"Mike Rapoport" wrote in message
nk.net...
You still have it partly backwards on droplet size.

Let me provide you with a more probabilistic view of the data you are
relying on, being "observed". Now in fact FAA funded a study of icing from
large droplets, as that was the "observed phenomena". Consider an
experiment where 50 pilots encounter large droplet icing and 50 pilots
encounter small droplet icing. At the end of the experiment, the group with
"observed" large droplet reports 49 incidents, while the small droplet
"observed" group has only 5 advocates.