This deck was derived from: http://netrunner.meteor.com/decks/LRzjefWSq64d4coSe
End the run is a crutch. So Sayeth Damon Stone, and while he's wrong, not ending the run is not always as bad as it appears it should be. When ice is damaging enough that it must be paid for, runs come a bit slower. When runs pay the corp, runs come a bit slower. And slower and less accesses tends to mean more chances to score out as NBN. Add in the fact that you're yellow and a "slow down" means certain victory. But, if they speed up, they speed you up as well. You just keep pace. You... RACE.
Paywall and Targeted Marketing mean you will almost always be packing cash, or at the very least recover from siphon. Taxing ice pieces mean you can make some servers worse than other, and 5 drawing assets mean you can almost always have an agenda or break R&D lock when you need to.
This is not a deck you will just win with without leveraging the skill of knowing the decks you are facing. Its controllish in that nature. You create paths that are harder to tread, and lead the runner where you want them to run. Versus criminal this almost always means deep servers on HQ, 3 ice, which tax and a tollbooth on archives (if sneak door seems to be in your future). Versus shaper, this means a tollbooth on R&D and on the remote.
You protect your sansan with ash if its required and allow the runner to pay for your traces. You should be taxing the runner all of the time, even if you let them have successful runs. They know they can't let architect fire. They know they can't let archangel fire. They rarely want to let newshound fire. The multi sub and sometimes high strength ice tax faust just enough. And 3x tollbooth means the runner will always be paying money somewhere.
The basic game plan is to establish a current that threatens the economical viability of the runner you're facing or makes runs give you money and leverage that into remote scores. That is the primary goal with this deck is to remote score. Never advance if possible. Use SanSan and astro if you have them, but you can't rely on them.
Thanks @thebigboy for the initial framework. I think I didn't stray too far.