I haven't advocated stopping doing them, I merely agree with the CAA that for the most part they are demonstration only with student follow through at PPL level.
A good instructor will carry out an EOL time after time without incident. The one time it goes slightly wrong however - windshear, student's unexpected response to ground rush, or any number of other variables - will you wonder if it really was a fruitful exercise in the context of who you are training or more a demonstration of the instructor's prowess? Hence the term party-trick.
It would be interesting to see some statistics - number of aircraft damaged and people injured carrying out practice EOLs, against the number of actual engine failures resulting in damage and injury.
R44s, Jetrangers, Bell 47s, Enstroms all auto a dream. H269s aren't too bad either. But the reality is most people train in the R22. Doing EOLs in an R22 is a more delicate matter, and I'm not entirely convinced its worth the risk trying to teach a pre-PPL student the procedure for what I see as very little benefit.
Does the instruction on vortex ring suffer through the student only experiencing the incipient stages thereof? How about the days when we had to demonstrate low g?
A step back, or a step forward?
Just because something has been done a particular way for a long while doesn't necessarily mean its been right all that time.
Just one man's opinion
But since EOL's are something many pilots will have to deal with in reality
Are they?
Do you believe there is a huge difference between a power recovery after completing a flare and levelling the aircraft, and settling it onto the ground and running on as per being taught in running landing?
I guess it also depends on what you consider teaching and what you consider demonstration with student follow through.