A Discourse on Computer Role-Playing

By Murray K

In which the author examines the strengths and weaknesses of computer RPGs and how technology is working to eliminate these weaknesses.

What is a Computer Role-Playing Game?

To begin a discussion of Computer Role-Playing Games (or CRPGs), a definition is called for. In general, when you mention CRPGs, the first things that come to mind are gaming series like Might and Magic, Ultima, and the range of TSR products. If it has spells, classes and character attributes, it's called a role-playing game. But I'd like to widen the definition a little.

First, let me state what I don't consider to be a CRPG. The genre had been fairly quiet for a while, until last year, when a single game made CRPGs cool again. That game was Diablo. But while Diablo was a great game, it had all the role-playing elements of Space Invaders. The total conversation with the towns people could have been put on 2 foolscap pages and there would have been room left for the plot summary. It had classes, yes, it had magic spells, oh yes, but the basic premise was to keep clicking your sword on the monsters until your mouse wore out. I think Blizzard had a secret deal going with mouse manufacturers!

However, I do consider Quake II to be a role-playing game. And so, to a greater extent, are Jedi Knight and Unreal. All of these games are so firmly set in the first person shootem genre that they are practically archetypal, yet I believe they are also - in part - role-playing games. At the very least, far closer to it than Diablo.

Please explain?

You are probably wondering then what my definition of a CRPG is. Well, it's very simple: I believe a good CRPG is a computer game that allows you to play a role. It is a game that allows you to enter the role of a character and act like that character, responding to the stimuli around you in ways that would suit that character.

I put Quake or Unreal in amongst the role-playing games because they allow the player to become immersed in the world they are set in. Diablo was a fun game to play, but it never made me feel like I was there. And isn't that what role-playing, in any form, is about? About being able to assume a different persona from your own and enter situations that are different from your everyday experience. And this is what these games provide.

A game like Quake has a definite advantage in creating a believable world for the player in that it is set in a 3D world you can see and move through in the first person. This gives it an edge over a game like Diablo, which uses a third person perspective. When role-playing, few people would imagine watching their characters from afar. However, convincing visuals are by no means the only or even the best way of transporting a person to another reality. The most obvious example of this is in table-top role-playing. You cannot see the woods or the tavern, but with imagination, it is possible to believe that you are there. Imagination, that is, coupled with that most important quantity in role-playing, a talented Game Master.

The Computer as a GM

The Game Master's chief purpose is to create the world in which the player lives and the other characters with which the player interacts. Secondary to their role is the maintaining of the rules and being the final umpire in arguments. However, no matter what the system, or how brilliant the rules, the game will not be interesting to the players if the GM cannot breathe life into his world. Likewise, the aim of computer role-playing games is to produce a world and people that the player can believe in and interact with. In short, the computer must take the role of Game Master. How successful are they in this task?

Where Computers Succeed

Computers excel in several areas. I have already mentioned one area in which computers are becoming more and more impressive: the creation of realistic three-dimensional worlds. At the moment, these are generally used in first-person shooters. However, before Doom even hit the shelves, they were used to excellent effect in some great role-playing games, like Ultima: Underworld or Eye of the Beholder. Since then, other attempts at adding 3-D visuals to a role-playing game have been made with mixed success. One from last year, Daggerfall, was a very ambitious attempt to deliver a full and expansive world (supposedly four times the size of England!), and to some extent it succeeded. The danger with these sort of games, however, is that they will have great eye candy and shiny new features at the expense of real game-play. Slowly, as 3D engines become more common-place and games have had to start distinguishing themselves beyond pretty graphics, better game-play is appearing.

Another area in which computers exceed human capabilities is memory. For starters, the computer knows all the rules (which can be quite a lot in, say, AD&D) off by heart. In terms of role-playing, however, this is a double edged sword. It means that you can be certain that your opposition is not cheating and it will normally keep you from cheating. However, it also limits the use of artistic flair and creativity. The computer can't ignore a roll to achieve a more interesting story.

The main use for the extensive memory of the computer is the ability to keep up with the many characters and events in a world. It is now possible to give NPCs their own lives. Just by giving each character a simple set of routines and reactions, it is possible to create a realistic community. While most GMs maintain a few main characters, it is clearly impossible to keep track of every NPC in the world. While obviously unnecessary (as appropriate NPC's can be created whenever needed) it does add to a game if there is enough scope for this sort of player interaction. The idea of completely independent NPCs has only recently become a feature in computer role-playing games, but is becoming more common.

It is also possible to maintain a completely consistent economy, in which trade, shortages and surpluses can occur. These effects have been modeled with varying degrees of success in games such as Daggerfall, the Ultima series, Fallout and even some not-so-role-playing games such as Wing Commander: Privateer. Once again, this is not essential, but does help create the feeling of a world going on in the background, independent of the player.

This is only possible because the computer can be the ultimate encyclopaedia, the GM who has 2 scribbled pages devoted to every one of the 10,000 citizens of the town, knows the names of all the local cats, and always remembers the exact cost of a dozen eggs(adjusted for the fluctuating market).

Where Computers Fail

So, at the moment, it is technically possible to produce a world rich with detail that can actually be seen, and a world that can feel alive and ever-changing without ever being inconsistent. It's about this time you want to reach out and touch something. That's when the problems start.

Many times in role-playing games, I have done spur of the moment things like throwing a rock or putting a board across a door. The trouble with most computer games is that unless an item has been put there specifically so you can do something with it, you probably won't be able to touch it. That means no grabbing the bed sheets and forming a rope out the window with them, or jumping onto the back of the ancient red dragon and hanging on for dear life. The human imagination is nearly limitless in the possible combinations of events it can bring together. The computer, on the other hand, tends to be limited by what programmers teach it, and they don't have time to teach it much.

As well as interaction with the environment, there is the one major hurdle that must be overcome to truly immerse the player in another world, and it is the hardest of all: NPC interaction. The computer may know all there is to know about an NPC, down to the amount of sleep it got last night, but the computer doesn't seem to be any good at making them cranky when they are tired.

A first person shooter doesn't have to worry about this because the only interaction you require is shooting and being shot. But in any "traditional" role-playing game, you need to be able to talk to the other people in your world. A good GM is able to invent conversations out of thin air, but most computer games rely on pre-scripted conversations, both what you can say and what the replies will be. It takes away almost all thought and all immersion to find yourself deciding between 3 conversation topics, especially when you know you are going to just go through all of them in any case, trying to find all the information you can.

Further Progress

For a computer to be able to produce an immersive experience on par with good table-top role-playing, it needs to overcome these problems with interactivity. The good news is that slowly but surely, these are being overcome. The really good news is, most the elements are almost here.

Total interactivity with the environment is the easiest of the problems to overcome. Basically, manipulation of objects involves physics, and physics is something that computers are good at. Teach a computer how a hinge works once and you will be able to put hinges anywhere you want. Then you can teach it how a door works. Then, with a little bit more work, you can teach it how a door can be blocked, and suddenly I can put a board in front of a door just like I would in table-top role-playing, or in real life.

This stuff is already being done in several games that are soon to come out. A couple of notable examples of this are Trespasser and Half-life. Trespasser claims to have a complete physics model for every object in the game. That means that if you want to get over a wall, you can knock a hole in it, stack up boards next to it or try hooking a rope onto something and climbing over it. This is the sort of level of interactivity that a computer role-playing game needs.

Harder than that is interaction with NPC's. Meaningful conversation can be very difficult to create on a computer. The original text based computer games were based around conversation to a large extent, but these had a limited number of possible paths for the conversation to go through. Basically, you typed something and if the parser saw a keyword, it reacted to it, otherwise it said something like "I don't understand". Either that, or it threw back your own words at you, like the computer psychiatrist ELIZA.

There are three possible ways around this. The first is to only let the player say certain things. However, as mentioned above, this can lead to players simply running through all the responses, and this can inhibit role-playing terribly. But if these conversations are well written, such games can be quite good. As long as the player is not too limited in his options and the topics make sense with the situation, this method helps the narrative and doesn't detract too much from the immersion.

The second response is to take the text parsers used by the old games and increase the number of keywords they can recognize. Memory space is, of course, finite, and every possible phrase cannot be catered for, but as long as the player stays within the conventions of the game, this can work well. The recent game Starship Titanic claims to have one of the most advanced language parsers and a very extensive conversation tree. In playing it, I quite enjoyed being able to type in whatever I wanted, and it lent itself to a quite good role-playing experience. At one point the characters actually seemed to understand sarcasm. There are still, however, the variations on "I don't understand", which can get quite annoying after you have heard them a dozen times.

The third way of creating interacting NPCs is teach the computer how to recognise concepts, and how to construct natural language sentences using those concepts. This is something that is at present restricted to research houses in large companies. However, five years ago, the computer graphics I can now cook up on my machine were considered state of the art and the software was only found in research labs. Eventually, I believe that some games will include concept based parsers and constructors, and that these will be used as the brains of NPCs. It seems a long way off, but I think we will start to see it in the next decade.

Future Dreams

I think that eventually, the 3D role-playing genre and the top-down role-playing genre will merge. It is becoming increasingly common to have 3D games that incorporate camera angles, and thus soon the RPG player will be able to view the action from the position they prefer. Hopefully, this merging will also lead to a merging of the best characteristics of both, i.e. the increasingly absorbing plots and character interaction of the top-down CRPGs and the increased environmental interaction of the 3D CRPGs.

I personally don't see any end to the levels of immersion possible as long as technology and game-design continue to advance incrementally. The only true barrier is going to be the amount of money that can be invested in laying background information, such as the terrain, environment and NPCs personalities. Eventually, it will not be cost-effective to produce a more realistic world to sell in a computer game. However, programmers being what they are, you can bet that it will be a long time before that happens.

But what I am really waiting for is the day they build a holosuite.

What did you think of this article? How useful was it? How interesting? Let us know!