The 'dusky' Long-tailed Shrike: a wolf in sheep's clothing?

In last week's post I looked at the curious case of the 'hepatic' forms of female cuckoos. The Long-tailed Shrike poses a similar puzzle: a minority of shrikes of this species have a distinct dark form, known as the 'dusky' morph. With today's observation from the Ngong Chung peninsula on Po Toi, I calculate that 90 of 900 observations of Long-tailed Shrikes in Hong Kong on iNaturalist are of the 'dusky' morph (there are various intermediate forms or intergrades, so I considered the round dark face mask as the decisive criterion). Records from eBird show a similar rate, close to 10%.
The dark morph appears to be especially common in Hong Kong and Guangdong province. It is often described as melanistic, but melanism alone should not affect 10% of the species. Moreover, the dusky morph is not simply a darker form of the usual plumage. In particular, the black facial mask takes the form of a round disc rather than a horizontal band. This point is noted in the eBird description ("melanistic form has more extensive dark mask and is darker overall") and shown in a fine photo by Derek Hon: https://www.inaturalist.org/observations/67244093
A fuller explanation may involve aggressive mimicry, whereby a predator takes on the appearance of a harmless species.
When I came across this shrike today, I at first thought we were looking at a Masked Laughingthrush. The size and shape were right, and the black facial disk suggested this species. When my companion with superior eyesight and equipment recognised that it was a shrike, I realised that we were looking at a 'dusky' Long-tailed Shrike. In fact, a few of the 'dusky' shrikes observed on iNaturalist were initially identified as Masked Laughingthrushes. If the disguise is good enough to fool birders, it is good enough to fool prey species into thinking they are facing a relatively harmless laughingthrush, rather than a predatory shrike. Compare this view of the shrike, https://www.inaturalist.org/observations/36365381 with this laughingthrush: https://www.inaturalist.org/observations/221759253
With this form of mimicry, the dusky shrike is effectively a 'wolf in sheep's clothing'. This will give a selective advantage to 'dusky' shrikes, but only as long as the disguise remains relatively rare, since otherwise prey species will come to recognise dusky shrikes as dangerous. This would be an instance of frequency-dependent selection.
We then have an explanation for the 'dusky' morph making up 10% of Long-tailed Shrikes in Hong Kong and Guangdong, where the Masked Laughingthrush is a common resident. If this explanation is correct, the frequency of the dusky morph should be lower where Masked Laughingthrushes (or similar black-faced laughingthrush species) are absent or rarer than they are here.
Aggressive mimicry of this kind is not uncommon in birds of prey. An example is the Grey Goshawk of Australia. The white morph of the Grey Goshawk mimics a harmless white cockatoo, and indeed uses cockatoo flocks as cover when hunting. I believe I have witnessed this, and Australian birders such as these have seen it on multiple occasions:
http://bioacoustics.cse.unsw.edu.au/birding-aus/2013-01/msg00283.html
https://donaldknowler.com/goshawks-get-cocky/
Thanks to Ton Yeh and Kalvin Chan for discussion of this puzzle (which is not to imply that they agree with any of the above).