Forrest Scott Wood said:
There is an opinion that the sliding action of the skate, and yes by needs sprung, is more beneficial for electrical contact and it is also somewhat believed that as a bit of an added bonus at no extra charge the sliding action offers a bit of benefit in keeping the rail clean-er than otherwise.
May be related to how a number of trolley and interurban lines here in US went from trolley poles with wheel collectors to sliding shoe collectors because the shoes offered more continuous contact than the wheels, which could bounce a bit at times.
There are potential issues with these sliding shoes. Among them snagging on uneven rail ends at joints, even with ends of shoe turned up; shorting across turnout/switch parts which have too little clearance between rails; having issues with switch frogs; snagging on the same if a groove is allowed to wear in shoe; and probably more which are not recalled at the moment.
As it happens, pantograph shoes have the same groove wearing problem, which is why you will see catenary wires zig zagging across track centerline so as to spread out the wear. There are also pantograph shoes which have a spanwise slot to carry lubrication of some kind of conductive grease or the like.
Back to the shoes on G trucks; several real-world electric operations which have not used the conventional outboard 3rd rail have used shoes on the running rails or more often on conductive rails between the running rails. But they are able to use sloped approaches and exits from the conductive rail which would not be workable on running rails.
I haven’t any available documentation to prove this statement but the characteristics of sliding contact shoes is well understood in the prototype world and used to be in the model world, maybe up through the 1950s before declining in favor of 2 rail, except for traction modeling.
Somewhere in a model magazine in the 1980s was article about a fellow who had a large layout in O scale which still used outboard 3rd rail current supply even on the big steam locomotives.
I wonder who that was? Hmm. There was a Bill Schoop who did O scale scratchbuilding of brass steamers but I don’t think it was him.
Oh, speaking of him, have seen a 1950s scratchbuild of on of the little B&O “Docksider” 0-4-0T (which I think they classed C-16) (they also had 0-6-0 called Docksiders) where Bill drilled one driver axle hole off-center by something like 1/64 or 1/32 inch in order to emphasize the rocking and hunting the very short wheelbase caused in the prototype.
Wow, that’s far more trivia flooding out than I expected.
Oh well, here it is, make of it what you will 
{EDIT: Hey! where did the paragraph spacing go??? I typed this with double spaces every few sentences like newspaper articles and now some web articles do}
Your talking about Frank Ellison or John Armstrong but I thank Frank and his Delta Lines.