Phantom power killed my guitar tone for the past year or so

[Omega_Void]

...but my intuition tells me that the problem was that the ground and minus pins were shorting.

Could be it. Actually, hot or cold pin would produce similar effect, just with opposite polarity.

...When you amplify the bleeding RF noise (that the ground shield was originally protecting the signal from) in the minus pin, it removes the low end, "good mids" and amplifies the "bad treble" that was not in the original signal.

This is wrong though. Adding DC to AC doesn't amplify the noise. You do lose balanced noise rejection, though 'cos it's basically now got one signal conductor and one ground conductor -- it's an unbalanced cable.

What would happen:

With phantom power on, one signal conductor allows current to flow into the transformer primary at one end, the other conductor grounds it. So there's a few milliamps of DC in the transformer primary, which saturates the magnetic filed in the core, causing one side of the waveform to be clipped. Hence distortion and an inability to create the large swings needed for low end.

(In a mic without a transformer, the DC would go straight to the voice coil and push the diaphragm outwards or pull it inwards. Ouch!)

If one signal conductor is shorted to the shield with phantom power off all that happens is you lose all the signal on that conductor 'cos it's dragged to 0V. Tone remains unchanged as the conductors are all at 0V when quiescent, so there's no DC in the transformer primary. But into an active-balanced preamp input you lose about 6dB of level, 'cos the voltage is only developing over half the impedance.

FWIW.