man, i really hate to jump in with this one, since you are doing pretty good on this one and your calculations, but there is a big side of this that you seem to have left out of your factors....basically- "rail sag"
now, you touched on it's effect when you pointed out the voltage increase @fs, but just missed it. here's a little insite into how all the different amps are working, and really most are different, so i cannot really give the examples, but i'll shoot. the biggest differences are in the effectiveness of the power supplies. both in the controlling, torroid efficiency, and power starage. so we'll throw out a few categories first, i'll jump into unregulated power supplies. those are set up so that they are being activated at full power at all times, and the only thing that keeps the current down on them is the resistance provided when the rail(s) reach charged. next, is the simple regulated power supplies- they measure the voltage that is stored and only increase the activity to keep the voltage above a certain level. and then there are the more "logic" power supplies, which work in different ways to keep not only a certain voltage range, but usually monitor the voltage drops at different points and adjust the voltage accordingly to maintain a certain wattage. this is typically what you will find in the "any ohm" rated amps. and then you move on to the effectiveness of the torroid/transformer and rail voltage storage. the rail sag is going to be very different from one amp to the next, and that will play a large role on your peak:rms difference. see, many amps will rate the peak power by calculating the max idle charged rail voltage vs. ohm load. others, will measure maximum sustained power, no matter how bad the output sounds. this is more-so what you see when the max is only slightly more than rms. there is also the vast difference in range of voltage sag. for instance, the crappy power supply in a sony explode can charge the rails to, say 36v+/-.... now the sustainable voltage of the power supply without the benefit of chrging, it can only maintain 15v, so your peak is going to be huge in comparison to the point where rail voltage will dictate that the signal is starting to be effected if increased. (this is where the root mean square, "rms" is supposedly on any amp) so, now you take a nice class d, like a directed, or ed, or whatever you prefer as a typical example..... we'll take 2 different ones to compare, both theoretical amps, but reflect real-world readings. amp "a" has a resting rail voltage of 36, yet it maintains between 30-32v in it's higher efficiency loads, and 26-28v at it's lowest and most stressed load, then you have your amp"b" that similarly rests at 117-18v and only dropping to 14-15v. i cannot see the calculations really working out on either of those, but i do not have the whole idea in my head as you do, and i'm really not trying to pick a fight, or discourage, just some things to think about/factor. i also know that defined rms is different from company to company, and even differntly on the same amp of different years, or lines*cough, cough* rockford*cough* one measured distortion could be at the
