I agree that the Rand data set is not at all bad, within obvious limitations. What I'm questioning is the applicability of conclusions drawn from that data set to a conflict that lies largely outside the parameters of that data set.

Treating insurgency as a subset of civil war muddies the issue even further, because whatever the conflict in Afghanistan is, it is definitely not civil war.

I'm not comfortable with calling this fight an insurgency or a counterinsurgency because I think that designation conditions us to see the core of the fight as the insurgents seeking to overthrow a constituted Government and the US intervening in support of that Government. I don't think that's an accurate perception at all. To me the core of the fight is our perceived need to determine who rules Afghanistan and to impose certain conditions on the way Afghanistan is ruled. I'm not saying that's right or wrong, but I don't see how we can accomplish a purpose if we aren't clear with ourselves about what that purpose is.