Here's the real rub: Many end-feds are fed with effectively zero transmission line loss because there isn't one. You can bring an end-fed wire directly to the tuner, and then if the tuner adjusts the match to SWR = 1.0, where's the line loss? There isn't any. In the old days, this is exactly what many did (and not just hams). As Cecil pointed out, the objective of the tuner isn't to make your transmitter happy (I don't think they have feelings) but to transfer maximum power from source to load. With a big enough tuner (very low-loss components, primarily the inductor and interconnection), tuner loss can often be so small it's nearly impossible to measure. "Small" tuners generally have more loss. So, the "difficulty" becomes only that return current, which is required or no current would flow at all, will now be carried by whatever happens to be connected to the "other" pole of the source (in our case, usually a coaxial connector which is common to the equipment chassis), so now you have the same RF current flowing through everything common to that. With a current-fed end-fed antenna (1/4-WL, 3/4-WL, 5/4-WL etc) this current would be substantial; but with a voltage-fed antenna (1/2-WL, 1-WL, 1-1/2-WL, etc) it can be very small . To help match the voltage-fed designs, which are the most popular because they will work with almost no counterpoise, a high-transformation unun is usually employed and even 9:1 may not be an effective ratio; 49:1 might be, and is still something reasonably easy to wind and compensate. A field strength meter can be your friend in determining the actual radiation efficiency of various antennas installed at the same location. I think a lot of amateur antennas, whether end-fed, center-fed or off center-fed, lack optimum deployment and therein lies their handicaps. I'd rather have an end-fed with one end at my hamshack window and the other end up 120' in a very tall tree than a center-fed horizontal dipole at 25 feet. I've done both, and the much higher end-fed performs better.