Let's look at a 10 card deck with 3 agendas in it. Agendas are N, anything else A. Here's a few possible orderings:
There are 10! ways of actually ordering these; less since there are only 2 "types" of cards and we're not differentiating. But whatever, here's a very small sample that should illustrate why DBS doesn't protect you that much. We'll assume there's a single DBS active on the field. For the first ordering DBs fires, you pick the non-agenda card. You're still vulnerable to indexing, maker's eye. R&D interface, not so much. The second ordering puts you more at risk than you were by forcing an agenda to the top that would have been deeper. The Third ordering is the same situation where you have to decide to make your hand worse 2 times in a row creating that agenda bubble potentially on the bottom of your stack (which is better than being flooded perhaps, but doesn't actually improve R&D's quality, you're still susciptle to maker's indexing and RDI). And the last one actively makes the runner's accesses better in just a few turns.
These decks were a bit more "dense" than a normal deck of netrunner cards. But 3 agendas in 10 cards is not highly unusual. If this you combine them into a 40 card deck you can see how DBS doesn't necessarily protect your R&D accesses because even if it lets you move an agenda that was on top, you still have agendas below it that might now be exposed.
NAAAANAAAN AANAANAANA ANNAANAAAA AAAANANAAN
(Same orderings as above, just now a deck, has a few more 1 pointers so think Jinteki PE)