The Value of Accuracy - Analysis of 2009 PGA Tour Results

[Richard Choi]

Going by the early results so far (especially at Torrey where there were some thick rough). I don't think the rule change is going to have ANY significant impact on how PGA Tour players tackle the course or upping the scores.

[JESII]

Until we see a firm course...then we should.

[Brent Hutto]

Until we see a firm course...then we should.

In a typical PGA Tour season, how many rounds does a typical Tour player have on what we around here would consider a firm course? Maybe 20-odd rounds out of a 100-round season? If it requires different wedges, thick rough and uber-firm conditions that's a lot to ask.

[JESII]

I don't think you need thick rough, just firm greens...

[Brent Hutto]

Pity, it's easier to grow thick rough. Especially when it rains.

At any rate, I wouldn't read too much into one tournament. Especially one at Torrey Pines. Heck, it's a muni anyhow!

[Jonathan Cummings]

There are only a few absolute I have consistently observed.  (1) I hit more greens, my scores goes down; (2) I putt less, my scores goes down, (3) I drive more accurately - no affect on my score.

Jonathan, this is very interesting.  I have just a couple questions.

Is it possible that you play on courses where the rough is not particularly penalizing? 

Tony - I have no answer to your "rough" question although my overall "N" is quite large.  (I've played more rounds and courses [more than 1000 the world over] than all but a few on this panel). 

Your answer (3) is most likely - there is enough randomness in my play to cloud any more than rudimentary Guassian anaylsis.  (If it helps you, I've been a teens 'capper forever.)

There is a difference between statistical complexity and stability.  I may be oversimplifying the casual trends I cite, but I can tell you catagorically that these observations are stable and repeatable using just about any sized moving window you want and over any part of my data window.  Try as I may I seem to be Peter Principle-ed into these stats.

BTW - I once used rater data from GW to correlate rating categories to overall score in an effort to quantify "usefulness" of specific categories when determining end result.  Now THAT had some interesting results!

JC

[Anthony Fowler]

As requested, I have repeated this analysis for previous seasons: 2009, 1999, 1989, and 1980.


On average, how much will a 1 st. dev. improvement in accuracy or distance lower a players scoring average (strokes per round)?

Accuracy
2009 -- .33
1999 -- .40
1989 -- .58
1980 -- .54

Distance
2009 -- .33
1999 -- .32
1989 -- .32
1980 -- .36


On average, how much will a 1 st. dev. improvement in accuracy or distance improve a players position on the money list (in ranking)?

Accuracy
2009 -- 14
1999 -- 27
1989 -- 48
1980 -- 39

Distance
2009 -- 25
1999 -- 24
1989 -- 28
1980 -- 23


On average, how much will a 1 st. dev. improvement in accuracy or distance improve a players greens-in-regulation (in pct. points)?

Accuracy
2009 -- 2.2
1999 -- 1.8
1989 -- 1.9
1980 -- 2.1

Distance
2009 -- 1.9
1999 -- 1.6
1989 -- 1.6
1980 -- 1.5


The value of fairways-hit to PGA tour players has been high and roughly comparable to that of distance for the past 30 years.  On the whole, the value of accuracy in terms of money and scoring average has decreased slightly over time, while the effect of fairway pct. on GIRs has remained constant.  This appears to arise from the fact that GIR's in general are less predictive of success on the PGA tour than they used to be. 

Focusing in on the effects of driving on the money-list, I think one could reasonably make the case that accuracy is less important than it was 20 or 30 years ago.  1989 appears to be around the heyday of accuracy, when Faldo, Norman, Kite, Strange, and Stewart were making a lot of money and winning a lot of big tournaments.  Nonetheless, I don't think the differences are not so stark as to raise serious concern.

I think many will be surprised to find that the value of distance has not increased over time.  If anything, it appears that ball-striking overall has become less important, which I personally think is great.  Perhaps the advent of faster, firmer greens has placed a greater premium on the short game.
 

[George Pazin]

I can't comment on your statistical approach (not with the grade I got...), but this seems to be the crux of the problem to me: attempting to correlate season-long performance with the sum of individual event results

To illustrate with an admittedly absurd extreme, I'd say, what if you hit no fairways in 20 events, but every fairway in 5. You won zip in the 20, but won all 5 other events. You'd be pretty much at the bottom on the accuracy list and the top of the money and wins list. (Make the examples a little less extreme and I'd say you have Tiger...)

It seems to me you'd need to analyze specific events and the performance therein to draw meaningful conclusions.

That's waaaaaaaaaaaaaaaaaay too much heavy lifting for me... (Mattspeak to the rescue again)

[Anthony Fowler]

George, I agree with you.  I don't have individual level data for each tournament or hole.  I'm sure it's obtainable, but probably not worth the time.  It would be great to see how the results for individual players change when their driving performance changes. 

However, the less precise approach that I have offered still provides a rough estimate.  Some players are generally straighter/longer drivers, and this variation correlates with season-long success.  As you suggested, the estimated value of accuracy may likely increase when we look at individual level variation across holes, rounds, and events.

As I pointed out before, Tiger was actually one of the straighter drivers on tour in 2009 given his length, so I'm not sure if he fits your extreme case scenario. 

[Jim_Kennedy]

AF,
Using the average of the group,  the top 10 money winners in 1980 would place 55th for accuracy and 63rd for distance.
The top 10 group in 2009 would rank 53rd in accuracy, a minor change, but 43rd in distance, quite a bit better than in 1980. That  seems to suggest that distance has become much more important in the quest for money.

Additionally, Dr Steven Otto, the Director for Research/Testing for the R&A has said:

"...in the 1980s, about 50% of a golfer's earnings could be credited to accuracy from the tee. By the 1990s, that figure had fallen to 25% and currently it is running at as little as 5%. As a result, there is concern that brute force in golf is taking over from skill.'

[Anthony Fowler]

AF,
Using the average of the group,  the top 10 money winners in 1980 would place 55th for accuracy and 63rd for distance.
The top 10 group in 2009 would rank 53rd in accuracy, a minor change, but 43rd in distance, quite a bit better than in 1980. That  seems to suggest that distance has become much more important in the quest for money.

Additionally, Dr Steven Otto, the Director for Research/Testing for the R&A has said:

"...in the 1980s, about 50% of a golfer's earnings could be credited to accuracy from the tee. By the 1990s, that figure had fallen to 25% and currently it is running at as little as 5%. As a result, there is concern that brute force in golf is taking over from skill.'

Jim, I am afraid we are talking past each other in this thread.  First of all, I reject the notion that looking at the relative ranks of the top players on these categories is informative.  For one, accuracy and distance compete with each other so the ranking of just accuracy ignores the confounding variable of distance. 

Second, how much can we really learn by observing extreme outliers?  Tiger Woods and Phil Mickelson are the top 2 players for many reasons.  For starters, they have the best short games in the world.  Just because they hit less fairways than the average tour player and hit it farther than average tour player, it does not mean that distance is more important than accuracy.  (And as I have explained before, Tiger is actually a very accurate driver when you control for his distance, as are most of the top players today.)

These regression results consider the entire tour (not just the top 10) and estimate the value of accuracy and driving for the average tour player.  If you look at the estimated effects of increased distance on money, you will see that the estimated value of distance has not changed significantly since 1980. 

I think it just so happens that Tiger and Phil, the two best short games in golf, happen to be long hitters.  That says nothing about the importance of distance.     

[Steve Lang]

   How do you figure in heart and spirit or even luck of a champion golfer?

   How do you figure in strategy and laying up to favored distances?

  How do you figure in missed 2-4 footers?

AF

Can you post the raw data or probability plots of distance and accuracy?  DOes the data show more than one mode?  Is is arithmetic or log normally distributed?

[Jim_Kennedy]

Anthony,
It's pretty straightforward. The top ten money winners in '80 occupied the same spot on the accuracy list as the top ten players in 2009. The top ten money winners in '09 were 18 places higher up the distance list than the top ten players in 1980.
It seems that the 2009 crowd is much more dependent on distance to make money then the guys from 1980 had to be.

As simple as that method seems, it does tend to be in agreement with what Dr. Steve Otto from the R&A found (I'll bet his methods would go right over my head too), and his finding seems at odds with yours, and dramatically so.



 


 


  

[Anthony Fowler]

  How do you figure in heart and spirit or even luck of a champion golfer?

   How do you figure in strategy and laying up to favored distances?

  How do you figure in missed 2-4 footers?

These and many other factors are in the error term of the regressions  .

Can you post the raw data or probability plots of distance and accuracy?  DOes the data show more than one mode?  Is is arithmetic or log normally distributed?

Distance and Accuracy are normally distributed (arithmetic) and unimodal.  The link below provides several plots of the data.
 
http://web.mit.edu/fowler/Public/accuracy.pdf

In some of the plots, accuracy and distance are normalized so that the season mean is set to zero and the standard deviation is 1.  The scatter plots of distance and accuracy show that the trade-off is primarily linear in the range that we can observe.  The histograms show the unimodal, normal distribution of accuracy and distance.  At the end, I included residual plots of accuracy resulting from the regressions.  The horizontal axis can be thought of as deviation in accuracy from a player's expected accuracy given his distance.  The vertical axis is a player's deviation in the outcome variable (money rank, scoring average, or GIR's) from the player's expected given his distance.

You can see that accuracy tightly predicts GIR's but the fit is weaker for score and money.  As I explained earlier, this is because the relationship between GIR's and score are somewhat weak and declining over time.

[Ken Moum]

As requested, I have repeated this analysis for previous seasons: 2009, 1999, 1989, and 1980.


On average, how much will a 1 st. dev. improvement in accuracy or distance lower a players scoring average (strokes per round)?

Accuracy
2009 -- .33
1999 -- .40
1989 -- .58
1980 -- .54

Distance
2009 -- .33
1999 -- .32
1989 -- .32
1980 -- .36


On average, how much will a 1 st. dev. improvement in accuracy or distance improve a players position on the money list (in ranking)?

Accuracy
2009 -- 14
1999 -- 27
1989 -- 48
1980 -- 39

Distance
2009 -- 25
1999 -- 24
1989 -- 28
1980 -- 23


I think many will be surprised to find that the value of distance has not increased over time.  If anything, it appears that ball-striking overall has become less important, which I personally think is great.  Perhaps the advent of faster, firmer greens has placed a greater premium on the short game.
 

First, I have to ask why you think it's great that making ball-striking less important is great?  I am a crappy ball striker who gets by on his short game, and I don't think devalueing balls striking is a good thing in pro golf, what's your reasoning?

More importantly, I think the two categories listed above are a big deal. Not because distance is more important than accuracy, but because the shift over the last 30 years has been significant.

During that whole time, improving your distance one std. dev. improved your money list position by about 25 spots.

But the value of improving your accuracy one std. dev. has dropped from 39 spots to 14 spots.  And this is during a period when fairways have gotten even more perfect--and perhaps narrower-- while hole locations have been made less accessible. If anything, being the fairway should have becmoe even more valuable.

I am satisfied that my perceptions are correct, and that we need a lighter golf ball.

K

[Jason Topp]

[

I am satisfied that my perceptions are correct, and that we need a lighter golf ball.

K

Ken - have you ever played with a lighter golf ball in wind?  I have with one of the old Cayman balls and hated it.  It was like playing with one of those wiffle balls. 

The floater ball was tried for a short period of time and rejected very quickly. 

I prefer the approach of just reducing velocity by a little bit or requiring the old 336 dimple pattern.

[Ken Moum]

[

I am satisfied that my perceptions are correct, and that we need a lighter golf ball.

K

Ken - have you ever played with a lighter golf ball in wind?  I have with one of the old Cayman balls and hated it.  It was like playing with one of those wiffle balls.  

The floater ball was tried for a short period of time and rejected very quickly.  

I prefer the approach of just reducing velocity by a little bit or requiring the old 336 dimple pattern.

jason, you're talking about a ball that weighs about half of what a normal ball weighs. The USGA spec is 45.93 grams, and Cayman balls weigh 20-35 grams.  Restricted flight range balls are 35-40 grams.

I am talking about one that weighs something like 42 grams, which is prett close to the weight of the ball that was tried and rejected in the 1930s.  But you have to remember that in the 1930s, the balata ball was already VERY hard to control. With today's multi-layer balls you'd see a very slight difference in controllability, but the ball would curve slightly more than the current ball, bringing back the balance I would like to see.

Note-- In looking up the weight of the Cayman, I discovered that someone has a patent on a ball in this wieght range because it optimizes distance for 80-mph swing speeds. See http://www.patentstorm.us/patents/5497996/description.html. It's held by Lauro. c. Cadorniga for Dunlop Slazenger.

For swing speeds over 100 mph, the current ball is optimum, according to the patent.

This is EXACTLY what i have been say for several years. A lighter ball would actually HELP the shortest hitters while reigning in the distance of bomb-and-gouge crew.

See this from the patent document:


A first of the inventive balls 2 had the core 4 of specific gravity of 1.07, weight of 32.5 grams, and diameter of 1.510 inches. The cover 6 was of a diameter of 1.680 inches, and exhibited a Shore D hardness of 63-65. The ball weighed 43.5 grams. The first ball 2 exhibited spin rates similar to those exhibited by the prior art ball of Example 1. In machine tests (FIG. 3), a driver club head velocity of 100 mph resulted in an initial ball velocity of 253.6 ft./sec. (FIG. 2) and a range of 253 yards (FIG. 3), a velocity of 90 mph produced a range of 215 yards (FIG. 3), and a velocity of 80 mph produced a range of 193 yards (FIG. 3). In a player driver test involving six players having slow to medium club head velocities (70-95 mph) the first ball, on average, exceeded the "Top-Flite XL Tour" ball (45.93 grams) by 5.0 yards and the "Pinnacle Gold" ball (45.93 grams) by 6.0 yards (FIG. 4).



A second of the balls 2 had a core 4 of specific gravity of 1.04, weight of 31.0 grams, and diameter of 1.510 inches. The cover 6 was of a diameter of 1.680 inches, and exhibited a Shore D hardness of 63-65. The ball weighed 42.0 grams. The second ball exhibited spin rates similar to those exhibited by the prior art and first balls of Examples 1 and 2. In machine tests (FIG. 3), a driver club head velocity of 100 mph resulted in an initial ball velocity of 253.9 ft./sec. (FIG. 2) and a range of 252 yards (FIG. 3), a velocity of 90 mph produced a range of 212 yards (FIG. 3), and a velocity 80 mph produced a range of 194 yards (FIG. 3). In a player driver test involving eleven players having slow to medium club head velocities (70-95 mph) the second ball, on average, exceeded the "Top-Flite XL Tour" ball and the "Pinnacle Gold" ball by 7.0 yards (FIG. 5).

General comments of the players in Examples 2 and 3 were that the ball herein described produced "softer feel" and "good flight". The players preferred balls of examples 2 and 3 to the ball of example 1.

[Anthony Fowler]

Ken, I think it is a good thing that ball-striking is less important out of my own personal taste.  First, the shots I enjoy watching tour players execute most are inside 100 yards.  There is a lot more creativity an ingenuity involved in these shots.  Second, short game appears to be more egalitarian.  Anyone can work hard and improve their short game, while only the genetically advantaged will have a shot at driving the ball as well as Tiger Woods, Sean O'Hair, Kenny Perry, etc.  I understand that some may disagree, and you are certainly entitled to your differing opinion. 

You raise an interesting question regarding the reasons for the decline of ball-striking in importance.  My guess is as follows.  Once courses are hard enough, even the best ball-strikers cannot rely on their long game to score.  When the rough is thick, fairways are narrow, greens are firm, pins are tucked, etc.; nobody is going to be shooting darts all day.  As a result, the short game becomes more important for scoring.  On the Nationwide Tour, many older veterans complain that they cannot compete because the courses are too easy and the bomb and gauge players are beating them up.  On the other end of the spectrum, the best U.S. Open players (Tiger, Phil, Furyk, Ogilvie, Els, Goosen) are typically the best chippers and putters, but not always the best ball-strikers.  Personally, I'd rather watch U.S. Open golf over Nationwide Tour golf any day. 

[Ken Moum]

Personally, I'd rather watch U.S. Open golf over Nationwide Tour golf any day. 

Me too. In fact, i actually agree with pretty much everything you said.

That said, my personal taste would be to require a little more precision among those who have the genetic advantage of distance.  Not a lot, mind you, but enough to tilt that balance back to where it was about 40 years ago.

The longest, straightest driver is always going to have an advantage, and I think he should.  But the current low-spin ball doesn't do enough, IMHO, to require them to be precise.

So we end up with 25-yard fairways and ankle-deep rough at courses like the one I play. It's a par 71, 6,300-yard Ross, and I have witnessed people playing with or against me drive it inside 100 yards on nine of the eleven par fours.

When I say I want to see ball striking, I mean I would like to see pros with a 3 iron in their hand, trying to get on in regulation. 

Maybe the solution is to increase the minimum length to 600 yards for par fives and 325 yards for par fours.

K

[Jim_Kennedy]

If you're a firm believer in the short game than Harrington's words should be listened to, and this coincides with what I said a quite a few posts ago:
"But if you're aggressive and you're going at par 5s maybe when you should be laying up and things like that, it certainly takes the aggressive club out of your hand," Harrington said. "You can't afford to miss those par 5s coming up short. ... Not necessarily those five-, 10-yard shots; it's sort of the 35-, 40-yard shots out of the rough, and it basically means that you'll see a lot more conservative play, a lot more laying up, and players will have to play a different way."

[Tags:]

Tags: