There are two flaws with rating QBs by game-winning drives. First, as Hitch noted, points scored in the early part of the game count just as much as those scored late. Second, and this gets to the heart of the matter, the worship of game winning drives is based on the myth of clutch performance. There is very little evidence in any sport for clutch performances, and indeed, the concept is not very coherent. If a player performs better in high leverage situations, it’s just another way of saying he’s not trying as hard in less important situations. Professionals are supposed to be giving their all at all times. They’re not supposed to take some plays off, then refocus when the game is close near the end.
Great players tend to have more great performances when the game is on the line not because they’re clutch, but because they have more great performances overall. When Brady is playing, his team is very likely to win, and by chance, some of those wins will come on the last drive. He will have more game-winning drives, but he will also have more games that are not determined at the end. It’s simple statistics.
Look at that link of GWDs that was posted upthread. Suppose we list the players not by GWDs, but by the number of games they've played that were not determined by GWDs:
Brees 191
Brady 187
Manning 162
Rivers 154
Roethlisberger 146
Palmer 141
Flacco 112
Ryan 109
Stafford 81
Russell 59
The ranking is not that different; the players with the most GWDs tend to be the same players with the most games not won at the end. Why? Because they played more games overall! The more games you play, the more chances you have for GWDs. That, plus the likelihood that your team will win, are the two main factors driving GWDs. By definition, GWDs can only occur in games that the QB's team wins. So the more games a QB wins, the more GWDs he's likely to have.
The same point can be made by determining the % of games determined by a GWD:
Russell 26.25%
Stafford 25.7%
Ryan 23.2%
Brady 21.1%
Roethlisberger 21.1%
Manning 19.4%
Palmer 19.4%
Flacco 18.8%
Brees 18.0%
Rivers 14.4%
This ranking looks quite different from the first two. Why? There really isn’t that much difference among most of the QBs. Russell and Stafford are very high, but that’s probably in large part because of small sample size; they haven’t played that many games. Rivers is low, because his team hasn’t won that many games. The other players are fairly close, within normal variation. When we take into account the number of games a QB has played, there isn’t that much difference is the number of GWDs, and what difference there is largely results from the fact that the best QBs on the best teams tend to win more overall.
In summary, the best QBs tend to have the most GWDs--at the end of their careers--because they played a lot of games, and because, being great QBs, their teams probably won a lot of those games. But you can't make comparisons of this sort with all QBs, because in the first place, if their career isn't over, they haven't had the opportunity to play as many games, and second, a good QB on a bad team is not going to win as many games.