Square your drill press table plane. It's easy.

Why no dial indicator:
"Two human hairs" wasn't entirely truthful. A hair is 0.001", and this table was out about 0.005" end to end, which is about five hairs, or one piece of printer paper. But the table is 16" long, so that means that it has a +/- of 2.5 hairs, or one piece of cheap notebook paper. As a percentage, that's 0.0003125% out of square. Before you suggest that I should improve this number, take note that even the smallest piece of sawdust is a few thou.

Not to dismiss or discourage improvement, but maybe we should find the appropriate level of trouble to go to for a given task.

*****
Method:
If you can smell a flaw somewhere in my reasoning, that's good. Richard Feynman once noted about science that it is a "belief in the ignorance of experts." Not that his opinion means anything. But really, there is a flaw in some of the procedure here, and if you picked up on it, I'm just happy that I was able to get you to think about it that hard. But let me skirt some of that critique with some good old-fashioned obscurantism, anyhow. Consider the context: we are squaring a drill press here, not writing planar topology problems for a specialist-level math text. I want to get the concept across, which is to get you to think of a plane as a triangle. If two points on that plane can be established as equidistant from a point of reference (the drill bit), then the following point can be adjusted to match, and the plane will be set.

On Q&D method #1, the back two points of the triangle, which we used to set the L/R adjustment, are not in line with the center point of our circle (as a diametric bisector). Thus, that line cannot be used to establish a fair basis for the height setting. So, if it reads that point three (the front one) must come down relative to the back two, that's exactly wrong, because our front point IS our point of reference. It would actually be the case that the rear points must come up, so they couldn't be set first. But anyone adjusting a drill press will quickly intuit this, so mentioning any of this becomes irrelevant, at least beyond that weird esoteric satisfaction that drives those who have been cursed by the compulsion to chase after mathematical perfection.

So what about the level 2 method? Again, despite trying to correct along two perpendicular axes, using a four-point based square arrangement is still impractical. Why? Because assuring that points l and r sit on the (new, different) bisector reduces too much of their front/back range, and that reduces their accuracy. As such (considering the dimensions of my table), the best suited approach here would likely be an equilateral arrangement which uses the front & center point. Strange how adherence to theory can restrict one, isn't it? This is why academics never get anything done; a "good enough" mentality can carry us far through accomplishment, especially when it considers context well, because it pays attention to what degrees of accuracy and precision are appropriate, which in turn wastes less productive time.

So, what is my point in all of this rambling? I don't know. I guess that I'm just admitting that it can't be done in only one step with any real-life certainty; adjust for L/R first, using any two arbitrary (but symmetrical) points, and then determine if F/B should be adjusted. In summary, these could be some great chalkboard discussion problems, but as for setting up your equipment properly goes, take these ideas and apply them only so far as your discretion finds them useful. One more thing though: I expect lots of "I just do it like this, and it's done- easy" comments on this video, and I'm sure that such methods can be defended for the sake of pragmatism. But-- analysis of theory provides us insight into places of discovery otherwise unreachable; it is not just useless daydreaming that narrows our focus into that precise beam, as such is our desire to improve asserting itself, and it is the foundation of all human achievement and discovery. Let it do what it does without ridicule, and dismiss it casually at your own expense, because it is to your own shame to mark out your limitations so vividly with sarcasm. Thanks for reading ;)