X20 Sensor, This is News

An 'X-Trans' sensor has less R and B pixels (11% less) - but that doesn't change their colour, nor the colour of any demosaicing.

The colour noise of the R and B channels could be 11% coarser spatially (i.e. in area, not in amplitude) - but then again, the dominant G channel would 11% finer - it's minor 'swings and roundabouts', and pretty small values compared to ISO variations for example (under 1/6th stop in ISO at worst).

Besides - who ever said '25%~50%~25%' is the optimum RGB ratio?

The 'X-Trans's '22%~56%~22%' RGB ratio might be better (I'm not saying it is).

Not to forget - the removal of the AA-filter in the optical path must compensate to some degree, not just in terms of increased sharpness/detail, but also in fractionally less optical absorption loss/increased transmission.