looked around for the woman, but she was no longer in sight. I started walking back to my car. As I reached the car, I heard sirens. I looked around again. An ambulance was approaching. It stopped by the injured man, and two EMS workers got out. They loaded the unconscious man into the back of the ambulance, got back in, and drove off towards the nearby hospital.

I was amazed.

I am recounting this true story because if you’ve heard nothing else about GTA, you’ve probably heard that the game lets you have sex with prostitutes and then kill them. That seems to be the starting point of almost every critique of GTA I see, even today.

Yes, it’s true. GTA lets you have sex with prostitutes. It also lets you kill them afterwards. However, as I hope my own experience illustrates, it also lets you choose a totally different path. The game sets up a complicated virtual city that obeys certain somewhat realistic rules, and lets you decide how to behave.

If I had chosen to do so, I could have killed the ambulance workers. I tried that later on. That time, a passing cop saw me, and before long I was being chased by multiple police cars. Unlike many video games, violence in GTA has consequences in the game.

Alternatively, I could have stolen the ambulance while they were trying to load the guy in the back. If you steal an ambulance in that way, you can then choose to take part in emergency rescue missions where you pick up wounded people and ferry them to hospital. You can also steal a taxi cab, and try to make money ferrying people safely around the city as quickly as possible. If your driving is too dangerous, they’ll bail out as soon as they can, shouting that you’re crazy.

You don’t have to shoot people to kill them in GTA either. You can run them over, deliberately or by accident. In fact, some missions are considerably easier if you drive straight into a gang of assailants rather than attempt to attack them on foot. Other missions can be failed instantly if you accidentally plow your vehicle into someone you’re supposed to be saving.

People can also die when vehicles blow up–which they often do, either because someone has shot the gas tank several times, or because you’ve booby-trapped them with explosives, or because you’ve managed to get your hands on a rocket launcher, or because an adjacent vehicle blew up and set them on fire. Sometimes a bystander will avoid an explosion, but get hit by a piece of debris. You can drive up onto a parking lot rooftop, drive at a ramp that points out over the edge, leap out of the vehicle at the last moment and roll across the ground, then watch as the car sails off the roof,  through the air, and crashes into a crowd of pedestrians below, crushing some of them to death.

You’re not the only person causing casualties either. The police shoot at criminals, and sometimes kill them. On occasion, when in hot pursuit, a cop car will hit and kill innocent bystanders.

This kind of detail, and the associated freedom of choice, was groundbreaking when GTA was released. When the second GTA 3D game was released, titled Grand Theft Auto Vice City, the degree of choice was increased still further.

In Vice City, you can get a job delivering pizza on a moped, hurrying to get the food there before it gets cold, then returning to the pizza restaurant for more. You can still work as a cab driver, but you can also get involved in a fight with a rival cab firm, put them out of business, take over their offices, and earn a regular income. You can buy a car sales showroom, then steal cars to order for customers. You can take part in demolition derby races at the stadium for prize money and prize cars. You can compete in illegal street races. You can use an ice cream van as a front to start a drug dealing operation.

As alluded to earlier, there are missions in GTA games, which advance the overall story in movie-like fashion towards some sort of resolution, and the end credits. However, you don’t have to attempt any of the main story missions. In fact, none of the action I’ve described so far has anything to do with the main story of any of the games; it’s all just incidental detail, part of the sandbox. Personally, though, I like story. I’ve played through the story missions of all five PS2 GTA games. So in the next part of this set of postings, I’ll talk about the media controversies around the GTA games, from the perspective of someone intimately familiar with them.

In the mean time, you might want to read the story Rage against the machines from Prospect magazine, which talks about the idea that video games in general are a brain-damaging addiction, and describes why that’s an outdated idea.

In defense of GTA, part 1: A short history of video games

You’ve probably noticed that a new Grand Theft Auto game is out, GTA IV. As usual, the release of a new GTA has resulted in a new round of articles criticizing (or outright excoriating) the game.

I’m a big fan of GTA. I’ve played every 3D GTA game from start to finish. As such, I feel I can provide an informed perspective on the game series. I see a number of annoying misconceptions and deceptions repeated time and time again, the most infamous of which is the claim that the game rewards you for killing prostitutes.

I’d like to explain why I continue to play every GTA game released. But before I can do that, I need to talk a bit about the history of video games, so that I can explain exactly why GTA was (and is still) so groundbreaking.

(In the text that follows, please excuse any lapses in chronology; my focus here is on general trends in game design, rather than the minutiae of which games were released when.)

The first wave of video games: the arcade

Video games are a comparatively new medium. While a few experimental games were created as early as the 1950s, it was in the 1960s that the first recognizable video games began to appear on university computer systems. In the early 1970s, these primitive games began to appear in amusement arcades.

Early games mostly belonged in one of two categories: sports games and shooting games. The first amusement arcade game, PONG, was an example of the former. It presented a stripped-down approximation of a familiar competitive sport, in this case tennis. Other games attempted to simulate baseball, hurdles, and other sports events.

Perhaps the earliest example of a shooting game in the arcades was Space Wars, an adaptation of a mainframe game. It allowed two players to maneuver spacecraft on a vector graphics screen, and attempt to destroy each other with missile fire.

In 1978, Taito launched Space Invaders, which introduced two vital changes to the shooting game formula. Firstly, it was a single player game, so players no longer needed to find a friend of similar ability in order to enjoy play. Obviously this was a relief to the kinds of people who played video games, but it was the second innovation that really changed gaming: Space Invaders presented the player with an enemy whose forces were apparently overwhelming.

It was a massive hit, and set the pattern for hundreds, perhaps even thousands of video games. Sometimes the enemies were Japanese aircraft, sometimes egg-laying aliens, sometimes undersea creatures; sometimes the screen scrolled horizontally or vertically, or even diagonally. However, the basic situation was always the same: large numbers of enemies were invading for no adequately explored reason, and as they moved around the screen you attempted to kill as many of them as possible. The genre became known as the shoot-em-up. While it soon became formulaic, and is no longer popular, for a while almost every other video game seemed to be a shoot-em-up.

In the 1980s, a few new primeval video game formulae were invented. The game Scramble turned the tables, making the human player the aggressor in an invasion attempt; this idea was repeated later in games like Zaxxon and R-Type. Mazes became popular, in games such as Pac-Man and Berzerk. Pac-Man also introduced many people to the game mechanic of evading enemies rather than destroying them, a formula also used in Q*Bert and Crazy Climber. Driving games made an appearance, from purist Formula One simulations to avoid-the-enemy variations with cars in mazes. Eventually true 3D graphics began to appear, with games such as Battlezone, a commercial game which was adapted into a military tank battle simulator.

One thing that was clear on entering any 1980s video arcade was that there was a tremendous focus on shooting things, and not a great deal of story-telling going on. Partly this was because of technological limitations, but mostly it was because the purpose of an arcade video game was to extract as many coins from customers as possible. Games therefore attempted to give the most intense experience possible, so that even a five minute gameplay session could feel exhausting. The easiest way to achieve this kind of intensity was with lots of violence, often enhanced with pulsating sound and visuals. Eighties games such as Robotron: 2084 and Defender still rate amongst the most intense video game experiences devised.

The second wave: console games

While arcade games battled to fit more and more killing on screen, video games began to undergo a parallel evolution in the home. In 1977, the Atari 2600 introduced the idea of a video game console which could play any number of different games, loaded onto it from cartridges. Prior to this innovation, home video game consoles came with a fixed set of games built in.

Cartridge-based console gaming lowered the investment needed to put a game into commercial production, and reduced the amount of money the customer had to risk to try a new game. At the same time, companies like Atari were making incredible amounts of money, so game designers were allowed to experiment with games that would have been unsuitable for arcades. And with no requirement to keep game sessions short in order to pull money from the player’s pockets, games could become longer.

So it was that another video game genre began to become popular: the adventure. Often modeled on pen and paper role-playing games, adventure games offered players the chance to take part in a comparatively lengthy quest, which often had some sort of narrative purpose behind it. For the first time, games began to explore why the protagonist was risking his life and why he was being attacked. Furthermore, almost all adventures had an ending in which the player could actually win the game.

It wasn’t long before some adventure games began to offer players true choice, with multiple ways to “win”–for example, by rescuing the princess, defeating the dragon, or recovering the treasure.

The third wave: home computers

During the home computer boom of the 1980s, the price of game distribution fell further, as software for home computers could be duplicated on regular audio cassettes. In addition, home computer programming languages put the means of game development in the hands of millions of people. This led to another explosion in the degree of complexity found in video games.

As well as graphical adventure games, there were adventures where the gameplay was represented purely as text. One of the earliest was called simply “Adventure”, and was developed in the 1970s on business computers which lacked any kind of graphical output. It was played at universities and colleges all over the world, before being adapted to run on home computers.

Text adventures were able to use the power of the written word to represent situations that the primitive computer graphics of the time were not suited to. In addition, because the player could type complex commands, it was possible to affect the game world in more sophisticated ways than was feasible with the four-way joystick with single fire button that was common at the time.

Hence for a number of years, text adventures led the way in showing what video games could be capable of. Most text adventures avoided violence entirely, encouraging players to find other ways to make progress within the game universe. Characters other than the player-controlled protagonist were commonplace, and they soon began to exhibit their own programmed personalities and act according to their own distinct goals. For the first time, games began to feel comparable in complexity to more established artistic genres; a text adventure game could feel like you were actually in a novel.

As the worlds modeled in the games became larger and more complex, many players began to find that exploring and mapping the world was an enjoyable and entertaining activity in and of itself. Games began to be advertised on the basis of how large they were and how many different locations they featured.

The death of the arcades and the rise of 3D

Meanwhile, the shoot-everything approach of arcade video games had run into a dead end, and the industry had collapsed. There were lean times for consoles too, as the limitations of their low priced hardware prevented their games from competing with those found on increasingly powerful home computers.

By the late 1980s, high resolution color graphics were commonplace on most home computer owners’ machines. This made it feasible to use full 3D color graphics in video games. One influential early 3D game was Wolfenstein 3D, which challenged the player to lead a heavily armed soldier into a maze-like Nazi encampment.

While there had been 3D action and adventure games on home computers as early as 1980, Wolfenstein 3D achieved notoriety for the level of violence depicted. The player was encouraged to kill hundreds of German soldiers, who were seen falling to the ground in a spray of blood. In addition, killing the “boss” at the end of a level resulted in an instant replay of his death. Although the game was controversial (and arguably tasteless), its high speed 3D graphics were groundbreaking, and it won many awards. It also kick-started the video game genre known as the first person shooter (FPS), still incredibly popular amongst Windows gamers.

The makers of Wolf 3D went on to make Doom, which ramped up the complexity of the 3D world. Rather than limiting the player to wandering in 4 fixed directions in a grid-like world, Doom provided the illusion of a true 3D world in which you could move in any direction at any angle. Doom also featured exploration-based puzzles involving locked doors and hidden switches. More controversially, it ramped up the violence level. It was another hit.

Before long, video game developers tried taking the kind of free-roaming 3D graphics popularized by Doom, and using them in story-based action-adventure games. The ultimate aim was to make a "cinematic" game; one that would feel like you were inside a movie.

By the mid 1990s, it was possible to model objects using polygons, and draw them at high speed on screen. This led to games in which both the world and the objects in it were truly three-dimensional. The launch of the Sony PlayStation boosted video game console power, enabling similar feats of programming in console games. But while game programmers could now draw and animate pretty much anything, the complexity of the resulting game worlds now became a major problem.

Next, in part 2: the complexity problem.

Waiter, there’s sex in my violence!

Let’s have a quick fact check here:

Grand Theft Auto: San Andreas is rated M by the ESRB, meaning it is rated as unsuitable for anyone under the age of 17. The rating label notes “Strong Sexual Content”.

The game features a mission where you rescue a prostitute who is being raped and murdered by two johns. It allows you to have sex with bikini-wearing prostitutes in your car, accompanied by bouncing suspension, steamed-up windows and explicit sound effects. You can also kill the prostitutes with a chainsaw afterwards, if you so wish. Or, you can wander into a strip club, watch the dancers on stage, then go into the back room and pay money for a private lap dance, depicted in 3D polygons.

None of that was a problem for the ESRB when they issued an M rating. However, if you modify the game you can make it show a scene where two adult characters have fully clothed, completely consensual sex, with no money changing hands, no exploitation, and no violence.

That’s apparently a horrendously damaging image, which means the game should have been rated for people 1 year older—because the difference between an AO rating and an M rating is that AO is for 18 and older, whereas M is for 17 and older.

As a sidenote, 17 year olds in America can have actual sex with real human beings in 43 states in the nation; I’m assuming they can legally do it with the lights on, perhaps even after taking their clothes off.

One of the previous installments of GTA featured a length lesbian S&M scene. Meanwhile, the (reportedly awful) game BMX XXX features topless women riding BMX bikes while their breasts jiggle, and full nudity; it’s rated M. No problems there.

So those are the facts. Now the House of Representatives has voted, almost unanimously, for the FTC to investigate whether Rockstar Games deceitfully misled the ESRB in order to corrupt America’s 17 year olds, rather than only those 18 and up.

Well, that’s money well spent. Thank goodness we don’t have any more important problems that need attention.