Ok, we have some varied ways of looking at this picture. We have Bob's World's three types of insurgency, we have Wilf's regular/irregular, which seems to focus on who is fighting, and Marc's conventional/unconventional, which seems to revolve more around how the fighting is done.

Now back to the Rand study. Three excerpts from the introduction:

Insurgencies have dominated the focus of the U.S. military for the past seven years, but they have a much longer history than that and are likely to figure prominently in future U.S. military operations. Thus, the general characteristics of insurgencies and, more important, how they end are of great interest to U.S. policymakers. This study constitutes the unclassified portion of a two-part study that examines insurgencies in great detail.
“How Insurgencies End” has produced several findings, some of which reinforce or explain conventional wisdom regarding insurgency and COIN. Others present a new range of dilemmas and opportunities to policymakers and planners. We derived additional findings primarily from the quantitative research. These findings reveal some useful insights into the relative success or failure of various methods employed by each side as they apply to insurgency endings. A few of these additional findings describe the impact of existing operational and environmental factors on COIN operations, thereby informing policy decisionmaking.
Recent U.S. experience in COIN has been especially tangled. Vietnam speaks for itself, as do Iraq and Afghanistan.
What I find fascinating and a bit confusing about all this is that nowhere that I can see in the document is there any attempt to define what an insurgency is. From the excerpts above it seems clear that the authors believe that the fighting in Afghanistan and Iraq are insurgencies and that the conclusions of the study have relevance to those conflicts. In the body of the study, however, insurgency is treated as a conflict between a government and its populace or some part or parts of that populace, hence the categorization of "government win" or "insurgent win". I'm not sure how relevant any of this is to a situation in which some part of the populace is fighting against an occupying invader, and the government is perceived as a largely extraneous subset or representative of the occupying power.

If we're going to survey insurgencies and apply the conclusions of the survey to the conflict in Afghanistan, we have to do two things at a bare minimum. We have to define what insurgency is, because until we do any discussion of insurgency is going to be too nebulous to be useful. Then we have to decide whether the conflict in Afghanistan is in fact insurgency.

The distinction seems relevant to me because I see a trend in American discourse toward treating the conflict in Afghanistan as a case where we are defending an allied government against insurgency, hence the "are we doing COIN or doing FID" debate. This seems to me to be at best disingenuous and at worst outright self-deception, and I suspect that one part of pursuing the war effectively is to be honest with ourselves about what we are doing.