The process of designing a SE tube amplifier can be somewhat random with the idea of using a LOW DCR power supply being a key concept.
I will typically look at the internet forums or other online sources for schematic or ideas that will catch my imagination, then modify the signal circuit to suit the parts I have in my inventory.
Two freeware design programs that I have found useful are PSUD2 and LTSpice4.
After I figure out what sort of circuit I want to build, a live circuit is constructed on a breadboard.
My thoughts on breadboards are this : I don't want to format an expensive metal chassis until I get the basic functionality of the circuit verified.
If the basic tube amp measurements are not correct (operating points, THD/THD+N, IMD, sine and square waves, bandwidth, phase etc.) or the subjective sound is not correct on an Altec speaker system, then the circuit stays on the breadboard until a satisfactory result can be achieved.
I rarely get a circuit out of the breadboard stage because most of the time the subjective sound is satisfactory - as it should be if the breadboarding process is well thought out.
IMO, the availability of inexpensive test and measurement equipment is within reach of almost everyone in the tube amp building hobby. Why not learn to use that sort of test equipment?
My test bench features a PC built to run ARTA freeware, BK 2120B dual trace 30MHz analog scope, Loadstar AF generator, DMM's, ESR meter and various capacitor/component testers. That test gear represents the bare minimum needed to operating efficiently as a designer of tube amplifiers.
Some in the tube amp hobby prefer to use "ear testing" as the primary arbiter of circuit performance but I find that approach to be somewhat flawed as it is easy to overlook simple design / build errors that result in a lot of wasted time.