So I decided to take matters into my own hands here and do a little experiment.
To recap, Waze was failing to find the optimal routing to Hwy 101, avoiding the obvious route up Maple St
to the freeway onramp.
I recently recategorized Maple St and the nearby Bridge St
from "street" to "primary street" (which they should be, anyway, IMHO, given their roles in the area) and bingo: the routing appears to work properly now, at least in the live map simulation.
So I guess the moral of the story is that the routing algorithm is only as good as the data it has to work with. With the aforementioned streets having been relegated to a status lower than they deserved, the routing algorithm avoided them, picking a route that scored higher, but was in reality less efficient.