I thought this article was very good, but left out the very important point about decreasing crime rates.

http://www.vox.com/2014/8/14/6003239/police-militarization-in-ferguson

Also, this: http://chn.ge/1rNg8DL

Mason adored your flow charts.

]]>I am Roya Farhadi, PhD scholar of Tehran university of entomology. My thesis is about prey and predator. I would like to develop a vensim model. As I surfed in net, I found a vensim model figured in your belog (http://metasd.com/wp-content/uploads/2013/04/Structure.png).

I would like to know how has developed this model and how can I contact her or him?

I would appreciate you for the time you spend for me.

Best regards,

Roya Farhadi

In engineering, it works: I’ve seen some observer Kalman (OKID) results that were almost miraculous, in the context of vibrations in complex structures.

I think there’s much to be learned from big data about the economy, but I’m not sure it will work the same way.

The first problem is the other one you mention, that it’s not even clear what the outcome is. More generally, lots of variables in the economic state space are unmeasured, or what’s actually being measured is poorly understood. There’s lots of noise, and in many cases strong incentives to lie and manipulate or hide data.

The second problem is that if you do manage to succeed with a non-operational model, it may not help you to redesign the system, or if it does, no one will believe you (a problem that we already have in spades).

Still, there’s much to think about here … econophysics may be the first sign that this works to some extent.

]]>*** I think that most economists would agree with you; but this still relies on the assumption that total output (or even per-capita), is still the outcome that we “should” be measuring and simulating.

Second; this is still a structural model even with those “updates” to emulate a more realistic model… shouldn’t we be looking at doing non-parametric models to help explain the phenomenon we’re interested in? Let the data reveal what the “model” actually is?

]]>