aa777888 - grass is a perfectly safe surface for EOLs - even a runway will roll you over if you get the technique wrong.
The British Military have conducted hundreds of thousands of EOLs to grass over 50 plus years and the number of accidents has been miniscule.
I, like others, used to train instructors which included sending them solo (with a stick buddy on the same course) that made for some interesting rides because not only did they have to fly their own EOLs but they had to correct errors I made to simulate student cock-ups.
All this done on grass and if you saw what we allowed the students to do and how close to the edge we needed to get them to build their confidence, you would not regard me as risk-averse. It's a different sort of risk to winching on a mountainside in the dark, wind and rain but a risk all the same.
recently the full-down became optional on the CFI check-ride because too many FAA DPEs were not current enough in Robinsons to save a CFI check-ride candidate's bacon, and their own
Sounds like people are struggling to me. If the accident ratio was as high as you suggest - where does the fault lie? Is it the unsuitability of the aircraft for the task, the ability of instructors and examiners to monitor and correct student errors, or a combination of both? Either way the teaching and standards are being diluted.