For a very long time I worked in multi-agent modelling. This is a field in which a large number of agents are given very simple instructions and the pattern that emerges from their individual behaviour does something that would be beyond the scope of any of the individuals in the pattern. This sort of model can explain a great many animal behaviours, which look intricate but can be modelled using simple rules.
This was useful in investigating human behaviour since it allowed the study of the influence of individual level behaviours on group properties. One of my favourite papers in the field, Accidental Atheists, looks at the distribution of religious belief in America. It's well known that some regions, particularly the "bible belt" have very dominant religious influences while others have a high proportion of atheists. It's not so clear why this persists, Americans have been moving around within the states for decades, so surely everyone would be all mixed together by now? One explanation is that people like to be surrounded by others of the same faith and so move to live with them, but this explanation is flawed as people's reasons for moving are fairly well studied and religion is a generally unimportant variable. The MARS model simulated an America in which people move more or less randomly, but sometimes change belief to conform to the local dominant religion, which using a few simple rules reproduced this pattern of religious regionalism. While there are flaws with this sorts of research I think it's got some really interesting implications. If nothing else, it's good for winding up atheists who insist their position comes from intellectual superiority by pointing out that their belief is just an accident of geography.
My point is that emergent effects are really cool and you can get all sorts of things coming out of quite simple systems. My own research had agents that would punish other agents for theft pop out of a simple "eat enough to live and breed" simulation. Games have these properties in spades. In general a game is ruled over by a very limited set of rules, but a set of properties emerges from them.
In Go, a stone is taken when it has no liberties (adjacent empty spaces). If two stones of the same colour are adjacent to each other they form one group, which can only be taken once the group as a whole has no liberties. It is not legal to play a stone into a location that has no liberties, unless doing so would remove some opposing stones thereby creating some. From this it emerges that if a group completely surrounds an empty space your opponent will only be able to play there is if they have removed all of the other liberties that the group has, by surrounding it. Furthermore it follows that if a group can completely surround two empty spaces (called having two eyes) you cannot remove it from the board, as you cannot legally play in one of the spaces without first having played in the other.
White can't take the black stones, because they have two spaces providing liberties (marked with Xs) and can't play into those spaces because the newly placed stone would have no liberties and wouldn't take anything. From this, a whole host of other properties emerge, for instance if the middle black stone was missing, white could play there to kill the group. Black could only remove the white stone by surrounding it, which would involve playing in both spaces marked X, leaving it with only one eye. That spot would become a vital point, if black played there first (as they have in the image) the group lives, if white plays there first it dies. The rules do not codify any of these things, the rule is simply that you must surround a stone to take it and can't play a stone where it's surrounded unless you take something in doing so. All of those other things emerge.
Emergent play can also occur because a game has many different elements that react unexpectedly. Magic is fantastic for this, though I suppose it's a benefit of the CCG genre in general. When Channel was first published it was thought to be a weak card, it offered the ability to deal yourself damage in exchange for mana, since mana was easier to come by than life in most situations. Then someone combined it with Fireball, a card which used mana to deal damage to your opponent on a one for one basis. Suddenly when the two cards were played together you could spend your life to damage your opponent's life, you only had to generate a one point life advantage and you were set to win the game! Over the years a great many combinations and tricks have been found with various cards, one of the things that keeps people coming back to the game is seeing new cards and coming up with ways to make awesome things emerge when they are combined. One of the reasons to seek out new opponents is to see what sort of situations emerge when your deck faces something new.
I try for emergent properties in my games. Wizard Academy is built on the notion that the spells would do quite simplistic things like "place a fire token on your space" and the things would act in quite simplistic ways like "when a fire token is activated place another fire token on a space adjacent to it" and that the combination of these things would make the game go. Space, Monkeys and Cannibalism goes the same way, with actions being straightforward but the combinations of them producing results. The human behaviour "load the highest damage available object into the torpedo tube" wasn't specifically designed so that if two humans end up in the weapons room and run out of ammo they start fighting to load each other first, but for a lighthearted game it was an event that provoked a fair amount of mirth and lead to a few new strategic options.
Which games do you think have excellent emergent properties?
A collection of posts by game designer Gregory Carslaw, including mirrors of all of his blogs maintained for particular projects. A complete index of posts can be found here: https://boardgamegeek.com/blogpost/58777/index
19 Jun 2013
- [+] Dice rolls