When virtual worlds feel more alive

Or why complexity science is relevant to gamers

In 1984 The Santa Fe Institute, then a nascent, interdisciplinary research body, held one of its first workshops. “Emergent Syntheses in Science” was an event which gathered together scientists, mathematicians, and philosophers. They were interested in one thing: emergent behaviour. Why, they wondered, is the collective behaviour of a bee hive not easy to understand by just looking at a single bee?

Emergent behaviour is everywhere. Bees making honey, ants colonising, diseases spreading, people forming neighbourhoods, the assembly of a cancer cell, an embryo forming in the womb. These are the self-organising systems we see everywhere around us. Forming, spreading, determining outcomes in the world; these are systems with no leaders, hierarchical structures or central control. These are dynamic systems which adapt in and evolve with their surrounding environments. Neither are they static; this behaviour is always surprising.

So what does this mean for game designers?

This could seem like a very distant discussion from the world of games. But, in fact, emergent behaviour is the stuff of stories. Because while this has huge implications for our understanding of the world around us, it also confirms a very personal, human instinct: we don’t always get to control what happens next. Within our day to day lives, actions are not scripted, outcomes are unpredictable, behaviour emerges.

Within our day to day lives, actions are not scripted, outcomes are unpredictable, behaviour emerges.

You couldn’t have predicted the things that happened to you today. You might have known you would read an article today, but you probably didn’t know that it would be this one. And this is, arguably, what makes life engaging, what makes it immersive. This behaviour is not predetermined, it is based on rules. Physics, weather patterns, social norms. But combined, and interacting with the myriad of other systems and variables, it creates complex behaviour. Replacing scripted design for simulated- based design has created a very new kind of game experience. Minecraft, Half-Life, Second Life, these are all games which have transformed the player’s experience. How does it work?

Let’s look at some birds.

Let’s take a deeper look at this idea and start with a very simple flock of birds. This is actually a classic study used by complexity scientists.

Here we have simulated thousands of birds, each performing a role within a much larger body. They have goals: to find a nest, find food, find their next destination to which they are migrating. Even though there’s no leader, they move in patterns that are clearly not random. Let’s digitally recreate this real-world behaviour, using a simulation based on simple rules.

Rule 1: All birds must travel at the same speed.

Rule 1: All birds must fly at the same speed.

Clearly, they are not flocking as we know they would in the wild. But each and every bird here is flying at the same speed.

Rule 2: All birds must stay close to their neighbour.

Rule 2: All birds must stay close to their neighbour.

Now we are a little closer to the shape of a flock. But we can see immediately that these birds have no concept of danger. They would not move out of a predator’s way or even each other’s; here we can see that they are crashing into each other. We clearly need to add in another rule…

Rule 3: All birds must avoid danger

Rule 3: All birds must avoid danger

Finally, we can see that there is a rule that the birds crash into each other or, indeed, other external parts of their environment.

Ta-da! Now the birds not only move at the same speed, and in the same direction, they now also form patterns as we see in nature; we have simulated the movement of birds flocking. Because this hasn’t been faked (or scripted) the birds would respond to their environment based on those rules; just as they would in the wild. This means that how you could interact with them would feel more real; you can’t fully predict their behaviour in an environment where multiple variables could affect their movement in a variety of ways. Simple rules, such as we showed in the flocking example, gives rise to complex behaviour.

Simulation based design: the key to a more engaging world?

This concept is at the basis of any simulation based design. Minecraft and other survival sandbox games are examples of how engaging these worlds can be to play in. But these all constrained by the size of their servers and so the dynamic nature of the things you can interact with are limited by the compute power they use. The idea of creating entire complex ecosystems within the worlds has only just begun. What if we could build worlds with hundreds of complex systems all interlocking and shaping how the player reacts? The world would have its own internal logic, its own expectations. But every time you play in this world you would experience something different, something surprising, something unexpected.

To see how you could integrate a flocking worker in your own game, check out our blog How to Create a Custom Flocking Worker in SpatialOS.