On the scale invariance of army unit deployments

I had not thought of that, but I kind of think you are right on this one. It appears that we have not been told the whole story. But then, what's new about that?

Well, it would be completely characteristic of BushCo's thinking to ignore logistical concerns in his XXstick-waving plan to "attack them more, MORE!".

It's simpler than that. Why assume the existing support infrastructure is at 100% capacity utilisation? There were as many troops in Iraq 18 months or so ago as they now propose, and I'm not aware of any reduction in support. The extra troops, after all, would have to pass through the system any - the peak utilisation is no higher.

There's already a whole Corps Support Command there, there has been since 2003, and one of those gives you a LOT of capability. Just ship more stuff.

You tell me!
Clearly if they were to add one soldier to the units rotating, there would be no need to add support personnel.
On the other hand, if they were to, say, double the number of combat troops, then one would expect a need for approximately double the number of support personnel.
The relationship is not exactly linear, some classes of support only require one set per theatre, you won't need to duplicate those, some might need more than linear increase, since the number of interactions increases faster than the number of units. But over that range an approximate linear increase in support troops with combat troops should be adequate for blog discussion (as distinct from actual planning real life stuff).

So... when the "surge" was brought up, it always sounded a bit pathetic, since the 20,000 total implied to me only a 2 or at most 3 extra brigades. But if they're sending in 5 more brigades; 20,000 extra combat troops, then they must increase the number of support units - that seems too much to be absorbed by any slack in theatre.

Presumably when they ramped down the combat units in the past, they also pulled back support units as well; so if they pulled back combat brigades, there was also a concurrent reduction in the number of quartermaster battallions and engineering units etc.

But there's absolutely no evidence that was the case.

Anyway, what is more important is that deployments are very nonlinear, because troops come in lumps (units). To deploy one battalion, you'd need far less than an equal number of support troops, indeed hardly any in-theatre. If you ramped that up to a brigade, you'd need more - but you'd get the brigade's service units free, as it were. If you went to a division-sized deployment, you need a serious amount, mostly provided by the divisional support brigade, and once you get to a corps then, yup, you probably do need as many loggies again.

But, if you want to deploy another division, you won't need many more until you max out the capacity of the corps support command, and the definition of a corps goes up to four divisions.

There's also the question of how you define "support". Quite a lot comes from reachback to bases in Germany and the US, infrastructure scaled to support a NATO land war. And combat engineers are not "support", certainly not in this kind of war.

Ok, my bad.
There is a lot of, for example, rotation of Guard units - locally you keep hearing about a company of transport or quartermaster troops coming back or heading out.
As the globalsecurity.org people note it is hard to get a genuine gross total number of troops from the military (which is probably proper, since it is intelligence of use to the opposition) and the different branches report differently.
eg it is not clear to me that when the Army reduced the number of combat brigades this year if it was because the marines took over Anbar.

Now I could probably go over to the Guard and Reserve websites and look at their unit by unit deployment list for the last 3 years and see how it scaled with the 6 month combat brigade rotations, but I can't be bothered.
Easier to post an assertion on the Net and wait for corrections ;-)

I would have thought a fair amount of support would scale with units, like forward deployed combat troops presumably need food and ammo resupply and that should scale linearly with the number of troops (or does the Army use tree structures for logistics and get log(N) scaling?)
I also assumed the engineering units sent out scaled roughly with combat troops, figuring they were doing ordinance disposal and road repair rather than scale independent schools and hospitals...

Ah well, if they had 3/4 of an army corps and just went to 1 full corp plus, maybe they don't need any additional logistics. I'd say that is why they picked that size bump, except that it is pretty clear that all they did was keep the units which were supposed to rotate out and send in the units already tapped to go in. Which completely determines the surge size.

Anyway, it was interesting, because I noticed the media phrasing of the plan had shifted from "20,000 more troops" to "20,000 more combat troops" which implied the total additional number deployed would be larger than 20,000, and if the discussion was being tailored to disguise the size of the surge then it was interesting to see how big.

We'll see how many troops in theatre they report in 3-6 months.

I always did wonder why the battalion of german troops providing security at Frankfurt airport didn't count as part of operation Iraqi freedom, I'm sure the actual members of the unit felt like they were on duty doing stuff.
Not politically convenient.
Like the recent prominent claim that the French were pulling out of Afghanistan made in the WaPo.

Hm, now ThinkProgress reports that CBS has a story saying the military will only make two combat brigades available for deployment to Baghdad and two battalions or marines for Anbar. With another 10,000 - or 2-3 brigades, on "standby" in Kuwait.

WTF is with that.

That's actually moderately sensible. If it all goes to shit on the MSRs, a reserve strike force in the south would be invaluable.