Well, you won't be shocked to hear that I think the spine whack is valid, and I find a knife that doesn't pass to be unacceptable. I use a light whippy wrist snap rather than a firm hammer-like whack (uh, we're still talking about knives here, right?).
One note up front. Liner locks (and to lesser extent, frame locks) usually have a short break-in period once you take them out of the box. That is, they have a period where the lock wears relatively quickly; once that's done, it only wears very slowly after that. I don't advocate taking a liner lock out of the box and spine whacking it right off -- the lock is definitely at its worst at this point. Instead, open the knife 30-50 times, medium-hard (NOT hard enough to damage anything, but enough to shake out any break-in wear). Then if it fails the spine whack test, I worry. I do not do the test until the lock is broken in.
I am with megalobyte in that I feel the lock geometry is the likely culprit for spine whack failures. I'll also tell you that where the liner engages on the blade tang (whether left, middle, or right) is not necessarily a great indicator of how reliable the lock will be. If the geometry is wrong, it's wrong, and the lock will fail. I've seen plenty of knives that lock up way to the left that I can't fail no matter what, and plenty that are over to the right that fail easily. In fact, it is a common occurrence that a knife that passes the spine whack test for months, suddenly starts failing as the lock wears and moves to the right. This refutes the theory that the mere position on the blade is what determines reliability. Why? Because as the liner wears and moves right, the lock geometry changes. The geometry has to be correct across the entire movement of the lock; it's geometry and not position that's the major factor.