Rtk vs rtpk, for wide open control points

I have learned that it is a good practice to have multiple control points, on any general project.
So, if I have a few minutes extra, I set another nail, rr spike, HIT, (hole in top, drilled for a more precise location on rr spike) or a bolt, or nail. You never know when somebody will decide to park over your favorite gps base station, or simply too close to it, for comfort.
So, here is my question: in theory, is RTK, or RTPK preferred for these shots?
Having come from the world of single frequency, static GPS, I'd assume RTPK should be slightly better/preferred. But, I have not been leaving a base and rover out all night, and all day, and reviewing the data, (especially little time frame pieces, to simulate real surveying situations) which, as I take it, is a common procedure, on the development team, and programmers group. I also know that as RTPK is still being refined, etc, without other extenuating factors, which one is, or should be more accurate? RTK, or RTPK?
I set one yesterday, and they were were within 0.016 of each other. (This has not always been the case) sometimes, they have been 0.035, or 0.045 apart)
Thank you for this great gear!
Nate

Due to absolute calibration being used with RTPK, it will be more precise if both the base and rover are aligned to north, otherwise there is no significant difference in the tests I’ve done.