Not trying to be a nit but this statement on its own needs some qualifiers to be a truism. For example, what % of current distance is being gained, and how huge is the resulting accuracy dispersion when one is going after that extra distance?

This might be true as a general (on average) rule, but many individuals may find it wont work for them if the changes required for them to gain distance result in significant accuracy problems.

Yes if you are swinging as hard as you possibly can to eek out a few extra yards and are experiencing significantly worse accuracy that is a problem. You are likely costing yourself distance as well because you're probably not hitting it as solidly.

I'm more talking about people being too conservative off the tee. Whether it is hitting less than driver or hitting the driver inefficiently so it doesn't go as far as it possibly could given their swing speed.

I picked 3 pros and calculated correlation of their respective stats from this year back to 2006:
a-driving distance rank and strokes gained tee-to-green rank;
b-driving accuracy rank and strokes gained tee-to-green rank.

Bubba Watson: a=-.33, b=.64
Phil Mickelson: a=.28, b=-.29
Luke Donald: a=.62, b=.50

Conclusions from this:
-When Bubba is more accurate, he scores better;
-When Phil is longer, he scores better;
-When Luke is longer or more accurate, he scores better.


I guess somebody could go thru and calculate everybody who's played for at least 5 years.

Using strokes gained tee to green seems suboptimal. Also using distance/accuracy vs tour average would seem better than rank. Broadies book has some strokes gained driving numbers for various players over different years. When I get time I will run some #s on them.

I just did a quick calc, it does not mean that much to me but using averages to convey a certainty on how each person should play is not logical.

Strokes gained driving is a derivative stat, no?

And I doubt changing from ranking to deviation from mean changes correlation much.

I would be curious to see how much better Phil is out of the rough then Bubba.

This is basically my point, e.g. specific improvements for one person yield more/less benefit than for others, for various reasons. But if you averaged them all out, then one improvement will float to the top.

Strokes gained is a derivative of strokes gained tee to green certainly, but what we are trying to decide is how much more valuable distance is compared to accuracy off the tee. Including the other 2 components of SGTG(approach shots and short game sans putting) is not very effective.

It would be better to isolate the strokes gained driving stat and analyze that. Broadie's book has complete stats from 2003-2012 on the following players

Tiger, Furyk, Donald, Phil, Vijay, Els, Sergio, and Stricker. It also has 09-12 data for Rory.

Here's some results

I ran each players strokes gained driving stat for the year vs their driving distance and accuracy as a % of the tour average.

For example, Luke Donald in 2012 lost -.06 shots per round off the tee. He was 3.5% shorter than the tour average but 7% more accurate.

What I found was that a 1% increase/decrease in distance was worth +/- .17 shots, while a 1% increase/decrease in accuracy was only worth +/- .04 shots.

As a comparison to Luke above you have Tiger in 2012. He was ~2% less accurate than Luke but he was ~5% longer than Luke.

So you have Luke at -.06.

Tiger was 2% less accurate so subtract (2*.04)

But Tiger was 5% longer so add (5*.17)

=

That gives you a predicted strokes gained for Tiger of .71

What was his actual SG Driving in 2012? It was .74 strokes gained per round

So 1% driving is +/- .17 shots and 1% accuracy is +/- .04 shots. I looked at a few years of tour data comparing distance to accuracy. What I found there is that generally, a 1% increase in yardage correlated to an almost 2% decrease in accuracy.

Taking this a step further you get every 1% increase in distance is worth .17 but you will generally lose 2% in accuracy which is worth -.08 (2*-.04) for a net gain of .09 shots.

So take the example that I was arguing a few days ago, where a top instructor was advocating he'd rather his student hit it 310 and a bit straighter than 325 and slightly less accurate. Based on the assumptions above that loss in distance represents a 4.8% decrease. Plug that into our equation and you get

-4.8 * .17 for lost distance = -.82
9.6 * .04 for gained accurace = .38

For a total strokes lost per round of .43. Just a wasted half a shot.

I don't have any Bubba data but here is some more individualized data for Phil and Luke.

For Phil 1% +/- distance was worth .080 strokes gained driving and 1% +/- accuracy was worth .027 strokes gained driving.

For Luke 1% +/- driving was worth pretty much exactly what the overall sample showed, .17 for distance and .04 for accuracy.

For Tiger the ratios are .12 for distance and .024 for accuracy.

How these individual #s are affected when you are only looking at about 7-10 data points is tough to determine. It's also pretty rare for players to change much especially with regards to distance. So not sure how much signal you can get from individual data.

You would need to know how much better Phil is vs Bubba(/tour) from the rough compared to how much better/worse he is vs Bubba(/tour) to be able to make any conclusion there. If Phil's strokes gained approach from the rough is .2 better than Bubba and his strokes gained approach from the fairway is also .2 better than Bubba it is a wash. PGA Tour has a rough proximity number but it is pretty worthless without distance/strokes gained measurement.

I find it a bit hard to believe that someone's strokes gained from the rough vs the field would be significantly different from their strokes gained from the fairway vs the field at equal distances. Even if there was a statistically significant difference I would then doubt it's big enough to offset what we see above regarding how valuable distance is.

Its hard to go from this.

Strokes gained = SG driving + SG approach + SG short + SG putting

To saying, hey give up some of your SG driving and it will improve your strokes gained overall.

Thx. Makes sense.

The only way that would make sense if strokes gained w/ driver was completely independent of the statistics averaging every players performance. I doubt that is the case. In other words, the benchmark should be the most objective measure obtainable. In this case strokes gained tee-green is the most objective because the ultimate goal of driving is to get to the green as efficiently as possible.


To be frank, we should not be cherry picking here. The study must be done for ALL tour players with at least 5 years' worth of data.

If you included Bubba you would have found that accuracy is more important to him in scoring than distance.

It really serves no purpose to say to player A "look at players B-K, they are hitting it x yards over mean and they are successful, and you are y yards below mean so shape up." The only meaningful metric for player A is "when you get more (less) accuracy off the tee, you score better (worse). When you get more (less) length off the tee, you score better (worse)." Whichever is more important is only relevant for that specific player, and not anybody else. You can average everybody's stats and get a general relationship, but it most likely will not suit everyone's individual behaviors.

In summary, if you had to say to 200 touring pros one of the following statements with the least amount of error residual:

A) "Sacrifice accuracy for distance and you will score better", or
B) "Sacrifice distance for accuracy and you will score better"

A is probably correct, but not ALWAYS correct for every player.

Making strokes gained with the driver independent of the other aspects of players games is exactly what the stat is intended to do.

The most objective measure of the "ultimate goal of driving" is to drive the golf ball better than your opponents and thus gain strokes from them. You get the best look at this by looking at strokes gained driving.

Where did you get the "ultimate goal of driving" is to get to the green as efficiently as possible? Setting yourself up for your next shot is important but how you perform the 2nd shot has no impact on your driving ability.

What point would it serve to hit your driver longer but get in a situation where you are hitting in a gap between your yardages, and failing to score as well? That is just one example. If I am hitting the ball longer but not scoring as well, then that means I am sacrificing distance for something else. That will be reflected in sg tee-green.

Maybe it's easier if you just present how the sg driving stat is calculated. If it uses tour averages then that excludes it from further consideration, for reasons already mentioned.

Since Broadie's book also had all of the players above SG stats I was able to get their strokes gained tee to green vs distance and accuracy as well.

Here is how those shake out.

1% of distance is worth +/- .23 shots per round and 1% of accuracy is worth +/- .06 shots per round. A very similar result to the effect distance and accuracy have on SG driving. It's also worth noting that distance and accuracy off the tee is a much worse predictor of SG Tee to Green compared to predicting someone's SG driving.

Going back to your initial 2 players that I have data on, Phil and Luke. Using distance and accuracy relative to average vs SGTG I get exactly opposite results from you.

Using my metrics I get
a-driving distance relative to average and actual strokes gained tee to green
b-driving accuracy relative to average and actual strokes gained tee to green

Phil Mickelson: a = -.62, b = .72
Luke Donald: a = .72, b = -.15

Conclusions you would make from my data is when Phil is shorter and more accurate he scores better while Luke scores better when he is longer but not more accurate.

Really though I would guess this data is just noise. When running regressions on sample sizes this small they are pretty meaningless.

Phil had 1 outlier accuracy year where he was literally 26% worse than the tour average.

Luke on the other hand had 3 almost identical driving years. Each year he was ~-4% in distance and +~6% in accuracy. His strokes gained tee to green those 3 years were

2003: .15
2007: .87
2012: 1.3

His strokes gained driving for those 3 years was in a much narrower range

2003: -.25
2007: .13
2012: -.06

With regard to gapping, i doubt that has too much effect. You are saying that someone should sacrifice distance off the tee so they hit 5 iron instead of 7 iron into the green because they are better vs the field with a 5 iron? Are they enough better to overcome the distance they have given up? I'm very skeptical of that claim. How can they be that much different skill wise between a couple of clubs. Proof would be good.

what is 1% of accuracy?

+/- 1% vs tour average

It's like you read nothing I wrote.

This seems to be going the route of the 1/10 tee not being faster than 1 tee wall of text you kept producing.

still means nothing

Going back to my analysis on SG Driving statistics. If you have a golfer who gain 0 strokes vs the field from his driving, keep his distance the same but increase his accuracy by 1% he will get .04 shots better. However if you were to keep his accuracy the same and increase his distance by 1% he would get .17 shots better.

for accuracy are you just using a yes/no for fairways hit? If so I don't think this is really the best way to do it because not all fairways hit are equal, but more importantly not all misses are equal and the misses should have much more disparate results. For instance, I'd much rather be a 50% FIR player with 40% of misses within 10 yards of the fairway and 8% within 20 yards and 2% greater than 20 yards than a 70% FIR player who misses the fairway by more than 20 yards 20% of the time.

For tour players their bad miss rates to where they are hitting 3 off the tee on par 4s and 5s are pretty low, but for the guy who is between a 10 and 25 handicap that might not be the case.

It is abnormal to assume a player who can only hit 50% of his fairways would have smaller misses outside of the fairway than someone who hit's 70% of their fairways. Golfers directional misses are pretty much normally distributed. I have also already addressed big misses(resulting in penalty shots and pitch out from trees) earlier and how they can alter strategy.

This whole discussion was regarding sacrificing distance just to improve accuracy by a little bit. That is still bad strategy outside of specific hole characteristics would make the trade off a decent option(water hazards, OB, trees, short par 4s where you could potentially leave yourself a 50-80 yard bunker shot which you severely want to avoid).

Nxt,

Do you still think that using 1/10 tee is not faster than using 1 tee only?

I think the main problem with breaking it down more and more specifically is we would be dealing with very small sample sizes.

bumping for the newer folks

My Twitter life would be better if I would have actually followed "I've decided not to fight it".