Mental models and predictions

So you’re reading this blog by a small furry animal with a secret laboratory, and you’re wondering, ‘what does this weasel know that I don’t know?’  Well, here’s something I know: how to predict the future.  I’ll introduce the topic here and drop some more hints in future blog posts.

Everything I need to know I learned in … kindergarten? … no, classes on stochastic estimation and control.  The future of a system is easily predicted if we have a good system model that reasonably well describes the system to be predicted, as well as measurements (with some inaccuracies or noise in them) of the current state of the system.  A proper system model includes some estimate of how wrong the model is, and how wrong the measurements are.  See ‘the map is not the territory’ in this article about mental models.  All models are wrong (to some degree); some models are useful.

A filter is a mathematical tool that combines model predictions and measurements to see where things are going for the modeled system, a few to several measurement cycles into the future.  Great measurements combined with a bad model will lead to bad predictions, while a good model can make sense out of bad measurements and still make reasonable predictions.  This is about control systems, but also about life in general.  If you understand how the world works, you can make predictions that turn out pretty well most of the time.  If you have a bad model, you will continually be surprised by events.

So how to apply this: read an article that makes a prediction about the near future.  Has the prediction had time to come true yet?  Did it turn out true or false?  Keep that in mind as you assess other articles.  Is there a person out there who makes predictions that are reliably wrong?  Bad model.  Stop listening to that guy; he will probably never get it.  Is there a particular set of beliefs that lead to bad models?  Stop listening to people who think that way.  All of them have bad models of reality, and probably always will.  Put more trust in people whose models of reality allow them to get some predictions right.

Is there a magazine or web site with a bunch of articles that make predictions?  Are the predictions similar, or is a diversity of views presented?  How did the predictions turn out?  That’s a guide to whether you keep paying any attention to that media source.  If they didn’t predict the past, they can’t predict the future.

Is there a political party that makes predictions?  Has that party enacted a policy?  Has the policy turned out like it was predicted?  “Our change to the tax code will help the economy” is a prediction.  Did it work out?  A politician in the US thinks the new leader of Country X will make things better/worse there.  Years later, are people now trying to move to Country X or desperately fleeing it?

Is there a public policy question that has some relation to science?  Are there multiple sides presenting different interpretations of what will happen in the future?  They are making predictions.  Have their past predictions actually turned out?  If so, then pay more attention.  They may have a good model, and if so, their predictions will continue to be valuable.  If not, move along.



3 thoughts on “Mental models and predictions

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s