Just read your link. I agree with him but he didn't address the the needs of modeler users in his various examples
The author rightly states that it takes 10 times the power to increase the volume noticeably with all other factors being equal. This is the exact reason a solid state amp needs to be oversized. You can't run an SS amp into saturation without creating square waves and ugly harmonics so you need extra headroom to avoid this.
For the SS amp to cleanly reproduce the KPA's output at the volume equivalent to a 50w tube amp, you can't use a 50w SS amp because any noticeable peak will clip the amp. A 500w amp will have the headroom to avoid this. This is based on the articles conclusion that a 10 times more powerful amp is only noticeably louder.
This is no big secret. I use a 500w SS amp and can drive it to clip at volumes equivalent to a cranked 50w tube stack. Others use powered PA cabs with 1000w ratings. Even the powered KPA uses a 600w amp. Speaker efficiency plays into this equation as well. FRFR speakers are less efficient, guitar speakers are more efficient.
Perhaps the 100w Class D amp you are using is really rated at 500w peak to allow additional headroom to avoid clipping?
Sorry if this is overly technical. I'm an electrical engineer so my explanation might be nerdy but does agree with your reference link when applied to FRFR operation.
bd