Malcolm the thing is that whatever voltage is applied to the HT output of the coil, as it is an auto-transformer, inevitably is connected to the primary via the junction at CB of the primary & secondary. I have tested several coils in this manner & they run on the respective vehicles just fine even when hot. I suppose in an unscreened system one could make a case for the coil to be mounted on an insulating block so there would be no incentive for HT to want to leak to the case? I have heard of some people mounting a spare coil in readiness for failure when the coil has got hot.
But a cold coil it seems should have an insulation of many gigohms at 5kV to allow for the very significant reduction when hot & it looks so far like the insulation drops by a factor of ten.
I'm back home now & easier to read things. I have a feeling that a fuzzy trace like that is a sign of a break in the cable or a breakdown in insulation but still allowing the plug to spark. Or might it be like one of those plugs in the article with a spark at the gap but also to carbon in the insulator or an internal break in the plug itself or leakage through a cracked insulator?