Visible spectrum. There's bound to be some object that glows barely, so some definite measure of luminosity in the visible spectrum should be set.
"They have a lot more data now and there are objects that cover every size from a pimple to galactic superclusters."
yeah, which is why they really need to come up with some definitions. Even if they end up tweaking them later. Up till now, I think they've relied on obvious differences. Planets were clearly not stars, asteroids were clearly not planets. Comets were neither. And then there was pluto.
I like your idea of the contest.
13 Jupiter masses are required for deuterium fusion.
Planets have less mass.
Brown dwarf stars have more and may only burn briefly (~10 million years).
That is the cutoff. (see my other posts re moons and low end planets)