According to the formulas, skin effect only begins to be important at RF frequencies. However, if you try, say, 10 ga. solid core (only--no other strands) as an interconnect on a decent system, I think you'll find that it does indeed make a difference at audio frequencies. This is a clear example of one of those "numbers don't tell the whole story" situations. Enough of those and you begin to look at audio and so-called "snake oil" upgrades in a different way. Some of them actually matter.
One of the things to remember regarding strands--if insulated--is that it increases the capacitance, which begins to tailor the frequency response. In a worst case scenario, it can cause oscillation and blow amplifiers (happened frequently back in the '70s). Moral of the story: There's a balance to be struck between the number of strands, their gauge, whether they're insulated or not, composition of the wire, composition of the insulation, etc.
Grey