A simple explanation of stretch tuning is that strings in physical instruments don't have "perfect" mathematical harmonics. This is because the rate that sound is transmitted in a medium such as the material of the string depends upon the frequency of the signal being transmitted.
I like how the author of the Wikipedia article didn't understand how/why but merely explained "what"
Normally you could look at a string and identify harmonics by dividing the length into integer fractions: 1/2, 1/3, 1/4, ...
In reality the reflections on a string will not line up in such geometric or mathematical perfection. Due to this fact the harmonics of lower strings are shifted from their expected positions such that when you play a higher octave it will phase with the harmonic of the lower string, even if it is tuned perfectly at 2x the frequency.
So stretch tuning is a completely unrelated subject because it is merely a way to cope with the imperfections of physical instruments by "mistuning" those instruments to reduce the undesirable phasing effect: although this is obviously a trade-off and it is impossible to achieve mathematical perfection using a stringed instrument.
With a synthesized piano sound we have no such trouble, while synthesizing the piano sound we can trivially produce mathematical harmonic ratios and so the problem never presets itself. Perfect harmonics, perfectly tuned scales.
Obviously such a "perfect" synthesized piano won't "sound like" a physical piano, as beauty comes from the unique imperfections in something like a physical instrument which in comparison makes the digital synthesized version sound "sterile" or "boring".
Sampled pianos however do re-create this effect with other limitations that belong in some "best tb-303 plug-in" thread or something on an equal level of "totally bogus waste of time to discuss" subjects.
Everyone knows you get the best tb-303 plug-in by plugging in a tb-303.
