erik_t

The answer in two parts: The first answer is that the entire stack is of course a lifting surface. You can flip out of control quite easily with a large axisymmetric fairing. The difference is that this configuration is giving you lift even at zero angle of attack. I wouldn't be surprised if the real thing mounted the payload at a small negative angle of incidence to ensure that there is zero lift along the entire stack when pointed exactly along the velocity vector. The second answer is that NASA/ESA/Roskosmos/SpaceX/etc have waaaaaaaay better SAS and control algorithms than we do Were I attempting to emulate this launch vehicle, I would try to very slightly adjust the angle of incidence of the lifting body, and I would have no shame in hanging big honking fins on the tail of the booster.

I must have identified this problem with a previous KSPI version, because everything is hunky-dory now and works as it should. That'll teach me not to check one last time before complaining...

I've gone with pretty hilariously overpowered 3.75m antimatter-reactor spacecraft of doom, but I can certainly check again.

I have encountered a somewhat esoteric problem with the DT-Vista when building a rather excessive Jool-mission monster-craft. When multiple engines are used on the same vehicle, the module that controls the engine operation caps the trust of the set of engines at 1100kN, rather than allowing each of them to throttle individually. At present, I've had to create my own dummy-engine copy of equivalent thrust, but of course the Isp-throttling is unfortunately lost. I haven't attempted to dig into the nitty-gritty of how the module is written, but might we see this deficiency corrected in the future? I can imagine it being a potentially simple fix.