My Photo

« Theory - Meta Post | Main | Rich Client »

June 30, 2008



Models are only useful if what they expect is what happens, the problem is when something else happens. As stupid as this sounds, it is at the very heart of the fallacy of models. A model might make sense in an artistic way, but for any real application you cannot ever say that it is correct, only that it is useful.

Let's give a real life example. If you model your weather observations with a nice curve, and omit the extreme temperatures that occur a couple of unpredictable days every year, you will miss the melting of the polar ice caps. Your model might make sense for simplifying the observations and might work most of the time for predicting the next day's temperature; but we will eventually face the painful consequences of the poor-grounded confidence in it.

I like the example about the rocket, it nicely summarizes the misconception about models. Saying that without models we could not have sent a rocket to the moon, is forgetting about thousands of unsuccessful rocket launches that preceded the one successful mission. It is the ultimate in unfairness to the people who have spent their blood and tears to put the rocket into the sky; also oversimplification and forgetfulness that builds false confidence in models...

Models are beautiful, you can hang them on your wall like a landscape painting; but don't try to climb into the damn thing.

The comments to this entry are closed.

Twitter Updates

    follow me on Twitter

    March 2016

    Sun Mon Tue Wed Thu Fri Sat
        1 2 3 4 5
    6 7 8 9 10 11 12
    13 14 15 16 17 18 19
    20 21 22 23 24 25 26
    27 28 29 30 31    


    Blog powered by Typepad