Originally Posted by
Presley Cannady
Whether or not alpha represents the structure of the insurgency in any way remains to be seen (and I can't even find the results of this research online, so it may be premature to say the least). It's possible this "theory" why alpha is stable around some point (2.5) may be no more than one working hypothesis amongst many. Gourley is coupling another unstated assumption, group dynamics, to this model in order to explain alpha and his talk completely skimped over that point.
The power law "works" because we intuit the probability of destructive events occurring decreasing with their destructiveness (nuclear terrorism on a Western target is harder and costlier than setting an IED in your own backyard). Everything else is a question of how it behaves in a scaling limit. The fact that it behaves like a power law at all means that it can only not be a power law asymptotically--some other term overwhelms the scaling exponent and breaks invariance--or the function is piece wise with the addition of more data.
Bookmarks