Why isn't "Total length of putts made" a more widely used statistic?

[rjsimper]

Is it because this is not something that you can easily measure for yourself, casually? Even if it is only a new ShotLink-era phenomenon, it seems to me that this is the best indicator of putting skill exhibited by a player, in a world where all putting statistics based on an actual round of golf are lacking in some area.  Putts per GIR is largely dependent on distance from the hole, total putts per round largely dependent on GIRs hit and fringe putts too.

I have had rounds where I felt like I didn't putt well and had 25 putts, and rounds where I felt like I putted lights-out and ended up with 30.

There have been rounds where I have made 4-5 birdies and never made a putt longer than 8 feet.

When Gay was spanking the lefty around Harbour Town last Sunday, they showed a stat on the back nine that said that Gay had made something like 180 feet of putts in the final round, whereas the other guy (I forget his name) had made something like 48 feet of putts.

Wouldn't "total length of putts made" over the course of a large sampling of players over 18 holes give you a greater insight into relative difficulty of one courses greens over another, the relative severity of greens, or the difficulty of certain pin positions?

[Phil Benedict]

I like this stat too because it's intuitive. 

[JLahrman]

Not really because it depends on how long the player's putts are.  If a player misses a green, but chips to 5 feet and makes it, I don't know if that counts as better putting than a lag from 50 feet to 2 inches.

I think you have to divide up putts as percentage made from the following distances:

2-5 feet
5-10 feet
10-20 feet
20+ feet

[Jeff_Brauer]

I think it works withoug further divvying it up.  At most, you could divide by the total number of putts to determine the average length of putt made.

I played with Pete Dye at Muirfield once and he made a bet based on total feet of putts made.  After three holes, he had me about 200 feet to 3.  The low number sure lets you know you didn't make any putts of length.

[David Kelly]

If two guys are on the same green and have identical 40 ft. putts and one guy runs it 7 ft. by the hole and makes it coming back while the other guy hits his first putt to 3 inches, who was the better putter?

[JLahrman]

I like to divvy it up a bit because anything where you add length runs the risk of being a bit spurious.  A 25-foot putt and a 50-foot putt are fairly similar for a Tour pro - they're not going to make them too often and they're also not going to three-putt them too often.  But when you start adding distances you're going to give twice as much credit to two-putting from 50 feet as you do for two-putting from 25 feet.

I divided up the distances for the following reasons:

2-5 feet - are they missing what should essentially be gimmes for Tour pros?

5-10 feet - are they making the more critical knee-knockers?  A four footer is a knee-knocker for me, but it's more like 8 feet for a tour pro

10-20 feet - are they making the mid-length birdie-type putts?

20+ feet - This figure should probably be expressed as the average number of putts it takes for the player to get the ball in the hole.  How much does this figure deviate from 2?

[Jim_Kennedy]

Ryan,
Total length of putts is important, but only when used in conjunction with other statistics as none of them reveal much on their own.

[JESII]

I would be willing to bet that Tour Pros three putt twice as often from 50 feet as they do from 25.



Shivas,

"...positive weight to pressure three footers..." in a statistic column...are you a lawyer or something?

[JLahrman]

I would be willing to bet that Tour Pros three putt twice as often from 50 feet as they do from 25.

They might...but I never see that statistic!  Also, three-putting twice as much from 50 feet might be 4% to 2%.  Who knows.

We could add one more cut in there.  Or redo the cuts.  I just personally get nothing from # of putts per GIR or total putts per round.

[JESII]

I guess they all have to be looked at with complimentary stats to mean anything, right?

[JLahrman]

I guess they all have to be looked at with complimentary stats to mean anything, right?

Yes, except for just simple percentages of putts made from various distances.  Trying to come up with one simple number to measure putting is very difficult, there are pros and cons to almost any approach.

[JESII]

If someone has one 6 footer and it has 8 inches of break and he misses it he's batting 0% from 5 - 10 feet...is that a fair representation of their putting from that length?

[Steve_ Shaffer]

I tried this once and had 40 inches.

[JLahrman]

If someone has one 6 footer and it has 8 inches of break and he misses it he's batting 0% from 5 - 10 feet...is that a fair representation of their putting from that length?

No but it will even out over time.  No statistic is going to mean much with only one trial.

And every statistic is affected by course difficulty.  If you only play tougher tournaments (ala Tiger) your GIR% is almost surely going to be lower.

[Michael Moore]

If two guys are on the same green and have identical 40 ft. putts and one guy runs it 7 ft. by the hole and makes it coming back while the other guy hits his first putt to 3 inches, who was the better putter?

I would say that in this instance it's a wash. Maybe the first guy was trying to make it and the second guy was just trying to leave it under the hole. Regrouping after a poor approach putt and saving par by pouring a seven-footer into the heart of the cup is something that every good putter does from time to time.

Total length of putts made is an interesting statistic that deserves better than to be dismissed as "how many bombs did you drop"? This is like saying that driving distance statistics "only" measure the number of drives that you piped down the middle of the fairway.  A made sixty foot putt is a wonderful shot that may be worth two strokes on the card and should be celebrated as such.

[rjsimper]

You all bring up excellent points.

I guess the following question is....

What, then, is the best stat to determine putting skill, or is there one?

The whole "98 out of 100 inside of 6 feet" is misleading too, as 75 of those might be from 1 to 18 inches.

To me the biggest shortcoming in this stat is that making a single bomb skews the data, but over enough trials I think it evens it out.

I've never liked putts per GIR, because many times an 8 foot par putt is harder than an 8 foot birdie putt mentally.

[Guy Nicholson]

It seems to me that a more relevant statistic would be "average length of putt missed." It's deceptively simple -- you have to discard what you make and focus on what you miss. The lower the number, the worse you are. And whether that 50-foot bomb actually goes in is irrelevant, as long as you make whatever comebacker is left.

[Brent Hutto]

Given the ShotLink data or its equivalent, it is a trivial computation to write out a function (curve) describing the relationship between distance and expected (average) number of putts taken for all Tour players. Given that curve, you simply look up each players mix of first-putt distances which gives you the number of putts that would taken on average by Tour pros facing that same mixture of first-putt distances.

If the resulting number for an 18-hole round were, let's say, 25.2 putts taken and the player actually took 23 putts during that round then he putted very well. Well enough to save himself 2.2 strokes over an average Tour pro on an average putting day facing the same distance putts.

Over a single round this admits considerable uncertainty as to whether his set of first-putt distances represented "average difficulty". You could easily have a skein of downhill breakers and all your 4-footers are sliders and end up taking much more than expected for one round. Over four rounds it would yield IMO a valid estimate of how well he putted. Over a season it represents the best assessment feasible for a player's putting accomplishments.

P.S. This statistic is so superior and so easily derived (by those in possession of the data) that I have little time for the much inferior alternatives in common use, even after nearly a decade of ShotLink's existence.

[Richard Choi]

Why don't they just divide the total length of putts made with total number of putts?

That should give you a pretty good indication of how well the person is putting.

[Mike Benham]

What, then, is the best stat to determine putting skill, or is there one?

Years and years ago, Golf Magazine had an article about how to measure your putting in a round of golf, comparing it your total score for the round.  Specifically to give you a putting score measured against par.

The formula gave the golfer points based on the length of their first putt, a +30 foot putt got 2 points, a 2 foot putt got 1 point, etc.  The theory was that you should ALWAYS make the 2 foot length putt in 1 stroke (par for this length putt would equal 1 point) while you should make the 30 foot putt in 2 strokes (par for this length putt would equal 2 points).

The table looked something like this:

Length of Putt     Points
0 - 2 feet               1.0
2 - 4 feet                1.1
4 - 6 feet               1.2

and so on ...


So you would add up your total putts for the round divide it by the total points based on the length of the first putt and multiply that number by 72 (par).

The resulting score should be compared to your total score for the round and rationalized to determine if your putting helped or hurt you achieve the score for the round.

As an extreme example, let's say the length of everyone of your first putts was 35 feet and you 2 putted each one, you would have a putting score for the round of "even par" 72.  A very good putting round.  If your total score for the round was an 80 ... that means your putting helped achieve that score.

To use this system you have to remember to step off the length of your first putt which actually helps you realize that you are closer or farther from the hole then you think and mark it on your scorecard.

I used this method in conjunction with other stats (GIRs, ParSaves, 1-putts, etc.) to analyze the round.

[Guy Nicholson]

Why don't they just divide the total length of putts made with total number of putts?

That should give you a pretty good indication of how well the person is putting.

Because under this formula, two-putting from 70/5 feet and 20/5 feet count as equal.

[Brent Hutto]

Why don't they just divide the total length of putts made with total number of putts?

That should give you a pretty good indication of how well the person is putting.

Not at all.

There's a curve that relates the number of putts it will take you, on average, to get the ball in the hole to the distance you are from the hole. It is not a straight line. You are roughly equally likely to one-putt from 60 feet vs. 75 feet. For that matter the odds of three-putting from 60 feet vs. 75 feet are similar. Yet that's a 15-foot or 25% increase in "length of putt".

Compare that to your odds from 10 feet vs. 25 feet (15 foot difference) or even 10 feet vs. 12.5 feet (25% increase). Those increases make a significant difference in your chances of making the putt.

Your proposed method represents the worst of these two alternatives. A foot is a foot is a foot, whether it's the difference between a 2-footer and a 3-footer or a 50-footer and a 51-footer. The only denominator that makes any sense to use for putting is one that reflects the actual degree of difficulty. Even if we neglect slope and speed (which average out for a given player over time or across players playing the same courses) you can not neglect the highly non-linear relationship between distance and number of putts taken.

[JLahrman]

There's a curve that relates the number of putts it will take you, on average, to get the ball in the hole to the distance you are from the hole. It is not a straight line. You are roughly equally likely to one-putt from 60 feet vs. 75 feet. For that matter the odds of three-putting from 60 feet vs. 75 feet are similar. Yet that's a 15-foot or 25% increase in "length of putt".

Compare that to your odds from 10 feet vs. 25 feet (15 foot difference) or even 10 feet vs. 12.5 feet (25% increase). Those increases make a significant difference in your chances of making the putt.

Your proposed method represents the worst of these two alternatives. A foot is a foot is a foot, whether it's the difference between a 2-footer and a 3-footer or a 50-footer and a 51-footer. The only denominator that makes any sense to use for putting is one that reflects the actual degree of difficulty. Even if we neglect slope and speed (which average out for a given player over time or across players playing the same courses) you can not neglect the highly non-linear relationship between distance and number of putts taken.

I don't understand any of that, but I agree with it entirely.

[Brent Hutto]

Bottom line, you can't add up "feet made" or "feet of putts" or any other thing that treats one foot just like another. Making a 5-footer one foot longer is not the same as making a 50-footer one foot longer.

[rjsimper]

Here's the example of how silly total length is:

You chip in 18 times in a row; therefore, you had the worst possible putting day. 

Not worst possible putting day, but you can't honestly claim to have putted well if this was the case.  Putting well does not have an infallible link to scoring well. If you hit every green in 10 shots, but automatically make your putt, wherever you are, then you've putted incredibly well, despite shooting 198.