As published for shadow.foreignpolicy.com on August 7, 2009:

There is a revealing story in today’s New York Times on the Obama administration’s efforts to identify meaningful metrics in the Afghan war. It has a quote from your humble blogging servant on how difficult it was to identify useful metrics in the Iraq war. The quote is accurate and in context (David Sanger is a pro), but it may be useful to provide even more context. The challenge is finding metrics that are valid and reliable and accurate.

Valid metrics correspond to the true situation on the ground and the true prospects for the war; when those numbers are trending up, the war is really trending up, and vice-versa. An “invalid” metric in this sense might be U.S. combat deaths. That figure could be trending up because you are losing the war (as was happening in 2006), or it could be trending up because you are finally taking costly steps to reverse the trajectory (as was happening in 2007).

Reliable metrics are ones that cannot be artificially raised (or lowered) by our actions or enemy actions. An unreliable metric might be enemy suicide bombing attacks. If that is your primary gauge of how things are going in the war then you can be misled in the short run because the enemy could launch a desperate flurry and lead you to draw the wrong inferences about the situation on the ground. The classic example of this was the Tet offensive in Vietnam, which convinced the recently deceased Walter Cronkite that the Viet Cong could never be defeated when, in fact, Tet marked the end of the VC.

Read the full article here