Simulation: the weblog


That US oil shock simulation again...

An earlier posting on this blog referred to a simulation of the effects of a major oil price shock, conducted by the US National Commission on Energy Policy and "Securing America's Future". I expressed some doubts at the time about the validity of the simulation.

However, there's no denying the impact of something like this if you pitch it at the right level. Recently, it has been quoted by the Irish news service Finfacts and the London Financial Times, as well as several blogs

Simulations have the ability to consider a wide range of inputs and a wide range of relationships between each input factor, and to produce a simple response to a specific question. That is why they are so powerful, so persuasive, and so news-worthy. They appear to offer answers to the big questions that worry all of us.

But the thesis of this blog is that simulations can be 'wrong' as well as 'right'. Indeed, it is misleading to think in these terms. They are not 'predictions'. (See earlier entries on this blog, about work by Gene Bellinger and Michael Drennan. We all have a basic desire to forestall the future. Simulations don't do that, but it's tempting to think that they do.

According to the FT, the president of Securing America’s Future Energy, a group which advocates reducing US dependence on oil, said: “There is nothing like watching, listening and learning as a group of former cabinet members and senior government officials sit in a ‘mock’ situation room responding in real time to a series of plausible and credible events. This is hopefully something that all champions of this issue can use to build support for serious action.”

Call me naive, but I just don't like simulations being used in this way. They should, as far as possible, be objective tools for developing our understanding, and not for channeling it, championing issues, or building support.

What would your attitude be if a scientist funded by the tobacco industry produced an experiment which demonstrated that, under certain conditions, smoking was good for your health?


Powered by pMachine