I do not anymore. Testing showed me that best seating (in a coarse sense) doesn't really change. It is what it is.
Fine tweaking of seating still 'shapes' grouping, but this is only up to a few thou either way, and does not produce big swings like full testing, unless coming off/into lands, or bridging a powder node.
The trick is finding best seating to begin.
I use a version of Berger's recommended(full seating testing). In this it's critical that you do the testing nowhere near any powder node. Otherwise, your seating adjustments can take you into 2 significant/different changes at once -masking actual results.
Keep in mind also that if you NEED a high starting pressure provided by jamming, for your load to perform(with some underbore, like a 30br), then best seating determined off the lands(OTL) may not be beneficial overall. In this case, might as well seat well into lands(ITL), powder develop, and then tweak seating to shape grouping.
If you take a shaping window into account it explains results like Mozella posted above(this is common). His best seating off-node may have been 8thou OTL. But in-node, best grouping shape could certainly occur +/- 3thou.
Some folks also get the notion that seating is a fine adjustment, but that's actually furthest from truth.
Full seating testing does way more to grouping than powder. They see fine adjustment, because typically they didn't do full testing. Most pull a seating depth out of their butts -for powder testing. Then when they go to seating adjustments from within a powder node, it don't take a whole lot to collapse that powder node. So they end up accepting butt pulled seating +/- a smidge for 'best' :
