Obviously GIR is going to be a better representation of low scoring than driving (DA) is. Driving is one shot removed from getting the ball in the cup from the GIR stroke. That's one more shot to screw up (the approach shot). That doesn't mean that the accurate drive was less useful of a skill to have in scoring low. It may just mean that it's damn hard, even for pros, to get two shots in a row (drive and then approach) perfectly executed on the golf course.

I'm certainly agreeing with you about GIR. It is a MUST to be on the green in regulation when you play in the PGA. When I play, on the other hand, it's a bonus.

Maybe an analysis of course conditions would be in order. It has been opined in a couple of articles here (yeah, somebody really reads those) that short grass, big bounces, and long rolls have more to do with driving distances. Are there any stats out there that measure that sort of thing for tour courses. I know that stimp meters measure the greens, but is there anything similar for the fairways? That would be an interesting analysis when compared to driving distance and scoring average.

Another thing to think about is the ability for a player to save a bad tee shot with their next approach. The players that score low and win a lot are the ones that can make the great up and down saves, regardless of driving accuracy. Because of this, they may not have to worry quite so much with accuracy and can go long if needed (see Tiger). The up and down saves statistic may be as good of an identifier of play as GIR.

Additionally, the best identifier of a winner is probably just the simple "putts per round." But I could be wrong.