I knew the issue of thermal ascents by making Solar balloon jumping. A balloon of 400 m³ leads to an inertia mass (mainly air) of about 460 kg. The game is using the kinetic energy of the balloon when it is climbing to lift someone when the rope is taut. The buoyancy is about 200 N (20 kg). So the jump is about few meters height. But if a thermal ascent occurs, the height can be far higher.
The problem is far more serious with utility-scale Magnus balloons inflated with air, whose the volume can reach hundreds of thousands of cubic meters. During the descent phase of pumping mode, if a thermal descent occurs, the balloon will no longer be able to recover, and the crash is assured, with destructive effects: the one-kilometer balloon contains 31,400,000 m³, which corresponds to an inertial mass of around 35,000 tons.