Originally posted by jon.partsch I'm no expert, but it seems to me that many people tend to overthink this. On a camera with Live View (i.e. contrast-detect focus), just photograph something contrasty parallel to the image sensor. I tape a piece of junk mail to the wall. Set camera on tripod, switch to live view, push shutter half way to focus, turn off live view and push shutter half way again - if the focus ring moves in phase-detect focus from where it was after focusing in live view, then you will know if you need an AF fine adjustment, and you will also know in which direction to adjust. Make an AF adjustment and repeat the live view vs. phase detect focus test until the focus ring no longer moves between focus methods. This seems to work great for me on any AF lens.
Your methodology is sound here, but this would not work for all camera/lens combinations.
In the real world when we half-press our shutter buttons (or use back-button AF), the lens may need to change focus from near to far or far to near. The action of achieving AF lock is not an exact science. If you do some testing with the camera on a tripod and an unambiguous focus point (such as a focus chart), you will generally find that out of 10 attempts the camera will only get about 7 of them identical. The rest will be very close but not the same point. You can only see this with a lens with distance markings. This is why it is important to do multiple tests, making the camera focus the lens from a minimum and infinity position each time. You then adjust the Fine-Tune AF setting to match the "majority vote" of your testing. Different camera/lens combinations will also yield different results in terms of the number of identical focus locks.
In practical terms you will probably not notice any difference using your method or my wine-bottle method, compared to a rigorous test-chart approach. But if someone has a lens that they feel is always slightly back or front focusing, the test chart approach is the best one.