You may well be right about personal preference, Dickybee, but as it takes seconds to do if you have the strings off and can't do any harm, then I will continue to do it if I feel a bridge (or the stop tailpiece) is particularly sloppy with no strings on. I certainly don't do it as a matter of course, but only if there's what I feel is excessive movement.

Whilst I'm not disagreeing with anything you say, you'd certainly need more than measuring the movement on a single example to definitively state whether I was wasting my time. And from my experience, I don't think I am.

I'm assuming that bridge and stop piece threads are deliberately quite loose compared to a standard nut and bolt arrangement, to allow for slight misplacement of the post holes when drilling. I've certainly come across quite a few guitars where the posts are just too close or too far apart or at a slight angle to each other to get a bridge or stop tailpiece at the right height without some sanding or filing. Some slight deviation from the ideal probably helps reduce sloppiness as you'll get the bridge or stop tailpiece pulling in or pushing out on the threads against the sides from the top and taking up some of the slack (in addition to the downward string pressure).

But having the post holes the ideal distance apart does allow the posts to move the most as there's no extra force applied when fitting the bridge/tailpiece.

As you mentioned, bridge height does come into play. The higher the bridge sits off the body, the less threads there are in the inserts and the more the posts can move backwards and forwards. And the higher the top of the post is from the body, the greater the distance it will be able to move.

On a bolt-on neck, you can adjust the neck angle, but with a fixed, neck guitar, you have to work with what you've got.

Maybe 10-15 years ago, Gibson increased the neck angle on their Les Pauls (and maybe some others) and you ended up with a bridge that sat way above the body. About this time I set up a LP for a friend that required the bridge to be roughly 15mm off the body (and that was with the lowest action I could get), and the strings had to be run over the top of the stop tail just so that the strings didn't pull down against the rear of the bridge as otherwise the stop tail would have sat up even higher from the body than the bridge did, and with very few threads indeed in the post holes. PTFE was certainly necessary here.

The tape can certainly make the whole bridge feel very rigid. TonePros (amongst others) make quite a feature of locking the bridge and stop tailpiece to the posts for rigidity and more stable tuning. If you buy into that, then having the posts as rigid as possible must also be of benefit.

I bought a used Gibson Flying V that had had a roller bridge and short Maestro Vibrola fitted to it (like Andy Powell of Wishbone Ash's guitar). It wasn't staying in tune very well. Despite having a roller bridge, I could see the bridge rocking backwards and forwards when the trem was used. And the string retainer bar just slotted onto the vibrato spring, and that could move sideways a bit with a bit of pressure from the palm of your hand. So I epoxied the two bits of the vibrola together and put PTFE tape on the bridge posts and now it stays in tune really well, certainly better than most Strat trems do.

I know quite a few luthiers don't like roller bridges, primarily (IIRC) because they often stop rolling due to general dirt and gunk build-up, and you can then get some rolling and some not, which is the worst of both worlds. So if you do have a roller bridge, then you need to keep it clean and make sure the rollers do roll.

They also aren't frictionless (as some descriptions on the web seem to indicate), and it does take a certain amount of initial force to overcome the static friction between the roller and the axle they roll around (once moving, dynamic friction is far less than static friction so they keep moving). To apply that force, there has to also be friction between the string and the roller. If the string simply glided across the roller, then the roller would never turn. Roller diameter and moments come into play here, and a bigger roller would be better and require less string to roller force before the roller to axle friction was overcome and the roller started moving, but you can't make the bridge too tall or too wide, so the rollers tend to be of a small diameter. So that can certainly reduce resistance to string movement, but not remove it completely.