On Sunday, May 17, 2020 at 6:20:53 PM UTC-7, wrote:
The tow cable pulls on the tow plan with some force (F) and angle (theta).. When the glider balloons, both the force and angle increase. This lifts the tail of the tow plane with a force Fsin(theta). When Fsin(theta) exceeds the control authority of the tow plane, the tow plane is forced into a dive.

Presumably, cable force has an upper bound of 80% to 200% of the glider gross weight (FAR 91.309).

It might be possible to design a tow plane release mechanism with something similar to a back release. I'm thinking of a torsion spring-loaded mechanism with some length (x) such that the tow cable tension exerts a torque xFsin(theta) on it.

During towplane upset fatal accident in Ephrata (1999) the rope did break, but the attitude of the towplane was such that a recovery was impossible. What I envision (which will probably draw the boo-birds) is a state estimation computer that monitors pitch rate of change, along with other flight parameters, to predict an upset in progress. If an upset isn't detected early there is no hope of recovery by any means.

Tom