I've held off comment on metrics because, while I believe you cannot improve what you cannot measure, choosing what to measure does to some extent define how you want to see the problem. Plus their are political aspects to the metrics used (obviously)

Personally would focus on a reduction in violence as in a Lowering of the number of civilian dead, versus a counter of the numbers of attacks on ISAF Forces. The persistence of ISAF in theatre is a given, as is freedom of action to move anyway within A'Stan.

Progress might be low/no dead civilians with no/low number of attacks. Failure might be any variation of the above.
I suggest the test being, no attacks and dead = peace. High attacks and dead = not some thing good.
Yes this defines the problem in military terms, because that is the nature of the problem.

I'd also be a bit careful of defining TAC, OP and STRAT objectives, in terms of metrics. Point being that successful TAC actions are irrelevant unless, their "synergy" is felt at the operational level. Point being we have to move on from defining things as "Tactical success" when they produce no operational benefit. Irrelevant success is not success.
I know this is VERY obvious, but I'm not comfortable with "tac success" being "success" unless it gets you somewhere.