With that it mind, what does 2X sampling multiply the axes by?
If I recall correctly, 2X is an MSAA mode and not available to SSAA. It just means their are two two samples averaged in the pixel rather the 4 shown in the example. Unfortunately, with only two samples, the blends effectiveness is extremely dependent on the angle of the object sampled. Hence, 4X MSAA shows quite an improvement over 2X.
Nvidia article is wrong, or at the very least semantically sloppy.
4X samples generally means doubling of resolution for both axes. 4*(x*y)==(2*x)*(2*y).
Yes, 4X SSAA doubles the resolution of each axis. This results in a quadrupling of pixel count (Hence the 4X). In cameras, moving from 2 Megapixel to 8 Megapixel is considered a 4 fold increase in resolution. Likewise, since resolution on monitors is also measured in two dimensions, a doubling of each axis is a quadrupling of resolution. It is technically the semantics of this article that are sloppy. Not that it makes much of a difference as Don spells out what is meant.