Friday, March 02, 2007

Civilization and Game Theory

This article in the Weekly Standard (via the American Scene) makes the important point that generalising from Grand Theft Auto to ‘computer games are bad for the young’ does not allow room for Civilization, 'a video game version of a classical education'.

When I first studied game theory it struck me that Civilization is probably the most complex (in terms of number of potential strategies) artificial challenge mankind has ever set itself. The number of possible strategies in chess is truly staggering, more than all the molecules in the universe, but for Civilization, with so many choices as described in the article, the number has to be higher. I doubt there's a game with more going on and, in that regard, Civilization has to be considered something of a cultural high watermark. It would also be interesting to see how players respond to a game they are unable to truly think through due to its complexity (just as in chess where players and computers can only think so many moves ahead); how do people guess their way towards effective or ineffective strategies?

It might be particularly interesting because it seems far more plausible that the sentimentality which must infest real life decisions will be mirrored in the imitation of real political choices that is Civilization than in the abstract game that is chess. There's a PhD in that one I think...


Jackart said...

I had to send my CIVII CD home. It was either a degree or CIV. The two were mutally exclusive

arthur said...

As a person who has sacrificed his degree for Civilization, I must disagree. Once you get the hang of the game, it is far less game theoretic than you think it is. There are lots of things you can tweak, but in reality the actual set of strategies are quite small.

The dominant opening strategy in Civs 1 – 4 is to expand till you hit a border. After the expansionary phase, you have the choice of rush attacking an opponent or not doing so. After that, it is mostly an optimisation problem: the choice is between 1) gaining a lead in technology, and overtake everyone in production/military tech/culture later on, and 2) building as many attacking units as possible and invade your neighbours until you fall behind in tech. How best to accomplish 1) and 2) is a matter of knowing the “math” behind the game, and has rather little to do with what your opponents are doing.

The only time when you have to react to opposing strategies, I think, is 1) when you are deciding whether to race for a wonder or a religion, 2) someone else has a military resource and you don’t (esp. uranium), 3) one of your neighbours are culture bombing you. All of these call for a response, but they are mostly detours from the 2 main strategies outlined above.

ian said...

I tend to agree with Arthur, although I would go further: you could argue that the Civilisation series (I've played I and II a lot, III a little, and also have a degree to write, so haven't dared buy IV...) in fact promotes a particular ideology, in that Democracy is held up to be the ultimate arrangement for your civilisation: it is the government type that enjoys highest production levels and zero corruption. I've had little success persisting with Communism, for example. There seems to be no attempt at a reasonable simulation of alternative modern political systems,making them viable in terms of game play.