Back in the real world....
Too bad that in the real world, with real speakers that can sometimes be real hard to drive, with real impedance, inductance and capacitance, there is a very real difference, even before clipping. But hey, if you're satified with a marginal amp, good for you.
Best example is when my friends went from a marginal class A/B surround receiver to a NU3000DSP on the Klipsch Forte's in their party room. 50ft of installed wiring, 2 other outdoor speakers in parallel and 3 autoformer volume controls. Not to mention the already well under 4 ohm min impedance of the Forte's and the use of an autoformer in their crossover. Yep. In the setup process we tried my old Yamaha RX-495 which drove the load very well (went into protect once), and my NU6000DSP (PA sub amp) which went into protect mode immediately when asked to supply bass at a moderately high level - it uses bridged amp sections and does not like loads below 4 ohms. It also blew out an NU1000DSP after two weekends. The NU3000DSP has been in service for about two years and still sounds pretty great....not suprising for an amp that will happily drive 2 ohm loads at full power. Overall, we're talking about a pretty major difference in sound quality that was plainly obvious to everybody, even at modest volumes below clipping.
Richard Clark provides a contest to compare amps. The first rule says that you can't go below the rated impedance of the amp. In the real world people do that all the time, usually by accident. Amps also have their own (completely measureable) non-flat frequency response, very few people have a PEQ and REW to flatten things out. And I'd be willing to bet that much like the '87 stereophile ABX amp test that used electrostatics (very flat impedance), they are using speakers that are pretty easy to drive.