I'm going to go out on a limb and say that following patrol routes is always a good definition of properly positioned.
A good way to get an F-117 shot down, for example.
I mean it's not like our bases get attacked often (as in ever) so it's only when soldiers leave them they're in danger, hence you could easily use a few UAVs to be in range of any given patrol route, while easily costing less than fighters being scrambled from an air field far away. Modern artillery repositions itself very quickly, though flight time is most certainly a concern (moving after calling in an artillery strike is a risky proposition). I'm unsure how much of a concern, as if a building is occupied, artillery repositioning and firing is going to beat fighter jets nine times out of ten, and the building in question is not going anywhere. Close air support A10 or Hind style is out of the question with artillery, obviously, but UAVs seem to do that just fine.
Artillery, at the moment, has the problem in the current fight with high CD, so many missions it could support don't get called. Risk of CD to high.
That's the loadout he quoted for the F-15s. I'm unsure what they're doing with it, destroying tank depots? Hitting an entire column?
Supporting a three day fight at An Najaf, 2004. F-15's did great, as did the F-18, F-16, and attack helicopter assets. Funny old thing, UAV's weren't up to that fight. I'll not comment on other operations.
Reducing multiple city blocks to rubble?
No. That is the Bone's and the Spirit's mission, if called for.
It seems like overkill of the highest order. The 500 lb bombs aren't that bad, but the 1-ton suckers seem... excessive... for insurgents.
I respectfully disagree, having more than once been involved in weapon selection
screaming matches decision making event. Your perception of the tactical utility of the GBU-31 is narrow. Depending on target and fusing, it is relatively discrete in its foot print. With no time delay and and HE warhead, on the other hand -- chutney.
Not every war, nor every phase of every war, is what Iraq is now, nor what Afghanistan is now. I'd be leary of trying to cookie cutter the current state of play when considering a requirements model, even if we are hypotheticalizing about here. (OK, probably not a word).
For example, in the weeks between leaving the LOD to Baghdad finally falling, 2003, CAS as traditionally used (fast response time, ticklish FAC to pilot comms, various restrictive fire control measures) is a finite asset with great demand from the ground commander running the close fight. The situation is not relatively static, as the current Iraq situation is, but the speed and flexibility of the autonomous pilot/flight of two/flight of four, rather than the centrally controlled UAV package, is a huge boon to the local tactical commander, where the fight is won or lost. Decentralized execution at its finest.
In 2004, which I am familiar with, Reaper was still in OpEval, so Iraq typical CAS sorties were
F-15
Tornado (limited bombs payload)
F-16
F-18
While now and again a sortie checked off of station with clean bomb racks, the vast majority dropped no, one, or two GBU-12 sized munitions. Now and again the 2000 pound load was used, but it wasn't that common. 9There was a strike, not CAS actually, on some underground hideouts near Fallujah that I recall used the bigger bomb, to good effect.) What they had over the UAV of the time was response time, man in the loop, speed, and an asset of low density/high area coverage. (They were also much less vulnerable to SAFIRE and SAM in the launch and recovery phase). By the way, the decision not to use 2000 pounders, vice 500, cost us at least one shot at killing, versus scaring, Zarqawi. So he lived another two years.

He got a lot of people, mostly Iraqis, killed in the interim.
The nice thing about Pred, and now Reaper, is that an armed low speed, long dwell time (hours and hours of eyes on target, no refueling breaks, etc) set of eyes can be pre positioned. Trouble is, stuff/trouble frequently broke out Some Where Else. The ability to respond to that was modest, at best. Fast a UAV ain't. Yet.
If you begin to fill the sky with Pred/Reapers, depending on your scenario, you can overcome some of that and still lose the ability to concentrate fires the way a two ship of F-16's, not to mention F-15E's, can. That is part of the required force mix. You also begin to negate the price/cost advantage.
UAV/Predator, when I was involved with such things, though often armed were primarily ISR assets. (Intel/Surveillance/Recon) Even then, however, some of us were fighting (along with the Spec Ops sorts) to get USAF and the Joint Air Community to acknowledge UAV as a CAS asset, or at least with occasional CAS tasking. Not sure how that has since played out, since we don't assign helicopters to CAS, even though they too are airborne fires. Funny thing, a lot of our CAS sorties, while waiting for a call for fire that didn't come during that mission, were also employed in an ISR role, albeit limited in some regards.
I am no longer in the doctrine wars, so maybe I ought to peak in and see how that has played out, in published Joint Doctrine.
In a dynamic environment, I don't find the UAV model to suffice. For the static environment, where the probability of needing a concentrated strike is low, yes, armed UAV's are a great tool. Hell, that is why they have, for the past 5-10 years, been funding magnets and grown so in use. For the current condition they fit many needs.
The argument isn't either/or, manned unmanned, it's the matter of force mix.
If you make a big enough, fast enough UAV with jet power and speed, you will begin to replicate the cost profile of manned aircraft, and not see the cost savings you tried to promote on the ratios previously mentioned.
I think Geni makes a good point, and with a set of UAV's now in development that are remotely controlled, F-35 is the last of our manned fighters/fighter bombers. By the time it matures and reaches the end of its life cycle, the order of battle will have, I expect, fast, unmanned, armed UAV's.
DR