The name isn't all that obvious, but from what I recall of the original talk by Jake Wobbrock, he intended the recognizer to be very simple and easy to implement [edit: and computationally simple] (as opposed to HMMs or the like). This is alluded to in the pdf of the linked paper: "we present a “$1 recognizer” that is easy, cheap, and usable almost anywhere in about 100 lines of code."

It seems like it's mentioned in the paper already (see abstract). Probably not the best name, but it is pretty catchy.