I'd like to see someone actually test this in OCTGN.
One deck uses every card legally possible to include into a deck in a house that has draw. The empty slots are filled with cards that provide some other form of reveal, recursion. All other empty slots are random cards added to fill out the rest of the deck form that house. Then it faces against a deck with no draw, no recursion and built in a highlander format. Lets see what the win ratio is.
If we remove the equation of Draw=Win from the context of the game, A Game of Thrones, then we can definitely say that it is an untrue statement. This game though frequently has draw effects on cards that also are versatile and do other things, or let you do something else in order to provide draw. Meaning you can potentially build a deck that is full of characters, events, locations, and attachments that all will allow you to still win challenges while they provide you with draw... so, depending on exactly how one wants to break down the statement, Draw=Win it could be a true statement, the cards that provide you draw in this game will win you the game against a deck which removes all attempts at manipulating your draw statistics.