I've put five videos on Youtube, illustrating flight dynamics for an Electra using the X-Plane simulator.
https://youtu.be/2JiGSJ5xPQg shows knife-edge flying, which was proposed as an explanation for the slow descent of N137US. It shows it at about 250 knots and about 150, which was the actual speed of Dad's plane. This is the only clip where I hand-flew the plane. The rest were all run by my harness, for consistency and repeatability.
https://youtu.be/kwVrHJ1obSk shows how I validated the simulator's turning circle calculations. Turn radius accuracy is critical in using the simulator for this research. I ran the plane in circles at bank angles from 10 to 35 degrees and varying speeds and heights. At each setting I ran five or more circles and averaged them for accuracy. The time required for each circle, along with the speed, gives me the circumference and therefore the radius. The bank angle and speed gives me the radius using the standard formula. The extremes of latitude and longitude gives the diameter and therefore the radius. These all agree to a high degree of accuracy. The elongation of circles at a distance from the equator turns out to be a computational pleat used by the simulator to cope with the "flat earth" problem, and has no effect on the short distances involved in my Flight 706 reconstruction.
https://youtu.be/62nRO2VFyGM shows the planned departure of Flight 706, with a gradual turn toward the west. This is shown for contrast with other videos, and describes where the turn starts in relation to the tower and runway.
https://youtu.be/R3ytdwRsxUM shows the increasing bank described in the CAB and ALPA reports, and shows what the result would be of such a bank. All the trial runs end at least 2,000 feet from the known impact point, and none of the runs ends in a 90 degree bank. These runs presume that the ailerons remain at the deflection used to initiate a smooth turn somewhat less emphatic than a standard rate turn.
https://youtu.be/9z_zaW5MvOU shows what it takes to achieve the 90 degree bank described in the CAB and ALPA reports. All these trials end from 2,700 to over 3,000 feet from the known impact point, and their final heading is 90 degrees off from what was surveyed.
I am tagging all these videos with N137US, which is not otherwise used on Youtube. When Youtube gets their indexes updated, that search term should bring up all these.
I have two more sets of simulation runs in mind. One is to explore what would be required for the plane to end in a 90-degree bank at the 4,000 foot distance. The other is to recreate what I think Dad's plane did. Based on the simulation results so far, it appears possible that I'll be able to propose a causal factor that has not been mentioned before. It will be very interesting to see whether this pans out.