Lest you all get a different impression, I am not at all opposed to using an inert gas or CO2 to displace O2 from the headspace. Let's say you have 2 liters of headspace. If you can go in and carefully purge/displace the air, you can reduce that 2 L of air to the equivalent of, say, 0.1 L. That is a big improvement.
The only point that I have been making is that (in my example) the equivalent 0.1 L of air has unfettered access to your wine (after a few minutes go by). But that may be a smaller volume than the volume of headspace in a well-topped off carboy. It may be a smaller amount of O2 left than if you used a vacuum pump.
For this application, you want to introduce the inert gas near the surface of the wine, and you want to do it without turbulent mixing. This means introducing the inert gas smoothly, at low flow rates. As I keep saying, you have tens of seconds to several minutes before molecular diffusion becomes important, so you can flush nearly all the air out with a slow, steady stream.
I do not have experience with Private Preserve and the like, but I worry that the stream might come out too fast, and induce turbulence.