Ah, thanks, I'd seen that thread but hadn't noted your particular interest in it - I can understand why you'd want to know more.
Initially it appears there's some further (useful) detail given on the ADS-B Exchange page, but a quick manual calculation just on the displayed altitude and time intervals shows a possibly different rate of descent than reported, which suggests to me it may be interpreting the RoD from the full set of data (and/or there's further data inputted, but which may be discarded for clarity of display).
As I've (locally) done a bit with the tar1090 code that ADS-B Exchange runs on, I'll see if I can find some time to have a quick look around what calculations it may be doing with the raw ADSB-out data, and from that what it displays...
FP.