If you can agree on a "all in one" metric, and a discount rate, then it's somewhat straightforward (albeit tedious) to go through and value each trade made by a general manager.
Similar process for drafts, although (unlike trades) drafts aren't a zerosum proposition, so I imagine that you'd want to benchmark against the expected value of a particular draft slot.
The "all in one" metric is a large hurdle (and perhaps impossible), but it's discussed many places, so I won't attempt to go into it here.
The discount rate is important, so that (for instance) a #1 pick three years from now isn't worth the same as a #1 pick now (since a GM has some incentive to win now, and might not even be around in three years). Conceivably, one could calibrate the discount rate on actual transactions made (although I've done a similar calibration exercise in draft picks over the same year, and the value that GMs ascribe to draft picks of different levels doesn't correlate very well to the measured value of draft picks of different levels).
