from Scientific American September 2001:

It may not happen anytime soon, but the Cumbre Vieja volcano in the Canary Islands is bound to erupt again—and when it does, it may well set in motion a tsunami wave more massive than any in recorded history. According to a new model by Simon Day of University College London and Steven Ward of the University of California, this mega-tsunami, which they describe in the September 2001 issue of Geophysical Research Letters, would rouse waves 100 meters high along the West Saharan shore, seas more than 40 meters high on the north coast of Brazil, and water walls towering 50 meters in the air off the coast of Florida and in the Caribbean. Britain, Spain, Portugal and France would also face sizeable waves.

but according to other marine geologists:

The risk of a landslide in the Canary Islands causing a tidal wave (tsunami) able to devastate America's east coast is vastly overstated.

That is the view of marine geologists studying ancient landslides in the area.

In typical Canary Island landslides, chunks of land break off in bits, not in one dramatic plunge, they argue.

as for your points -

Islands with a gradual rise in the sea floor suffered the most damage. This caused me to wonder about the continental shelf along our east coast. There are places over 100 miles from shore that are quite shallow, on the order of 100-200 feet. Would this make a tsunami higher or would the very long gradual decrease in ocean depth slow the wave through friction, causing it to lose energy? In other words, would the continential shelf make the wave more destructive or would it actually serve to weaken the wave through friction? It sounds like a good thing to simulate.

.. wouldn't it depend on the shape of the coastline as well? For one thing, you're talking about isolated islands, not an extended coastline. Could that make a difference? ..Also, wouldn't the specific profile (the actual morphology of the sea floor gradient in the separate cases) be important as well? Could those differences be significant?

And the biggest question of all: am I talking out of my backside? I'm no geologist .. the original modelling must have taken the sea floor profile into account, and the dispute seems to be based only on the question of whether one big chunk, or lots of little ones, would hit the water. Probably a geologist like Nick Hoffman is the one to ask for serious opinions on this one.