I hadn't heard of this theorem, and I can't find a proof anywhere (I suspect it's more like a rule of thumb), but the most precise formulation I can find is:

When a big scare card hits, b/f is better than c/c.
(2p2)

I think here c/f is also better than c/c, and clarkmeister doesn't tell you whether b/f is better than c/f - that does depend on how much of opp's range is scared of the card.