His analogy of a magma tube being like a plugged pipe in a pipe organ was spot on. These long period events are the mountain pressuring up to blow. A lot of his colleagues at Galeras bet their lives on Stanley Williams' theory that low gas emissions meant low eruption probability. Six of them lost.
I'd love to see his long period event model plugged into a supervolcano system, not that it would matter a hell of a lot if Yellowstone lets go. There won't be anyone left to say "I TOLD YOU SO!" to.
Calderas are entirely different from individual volcanes. We've never witnessed an entire large Caldera erupting (though we've seen large eruptions of individual volcanoes at large calderas like Rabaul).
It seems exceedingly unlikely you'd see a large caldera blast without quite a bit of obvious warning signs over a long period of time.