The way I'm defining "chassis twist" is not the same as visible body roll. Not directly anyway. Twist would be the difference between front body roll and rear body roll, and they are going to be slightly different. Not a lot in an S197, because the chassis is pretty stiff torsionally (so the amount of twist is low) and because there are no easy visible references for noticing this twist.
I promise, only a few numbers. The S197 chassis has about 15,000 ft*lbs per degree torsional stiffness, and driveshaft torque might be 800 ft*lbs times 3.3-ish 1st gear times maybe 2.0 either in the torque converter if automatic or as an impact factor for the sudden loading of clutch engagement with a MT. So with around 5000 ft*lbs applied against a 15,000 ft*lbs/degree stiffnss, that only gives about a third of a degree of chassis twist. Too small to see with the unaided eye even if you knew to specifically look for it.
Something must be happening locally involving the combined loadings of the LCAs, UCA, PHB, and ARB at much higher loads than the chassis was designed for. Purely speculation on my part at this point, but it's at least possible for some "elastic buckling" to be going on in the chassis itself, enough to then overload the brace but without leaving permanent deformation or distortion in the chassis as evidence. Even if an 'oilcanning' sound happened, you'd never hear it during a run at the dragstrip with both cars' engines roaring and you with a helmet on (and possibly earplugs as well).
Norm