Ken, and particularly Igor...
Fantastic sequence shot! I think the end of the trace shows what happens. Here in the USA we'd call it dirt-tracking, I guess. For those in other continents (Igor, HI!), it is more like a race car drifting a corner with much oversteer rather than neutral slip.
What I mean is that in the last 'stops' as the model turns vertical it looks like it is slipping 'toward the wheels'/ away from the "center" of the "corner," rather than tracking as it does earlier.
I noticed this slip after moving from around sea level to ~5000' (~1,5 KM) altitude. Turning to level from vertical dive looked like a two-step action -- model turns horizontal, then sinks in that horizontal a few inches. Not much, but that close to the ground, a bit stimulating.
By the way, Wild Bill Netzeband, ageless guru for CLPA in USA, holds that the max g loading we might ever see is around 35 g. That would be a 10'(~3m) radius... It is very important to manage energy. Too much energy can disspiate if we try for consistent, too-small corners. We are not flying Carrier or RC 3D prop-hanging. I guess it is another plus for long lines: There IS more room to show a flat side between "corners."