The document Fly posted discusses auto rotation as one descent and flip-flop spiral etc. the two different ways I showed in my testing. then you will have sliding and rocking. it varies. you keep saying "it tumbled" end of story. that's not how it's done.
all the data is there for auto rotation and non rotation. 5,000 feet descent values. wind data. different size leaflets.
The leaflets have been dropped for years,decades. one would think the data would be more accurate than a test done 18 times?