Quote Originally Posted by William F. Owen View Post
The idea that War or conflict might conform to an elegant model is pretty much a nail in the coffin of the concept. War simply does not work that way. What if he included data from WW1, or the Crimea? The data he uses is from 21st century insurgencies and civil wars - which are characterised by single murders and shootings. The whole thing fails the "so what" test.
I hope these guys put their research up for scrutiny soon. Spagat's presentation hints that they vary their conflict samples so widely that its a meta-analysis of anything from today's wars to Japanese invasion of Manchuria to Chicago Vice Lords beefing with the flavor of the month.

I don't expect war to conform to an elegant model, at least not a useful one. I'm not even sure individual classes of war we can identify would yield such a result. Almost all social science models are highly conditional, and if any overarching model exists--I seriously doubt it boils down to something as analytic as a power law.

I bet you'll find that the number of people killed per domestic hand-gun attacks is 0.75 world wide and has been for 100 years. - so what?
Gourley claims that the scaling constant (alpha in this case) should be derived from another set of tools in group dynamics, and that if things are left to themselves this should stabilize at around 2.59. Conceding truth to his claims regarding a mathematical pattern to warfare, I'd be more skeptical of Gourley's claims that in group dynamics exist methods methods to engineer alpha to achieve a desired result. Gourley himself is skeptical, as this is where his "I don't knows" crescendo.

- see my signature.
Well said.