Quote:
Originally Posted by shemp
Strokes gained is destined to being circular unless you accept it abstracts away information (I know, blah, blah, blah)-- and that will take you seeing some cracks, and that will likely start with how strokes gained extends backwards off the green towards the tee. It's a seismic shift in what appears to be an ideology.
Apropos nothing. Two players A & B arrive on the green 35 ft from the hole (whatever the 2 putt distance is). They both putt off the green into a hazard. Player A chips to 35 ft and makes his putt. Player B chips in. Who putted better?
Maybe "laughable" is wrong. How about meaningless?
Quote:
Originally Posted by shemp
To analyze a single round of golf based entirely on the numbers in shot tracker is inferior to using both those numbers and what you actually observe.
This last part is almost certainly true, but that's about all.
I've actually implemented the strokes-gained algorithm for my own use and that of others who aren't PGA Tour Professionals. I've got over 2 thousand holes of my own in my database, each and every one of which I charted the 1st putt distance with 1 thing in mind: how does the algorithm perform for this specific putt? So I've got a pretty good idea about the "cracks", and whether or not they're "seismic".
Yes, it's possible that the specific putts each player had on Sunday were such that the true difficulty and expectations were far different than strokes-gained would have you believe, but you've done nothing to quantify the difference -- it's only 1 round after all, so shouldn't be too hard -- and I'm extremely skeptical that if you did it would be all that significant.
I didn't see most of Sunday because I was out playing (and shooting my lifetime best round), but from what I can gather, Rory made a lot of long 2-putts (due to a conservative tee-to-green strategy) that you're claiming were easier than strokes-gained would imply, as he hit to safe spots. In my experience, long putts are very rarely the ones where expectation is greatly different than distance would suggest. This shouldn't be surprising, as the longer the putt, the more distance control becomes the predominant factor in how close you are to the "ideal vector". The individual putts that really break the strokes-gained algorithm on a regular basis are things like downhill sliding 4-footers that are much more difficult than the 90% strokes-gained would suggest. However, it doesn't seem to me that there were many of these in the round in question.
So what's so hard about admitting that strokes-gained isn't perfect, but it's pretty damn good, and this instance is probably not one of the outliers where it's not, but there can still exist other, separate concerns about how players should make their way around the course in different tournament situations.