All In The Game, In Theory

There's a fascinating article in the New York Times about Bruce Bueno de Mesquita, a game theorist who models the way networks of influence behave, and predicts the outcome of complex situations 'where parties can be described as trying to persuade or coerce one another'.

His work leads him to make some bold and far-reaching predictions about the outcome of currently unresolved political conflicts and standoffs. Some are optimistic (Iran will not make a nuclear bomb), others less so (most nations will welsh on the commitments they make to reduce carbon emissions at Copenhagen).

Bueno de Mesquita's track record in making successful predictions appears to be rather better than, say, Nostrodamus'. So why isn't he better known and continually in demand by corporates and politicians the world over? Bueno de Mesquita himself puts it down to his being a lousy salesman. But the Times also suggests that there's significant resistance to the idea that a mathematical model - however sophisticated - can take the place of intuition, inside information and gut instinct - the stock-in-trade of many political analysts when making predictions.

At GlobeScan we use statistical models of a different (and rather simpler) kind in our corporate reputation work, and this resistance is something I've noticed on more than a few occasions when presenting to clients. If we're looking at patterns in survey data to identify how issues cluster together in the minds of stakeholders, or using regressions to identify what factors really do drive an organisation's reputation, there's often a sceptic or two who will question whether we can really believe what the modelling is telling us. 'Is the way we're seen to treat our employees or our environmental track record really more important than our ads in driving our reputation?' Valid questions, for sure - but like Bruce Bueno de Mesquita, we've seen enough over the years to have confidence in our models.

I hope he's wrong about Copenhagen, though.