Now, no one ever made a QSO with a computer model of an antenna, but software like 4NEC2 can be useful to gain some insights.
Here's an attached simulated sweep from 80m to 10m of Dave's 72 ft wire. The model assumed:
72 ft long, 2mm diameter copper wire, sloping from 3 ft to 35 ft
17 ft counterpoise, same wire, sloping down from 3 ft to 1.5 ft
counterpoise in the same plane as, and in same direction as, the long wire
average ground conditions
Notice how the reactive part (red curve and axis scale) swings both inductive (+) and capacitive (-) with frequency. Notice also the reactance zero crossings (resonances) where the resistive part (blue curve and axis) is either at a peak, or an some intermediate value between the peaks.
The resonances where the resistive part is at a peak represent frequencies where the wire is behaving as an end-fed multiple of a half wave wire. The other resonances correspond to frequencies where the behavior is that of an **odd** multiple of a quarter wave, end-fed, and generally easier to match.
BTW --- 4NEC2 will calculate the L and C values of a matching network - Tee, Pi, or L (high or low-pass configuration) that will match a given feedpoint.
Dave, when you get a chance to do a sweep with your Nano, we'll all be curious as to whether the general appearance of your sweep is at all similar to this simulation.