Guest posting: game theory and Iran

Hello again to missmarketcrash's readers. I have the privilege once again of making the occasional contribution while she's on holiday.

Today, in honour of missmarketcrash's Rhode Island resort, I was very interested to read this article from her near-neighbour the New York Times.

My own work is about using behavioural economics to understand, predict and perhaps influence some aspects of decision making.

So it was fascinating to learn about Bruce Bueno de Mesquita (must remember not to compress this to 'Bruno' which is what I nearly typed) who does something similar but based fully on rational actor theory. He uses game theory techniques and a software system to project how power plays and coalition building will result in political or business outcomes.

Now lots of people are sceptical about the effectiveness of this kind of prediction. It's clear that the results are highly dependent on the input parameters - in the Iranian example, he suggests that Ahmedinejad's "desire" is 90, "power" is 5 and "resolve" is 90. To be honest this sounds like nothing quite so much as the strength, stamina, intelligence and charisma scores in Dungeons and Dragons, but you do have to start with some numbers if you're going to build any kind of model.

The article may be oversimplifying his approach, but these figures seem just to be assumptions not based on any empirical measurements. The proof, in any case, will be in the pudding, but it's a little slippery to get hold of any testable predictions because lots of it is covered by commercial confidentiality - or worse, national security. However:

...a study by the C.I.A., now declassified, found that Bueno de Mesquita’s predictions "hit the bull's-eye" twice as often as its own analysts did.

Anyway my interest is piqued because, despite the suggestion that this is based on a rational agent model, it appears to use some very psychological variables which aren't incorporated into standard game theory: preferred outcome, desire and resolve. I suspect that behavioural game theory might be a better description of this.

I will be following Mr Bueno de Mesquita's work with interest - we will have an interim result by October as he has predicted a student uprising in Iran.