Surely the cost argument should be something like cost per target kill. That should give a clear picture of which is the more effective system and it follows that you will get more bang per buck.
The only difficult part is the assumptions made and that's the bit you have to get right.