More on Strokes Gained - Putting

One thought I have about this new concept of Strokes Gained - Putting is that it's sort of a partial version of an idea I've had for a while. A year or two ago I had the idea of giving each shot you make in a round of golf a "grade" on the par scale. The idea is that par-level shots will get you a par on a given hole, while one birdie-level shot lowers your expected score on the hole by one, and a bogey-level shot raises your expected score by one. So, for instance, if you hit a solid drive down the middle of the fairway, followed by an iron to four feet, and then miss the putt and tap in for par, you went par-birdie-bogey-par. But there's a problem with that way of counting things, which is that it's not always clear where the difference from par should go. Suppose, on a short-to-mid-length par-4, your drive goes long and straight, leaving a short wedge into the green. Then you hit that wedge well, to ten feet, and then sink the putt for birdie. Or, conversely, on a par-3 you hit your tee shot into the fringe a dozen yards or more from the pin, and then a so-so chip shot to maybe seven feet, but then you miss the putt. Which shot in the first case deserves the birdie, and which in the second should take the blame for the bogey? It's hard to say; in both cases it was more of a team effort.

So my idea is to designate shots as having (potentially) non-integer values relative to par. The way this works in practice is to visualize a topographical map of sorts over a golf hole of expected average strokes to get into the hole from each position on the course. For a given shot, you take the expected stroke average from the point where that shot ends up and subtract from it the expected stroke average from the spot you hit it from, and then add one. So if you stand on the tee of a par-4 hole with a stroke average of 3.95, but then you hit a beautiful drive down the fairway to where you'd expect to take 2.73 shots from there to hole out, that shot was worth -0.22 strokes relative to par (obviously, negative numbers are good; you could flip that, if you wanted to, but why bother?). A whiff would be by definition +1.00 strokes, since you end up where you started. Holing out from a spot with X expected strokes would be -(X - 1) strokes relative to par.

Strokes Gained - Putting is basically the same idea, just for putting only. They start with an estimate of the chance that a given putt will be holed out, based as I understand it almost exclusively on the length of the putt. Then the idea is that, if there's a p chance you make the putt and you do make it, you gain (1 - p) strokes on the field, and if you miss it then you lose p strokes on the field. This is very similar to my idea, except slightly disregarding the possibility of three-putting. If there's a putt with a p chance of sinking it, then you have 2 - p expected strokes to hole out from that spot. If you hole out, then obviously that stroke was worth -(1 - p) strokes relative to par. If you miss it, assuming you go to a spot with certainty of holing out, then that stroke was worth -(-p) = +p strokes relative to par. It's the same thing, except with the minus signs flipped.

Ultimately this is the authoritative way to analyze the quality of golf shots. If you could create a perfect map of stroke expectancies from each spot on a golf course, you could compute things the way I've sketched out and then say that a player was x strokes under par on the year with his driving, y strokes over par with his irons, and z shots above or below par on putting. But that would be really, really hard to do effectively. In fact it's hard enough to do with putting. Anyone who plays the game of golf knows that there are lots of things aside from length that affect probability of making a putt; a fifteen-footer from slightly below the hole with two inches of right-to-left break is much easier than an eight-footer from well above the hole with a foot and a half of left-to-right break on it. But it's damn hard to quantify that: instead, they've just assumed that length is king, and everything else will average out in the long run. And they're probably not far from right about that. Moreover, they don't even account for the fact that, even if there is a 2% chance of holing out a putt from 70+ feet, the odds of three-putting are sufficiently high that you probably have more than 2.00 strokes left on average.

The point is that it's tough to make the map of stroke expectancies that I envision as the perfect way to judge a golf shot. It's hard enough when all you care about is what happens once you get onto the green. I like that they're trying, and I would be really interested to see if they ever expand this concept to more parts of the game.