Jerry, I assume you are referring to your longwire setup.
If so, the external coupling ferrite and parallel tuning capacitor
will not be very critical. As long as it peaks the wanted frequency,
optimisation in terms of rod and litz specification, number of turns, position
on the rod will not make a lot of discernable difference. Moving the
coupling rod relative to the receiver will, as you have already ascertained,
allow you to find a rough optimum. Changing the L/C turns/capacity
might improve selectivity, but probably not noticeably. More important
is avoidance of too close coupling, which might "derail" tracking of
the internal varactor C in the radio.

Optimising the quality of the external L winding would become more
relevant if it were to be directly connected into the radio so as to tune
with the internal DSP varactor, in lieu of the internal ferrite winding.

The next step up in terms of selectivity would be to add a Q-multiplier
or regenerative RF-amplifier to the coupling ferrite L/C. (We would
need expert opinion to know whether that might disqualify the setup
from uldx criteria, however.)

I dont have the real estate for a real longwire or beverage, but your
observations are beginning to tempt me to experiment further again :-)

Happy hunting!

Michael UK