Prop de-ice is a serious business. I reckon that many of the reports of performance loss due to structural icing are in fact caused by loss of power due to prop icing.
I've been into icing conditions (safely, with warm air below etc) to see the impact on the IAS and while I had difficulty picking up more than about 5-8mm of rime/clear in the time I had, I saw a negligible speed drop. My prop is TKS de-iced and the fluid also sprays all over the front window as a result. It's a relatively cheap installation (about £3k) and obviously very effective.
The problem with ground running as described above is that the RPM is low and the prop blades are doing a very low mach number (or TAS if you like). At full cruise RPM, say 2400 or 2500, most of the blade is doing the bigger part of mach 1 and the corresponding aerodynamic heating protects the outer part of the blade from icing. You can still get icing lower down the blade, and those parts are traditionally protected either with TKS or by electric heating. But during a low speed ground run the whole prop can get covered with ice.
The SOP for icing conditions includes going to max RPM, for the reasons above.
Supercooled water can exist (stratus cloud) only between 0C and about -15C and thus any airframe component which heats up by more than 15C will not collect ice. This happens (assuming a surface parallel to the airflow) at about 350kt TAS and, as well as comprehensive anti-ice systems, is what protects jet aircraft and gives them the tremendous mission capability compared to what we have.