This experiment works because the hot water rapidly melts the ice on which it is sat, forming a very good liquid coupling from the ice to the vessel. This then freezes quickly into a tight join, which conducts the heat away from the hot liquid MUCH more. The water in the "hot" vessel with the coupled base loses heat maybe 1000 times faster than the "cold" vessel sitting on a nice insulating layer of fluffy snow, and can in this instance, overtake it and freeze first.
People who run barefoot over hot coals (1200 degrees C) look ever so impressive and empowered, but I bet they can't walk over a hotplate at a mere 70 degrees C if it has a layer of oil over it - same insulation / conduction phenomenon.
BTW that wikipedia entry is a blatant attempt to technobamboozle laypeople. "Continuum mechanics", "partial differential equations" and CFD's trouble with turbulent flow has absolutely nothing to do with this; they are mathematical techniques used to describe or solve known heat transfer problems, they are not physical phenomenon that we don't understand.
People who run barefoot over hot coals (1200 degrees C) look ever so impressive and empowered, but I bet they can't walk over a hotplate at a mere 70 degrees C if it has a layer of oil over it - same insulation / conduction phenomenon.
BTW that wikipedia entry is a blatant attempt to technobamboozle laypeople. "Continuum mechanics", "partial differential equations" and CFD's trouble with turbulent flow has absolutely nothing to do with this; they are mathematical techniques used to describe or solve known heat transfer problems, they are not physical phenomenon that we don't understand.