Emergence and Simulation

For some reason when people talk about emergence in games other people get scared. I don't really understand it. I think people have some ill conceived notion that emergence involves tons of simulation and provides at best unpredictable results. It is hard to tune a game that you can't predict. I, of course, have a different view on things. I believe emergence in games is a beautiful thing and that it elevates a game to something more than the sum of its parts.

So, what do I mean by emergence - what is this magical mystical essence that evolves games beyond the ordinary? That is a tough question to answer succinctly. I think the essence of it is that complex behaviors emerge out of simpler ones. In games, this is mostly applied to solving problems. In fact, talking in terms of problems instead of puzzles begins to put some perspective on emergence. Puzzles only have one correct answer while problems can be solved in various ways. The essence of emergence is to allow players to solve problems in ways that make intuitive sense to them. This mean, first off, that players understand the game world. They understand the rules of this world and the effects of their actions. This effects might not always be predictable but that should be built into the rule. It is fine if the player's wand has some random effects on toads as long as the player knows when he points at a frog and uses his wand that it is likely to be something random. It is fine for the player to have to discover rules on his own as long as they are consistent. Consistency helps the player fully realize the game world and his place in it. It helps him learn the rules of the game world and how he can affect things and how he can solve problems. So, I think that emergence should come from simple interconnected consistent rules and systems.

So, this leads me to simulation. A lot of advocates of emergence also advocate heavy simulation. They believe it provides more opportunity for emergence. I get it. It does, because it makes the rules more believable. The problem is with heavy simulation you usually also involve heavy unpredictability. And this becomes a nightmare when you want to tune, debug and polish your game. Unpredictability can easily feel sloppy and that isn't something you really want your player to feel. I advocate a more relaxed method of developing systems. Develop the rules of your system first. For example, if you are building a fire system, enumerate all the things fire can do and all the things that can affect fire. It isn't super important how these rules are implemented but more so that they are implemented to the effect that they accomplish the rule and that it is consistent. I advocate simple rules that have complex interactions. In other words rules that are connected to objects in a complex web. I view these rules are connections between objects. For example, Water douses Fire is a rule that connects Water with Fire. The more connections you have to and from Water and Fire the more useful these objects become and the more ways that behaviors can emerge in new ways.

Simple interconnected consistent rules.

I could talk a lot more about this and in fact probably will at some point, maybe I'll include some of the charts I have done... that should be exciting.