It was a crapshoot in the kemper days and those generally were gainstaged consistently. When you add in the need to adjust the gain anywhere from +20dB to -20dB it just makes you think “well I might as well just use an algorithmic model and dial the whole tone in”. Way better workflow than going through a list of models and trying to work out which ones might work and what they’re intended to sound like.
If you are wanting to dial something in, captures are a poor choice IMO. The whole point to me is that the amp has been dialled in very specifically already to produce a certain sound. It might sound cool in other ways, but you’d hope the amp has already been dialled in a particular way to achieve a sound.
There is no benefit to gain staging being all over the place, the best you can hope for is something that sounds good (with accuracy becoming irrelevant). Anyone who disagrees compare the experience of Amplitube to Helix/Fractal/Neural DSP. Gain staging being all over the place without users knowing what to do=amateur hour.