http://www.tafm.org.tw/Data/011/128/120405.htm
There's a pic for subclavian steal. Instead of going north, blood coming up one verebral goes down the other because there's less resistance downhill into the subclavian artery and inadequate flow coming from the usual blood supply from the aorta.
There's no reason that would happen acutely with G's or a choke, or transiently. It's usually due to plaque buildup or compression from a funny rib, rarely a tumor pressing, and its rare in any case.
"Also I have seen patients severely hypoxic functioning as though little was wrong - again the body can adapt to chronic conditions over time, but not acutely."
This isn't so much an issue of adaptation as it is oxygen requirement and content. You require oxygen to your brain, continuously--and a lot of it. However, there's usually plenty of reserve, thru 1) more than necessary blood flow and 2) more oxygen content in the arterial blood than is necessary. Usually blood is 98% oxygenated or so, and if you look at the venous blood saturation, they're generally >70% so only 30% of supply was extracted. (That's a general consumption, on average I bet brain venous sat is lower than most veins because of the high metabolic rate; heart is also hungry).
So, if you reduce the blood flow, and the body simply extracts more of the oxygen from the blood it does get. Reduce the % oxygen saturation to the mid 80%s and you're usually fine because 80% saturation is still a lot of oxygen. If you only went from 98 to 78 when healthy, you can go from 88 to 68 when unhealthy. Brain doesn't use more oxygen when excited, just muscles, heart, etc, so hypoxia or poor perfusion is going to hurt physical more than mental performance initially. For these reasons we don't even give oxygen to lung patients until sats drop below 89% (for chronic use; in hospital, RN's like everyone >95% so they can worry less, but moral is, you walk around 89 from emphysema all day long and your doctor doesn't care).
In contrast, when you cease blood flow to the brain in a choke, you're going to use what you've got left in the head in seconds. Instead of constantly refreshing blood every 2 seconds you use 20% of the oxygen in, you get nothing new, so that oxygen content goes 100%, 80%, 60%, 40% over 8 seconds and that's goodnight (made up numbers, but more or less the story).
Are chronic lung patients adapted? Sure, usually with higher blood counts and some tolerance. Extraction increases, and some of the means to do that take time. But take someone used to 85%, make them 100%, put them in a choke, and they're down just as fast as a regular person.
"I wonder if you can train for blood loss and reduced blood flow - an ability to function another minute or so in a state of severe acidosis, hypoxia and hypoperfusion. As sick as it sounds it kind of sounds like fun training."
Sure, just bleed yourself out and run the marine obstacle course. I'll be sitting that one out. Patients who are sick enough to have poor perfusion are usually very unhappy or on their way out and I won't be joining them voluntarily.