Predicting the future: who pays wins?
October 16, 2007
Good Magazine‘s article on The New Nostradamus, Bruce Bueno De Mesquita, and rational choice theory is deeply thought-provoking. He claims to be able to model and accurately predict the outcome of conflicts. Certainly the successes and success rates quoted are impressive. But the article leaves untouched some serious issues behind the control and application of the modelling process.
past clients include Union Carbide, which needed a little help in structuring its defense after its 1984 chemical-plant disaster in Bhopal, India, claimed the lives of an estimated 22,000 people
Whilst this is somehow impressive – and the understatement in “a little help” is also impressive, there’s no indication of how company and community needs were balanced. Union Carbide (presumably Dow after the takeover) still maintains the Bhopal Information Center as their “[…] statement regarding Bhopal Tragedy in addition to historical and legal information about the incident”, and to make sure their side is heard and a line drawn under the matter; remarkably, this appears as a Sponsored Link on Google! Talking of balance, there’s also the International Campaign for Justice in Bhopal (ICJB), which suggests that all is not as it might be. Still, however, no mention of the impact of the model.
But there are limits to what his company will do. […] “We have a corporate policy that we will not, on a commercial basis, use the model in campaigns,” he says. “We don’t think it’s appropriate to manipulate the democratic process. We won’t take a client who wants to manipulate U.S. government policy, even if we agree with the manipulation. And we won’t take a foreign client whose objectives are contrary to the objectives of the United States government.”
- So the model might be used on a non-commercial basis? As, say,
- a favour?
- within his company to predict where that company should position itself?
- on behalf of the (US) government to determine policy?
- on behalf of the (US) government to determine action, such as when to call an election?
- There appears to be protectionism at work: we won’t try to influence domestic (US) government, but foreign governments are fair game.
- Would they take a client whose objectives are contrary to the objectives of a business which operates with (US) government blessing?
- Is it acceptable for a commercial organization to influence the public using this model?
This is not to claim that these possibilities are unreasonable in themselves, just to ask to what extent there should there be debate about them: in what context is the model applied?
A comparison of Bueno de Mesquita’s work with Isaac Asimov’s fictional science of psychohistory, introduced in his Foundation book series, is inevitable, but, at least to date, I couldn’t find anything approaching a close look at the connections. Asimov’s view was that “[… he] assumed that the time would come when there would be a science in which things could be predicted on a probabilistic or statistical basis.” (source)Psychohistory was propounded, within the context of the books, as having two underlying assumptions:
- That the population whose behaviour was modeled should be sufficiently large
- They should remain in ignorance of the results of the application of psychohistorical analyses.
On the first point the article effectively appears to say that Bueno de Mesquita’s model can work with individuals’ decisions. This makes it an interesting counterpoint to the Asimov model. Perhaps both are possible, although over different timescales?
The second is, however, of great significance. Is the population at large to know what the model predicts? If it does, can it react? Points about the impossibility of predestination working, if the subject knows about the prediction, have been made elsewhere (which I might come back to if I remember), as the subject can falsify the prediction, including predictions of elections when they are made public.
A more general theoretic issue is, What happens when another party is using a similar (or the same) model? Will they come up with the same solutions? Or will an additional layer of complexity be added?
On a minor note, describing Bueno de Mesquita as the “new Nostradamus” doesn’t do him any favours, beyond providing him with a newsy article title. Nostradamus’ writings aren’t especially well-understood, but they are widely misused.
Nevertheless, the model does seem to offer some excellent win-win solutions to some very difficult – and indeed dangerous – real-world situations, so it will be used. However, given that it does not predict perfectly, what is the danger that a decision will be taken with catastrophic results?The model is like fire: it isn’t the concept of fire that burns, it’s the way it’s used, whether deliberately or not.Thanks to a colleage for passing the article on.