Regarding the TK22 carb kits:
The K&L kits had decent jets and needles, and the Keyster kits did as well. So far only the no-name kits with the plain boxes and keyster part numbers had bad needles. I don't recall if we used the pilots and mains from the no-name kit yet, so those parts are yet to be evaluated.
At idle, the bike should be running almost entirely on the pilot system for fuel. The slide cutaway exposes the needle jet to atmospheric pressure, thus there is little to no vacuum to pull fuel up the main. The pilot system lets in air and fuel, but the metered air is for the purpose of controlling the amount of fuel flowing in the pilot system. (Using this method to control the fuel is nice since fuel won't gum up the air screw, since there is no fuel there.) The amount of air actually passing through the pilot system is almost insignificant relative to the air coming through the main slide opening. You can verify this by monitoring the vacuum on a manometer while turning the air screws. It does not alter the vacuum by any measurable amount (relative to the unchanged cylinders). Whereas even the slightest change of the slide adjustment for a given cylinder will alter the vacuum by a large amount (relative to the unchanged cylinders).
So basically, the air screw is a means for adjusting the amount of fuel in the pilot system for idling. But it is not adjusting, by much, the amount of air being burned at idle. The air screw is affecting the amount of vacuum available to the pilot jet for the purpose of pulling up idle fuel. Letting in more air, through the air screw, reduces the amount of vacuum pulling up fuel through the pilot, without altering the total amount of air entering the engine at idle, (by much).
If the pilot system is very lean, to the point that the main circuit is required to provide some of the idle fuel, then the air screws will have much less effect on the mixture. Then closing an air screw might not stall a bike. I usually only see this on bikes with a vacuum leak or fuel level issue, or if the idle mixture is far off for some reason. As a matter of fact, this may have been happening on the bike we were jetting with the faulty needles. After going back to the factory needles, suddenly everything started working properly again.
None of the pilot/idling issues will have too much affect on fuel mileage. That is mostly a factor of the needle and main jet, and maybe slightly a factor of the pilot jet, for slower speeds. Cruising is mostly a needle adjustment area of tuning.