Is there something that you feel like you are missing by running conservatively?
As long as your equipment is good enough to have a reasonably quiet noise floor, there is nothing technically unsound about lower volume recordings in the 24bit realm. I generally try to run any kind of recording so that I never peak much over -12dB.
Riding the levels in real time is not something I would do, especially in 24bit. As far as adjusting the levels in post at some variable rate, I know there are different philosophies and some people routinely use compression etc. My own preference is if the original presentation is very dynamic, I think the subsequent representation should retain that same dynamic element as well. If it is so extreme that it is difficult to listen to, I might try to compress some parts, but I don't like a lot of production on amateur recordings.
The issue of one instrument overpowering the others is something you might can address with variable gains in terms of making a more listenable recording in terms of end to end volume, but it is fundamentally a "mix" flaw and with two ambient mics, there is only so much you can do.
I don't know if it would help, and can guess it might in fact make things even worse, but I think between a combination of the fact that you are running two distinctly different signals into each mic, and I assume that the Fender is basically in front, it might be interesting to at least see what you get from a Mid-Side recording.