Well, that can't be true indefinitely or you would have an infinite sequence.

Unless you posit as an article or faith that absence of foreshocks just means that the foreshocks were below detection limit (which might mean negative magnitudes in some cases) I think you would have a hard job substantiating your case.

It would be interesting to put some figures together on %s of events with detectable foreshocks; one would imagine there must be papers on the subject.

Equally, what % of small to moderate earthquakes actually are foreshocks? It has to be less than 10% by simple application of the Gutenberg-Richter law.