John,
Valid point about the tail rotor. With regard to main rotor efficiency, it may become more efficient (lift vs drag ratio) at a lower RRPM but it will still produce less total lift for a given pitch setting. Lift is proportional to the square of the airspeed over the aerofoil, all other factors remaining constant.
I would think that if you were still climbing at 500 fpm at zero airspeed with a stuck collective, then you are in a bind that calls for drastic action.
In that situation I would reduce RRPM with throttle as much as I felt comfortable with (towards 90%). I would experiment with manouvreability at this reduced RRPM, considering the possible increased chance of retreating blade stall or stalling regions of the main rotor disc during manouvres. If that was the case, (probably evident by massive vibrations!) I would leave it at 97% (bottom of the N2 green band).
Then I would get it down to the ground by turning it on its side at slow speed (a tight descending turn), setting up the highest rate of descent I could. As the ground comes up and as late as possible, roll it right way up (which would rapidly take care of that ROD), lose any speed you have and roll off the throttle, all as quickly as I could before it started climbing again. Wait for what would be a significant but hopefully well survivable impact.
I would definitely experiment with it once or twice before committing myself to landing by rolling off the throttle completely. The 20kg of fuel you might burn in 15 min of experimenting in my opinion is not going to make your predicament much worse.