The three finger test as I use it (and dare I say, as MC uses it) is more of a process method.
I have a dull knife and begin grinding side for side. Once it begins to drag on the stone I will lightly use a three finger test to see if it is catching. This tells me I have either ground to a burr on one side, or am within a few passes or grinding a fresh apex, or have just ground a new apex. The advantage to using this is you don't have to squint, hold the tool up to the light, etc - you can check pretty much on the fly.
Don't just grab a knife and press your fingers into it thinking you're doing it correctly, watch a MC video and see how he applies it. He did not invent this test but he certainly popularlized it.
As the edge gets more refined on higher grit stones, it should still be catchy, this is an important part of the equation. Care/experience is needed, the more you use this test, the more it will tell you. I consider it a near perfect do-anything edge when it leaves small cut lines across the fingerprint whorls, oftentimes not immediately visible until the skin dries out. I can tell by feel pretty accurately without having to actually allow it to drag, but if I do, and the edge is right, this is the result. A wicked sharp, but catchy edge that can pressure cut and draw cut equally well. I do not normally use this test on plane irons and chisels etc, nor would it be a good idea to use this on straight razors. Though I have used it on these tools, it is better to test them on wood and hair respectively.
MC also uses it as a general test of sharpness, paraphrased "if it is three finger sticky and shaves arm hair it is sharp". Not as sharp as it might get with a more specialized finish, but in a general sense as a description of an off hand cutting tool.
Side note, I have used this test on freshly sharpened blades for the paper cutters at work, just to see if there really is an upper limit to what can safely be done, and it still works. Not that I would recommend using it this way - it really isn't safe to use on tools you cannot readily hold in your hand to carefully moderate the pressure. Bringing your unsupported fingers to a blade is not how the test is supposed to be done.