I've discussed this before, but let me recap the rudder issue and knife edge flying. The Electra can knife edge fly at about 250kts IAS. At anything less, the rudder cannot depress the tail against the airflow to lift the nose and cause the fuselage to act as an airfoil.
For any given speed under 250kts, there is some angle at which the rudder can just barely hold its own against the airflow. At any lower angle the rudder can counter the bank and roll the plane level. At any higher angle the rudder is insufficient and the plane will drop its nose. Sim runs tell me this angle is about 60-65 degrees at 160-ish kts. In contrast to knife-edge flying, I call this bevel-edge flying.
My Dad was at about a 57 degree bank at about 160 kts with a rudder whose bank-countering functionality was degraded to just about nothing. They didn't know that; they never trained for it; no one trains for this. But once the rudder took hold and the bank started to flatten, the rudder would have increasing effect. A full left rudder at a 60 degree bank is nothing at all like a full left rudder at 35 degrees. Once the bank starts to flatten, the effect of the rudder will increase exponentially.
This is a little like rolling a ball along the crest of a berm. Once it starts to move to one side or the other, it will accelerate quite quickly. As long as it's on the crest, it might stay there. And simulating this particular instance requires quite a long run along the crest, which then exponentially cascades. I hope that's not too obscure. What it means is that my chances of replicating this with the simulator are virtually nil. I can animate it, but I doubt I can simulate it.