By the way, my own little hypothesis about output impedance of SET amps varying over the signal cycle could be completely wrong. Transformers do store energy, which is why the primary easily swings well above B+ voltage. So the transformer itself might keep things constant over the signal swing. Not 100% sure about this.
However ... sharp Class AB cutoff transitions can generate ringing in the output transformer, which is why setting bias in a PP pentode amplifier can be fairly critical. Maybe the real reason partial triode, or so-called Ultralinear, was developed so the AB transitions were better behaved.
But the same applies to transistor amps, as well, setting bias can be temperamental, with the added entertainment that incorrect bias can destroy the output stage (transistors have a positive temperature coefficient and can "run away", carrying more and more current until the bond wires melt and the whole thing shorts out). Fortunately, vacuum tubes are designed to operate at high temperatures and don’t have temperature coefficients like transistors.
(Transistor parameters vary with voltage, current, and temperature, which is why a lot of local degeneration and current sources are used to stabilize the circuit.)