The magenta problem is a combination of:
- Gamut volume clipping, at the top and bottom of the gamut volume, resulting in hue skew as the colour ratios end up broken.
- Inherent nonlinearity with of the underlying photometric xy model with respect to desaturation approaches, hence implied Abney mixing on any gamut clip along saturation. Any form of additive achromatic light or implied additive achromatic light via a gamut transform and area clip will manifest this problem. It’s somewhat avoiding the discussion of a more “correct” physically plausible desaturation.