As a consequence of the issues raised in the recent debates, I'm
amazed that there seems to be widespread assumption here that the gain
of a tube output stage ( SET *or* P-P ) is simply derived from a stock
equation.

In the gain equation the plate load determines the upper term.

Where the ( refelcted ) load has a variable impedance ( such as a
loudspeaker ) then this term is also variable.

Hence in the real world - it's impossible for such a circuit
configuration to have a flat frequency response. From first
principles.

A flat frequency response can only exist into a pure resitive load.
Speakers are *not* pure resistive by any means.

The response can be *improved* by negative feedback but the response
errors can never be eliminated.

This only applies to circuits with the load in the plate circuit.
Loads in the cathode have no such problem by reason of the circuit
operation ( some understanding of electronic theory is presumed ).

Graham