Monday, July 10, 2017

Links Magazine’s Take on Who Should Pay for USGA National Championships

An article in Links Magazine (Schupak, Adam, “Pay to Play,” Summer 2017) asked who should pay the costs incurred in hosting a USGA National Championship?   The article argues the USGA should ease the burden on host clubs by increasing its subsidy to tournaments.  The argument however is not supported by cogent reasoning or empirical analysis.  There is no doubt the costs and revenues of USGA tournaments are in need of review.  This article does not provide that examination as explained below. 
The article focused on the costs incurred by the 2015 Men’s Mid-Amateur Championship held at the John’s Island Club in Florida.  It also mentioned a couple of other tournament, but did not describe those in depth.  Such a small sample size limits the credibility of the article’s position.  The article did state the 2013 Men’s Mid-Amateur cost the Country Club of Birmingham $300,000, but two years later John’s Island Club spent $650,000. The article is silent on what caused the $350,000 increase in costs.  It would have been important to review the contract between the USGA and John’s Island Club to determine which expenses were required by the USGA and which were optional expenses incurred at the Club’s discretion.

The article describes the current state of hosting costs:

The cost of running even one of the USGA’s 10 amateur championships has grown exponentially.  The starting price can be as modest as roughly $150,000 for the U.S. Women’s Mid-Amateur, balloons to $750,000 for the Walker Cup, and gets close to $1 million for the U.S. Amateur when a larger footprint for worldwide media an television compounds are factored in.

There is no reference to where these cost numbers come from.  This is a serious omission since the high costs are an essential part of the argument for increased USGA financial support.  If high costs are not documented, the article’s thesis cannot be proven.

The costs are purported to have risen exponentially, but the article only presents longitudinal data for the Gold Mountain Golf Club in Washington.  The cost of the Amateur Public Links was $150,000 in 2006.  Five years later the cost was $180,000 for the 2011 U.S. Junior Amateur.  That is only a 4 percent annual increase in current dollars and less than that in real dollars (i.e., costs adjusted for inflation).   The example from Gold Mountain was presented to demonstrate the escalating costs, but actually did just the opposite.

Now costs may have risen dramatically at other events or in years past 2011.  The article, however, presents no evidence to that effect or any explanation of why costs have exploded.  Has the USGA placed more requirements on the host club?  Or have host clubs increased the costs by trying to one-up previous host clubs?   These are important questions that go unanswered.

The article assumes the USGA places an inequitable burden on host clubs and suggests the USGA should pay more of the cost.   Mark Mulvoy, Chairman of the John’s Island Championship Committee makes the following argument:

Why does a non-profit (the USGA) have over $300 million in the bank?  Our members help fund educational programs and build museums and wings of hospitals.

Mulvoy’s complaint is not compelling.  The combined net worth of the members of the John’s Island Club dwarfs the USGA’s cash reserve. So why should the USGA subsidize the top 1 percent on the economic ladder?  It does not help matters that Mulvoy spent $100,000 on a “lavish gala” (as described in the local paper) for former Mid-Amateur champions in February 2016 when the tournament was still seven months away.   In the end, the John’s Island Club Mid-Amateur Committee only donated $7,500 to a junior golf organization.
Lacking any empirical evidence to support the article’s thesis, the author went to David Fay, former Executive Director of the USGA for support.  Fay came up with a platitudinous solution to a problem that may not exist.  Fay suggested the USGA should guarantee that no club that hosts one of its championships is ever put in a financial bind.  Fay’s proposal has two serious flaws.  First, if the USGA guarantees no financial harm it creates a moral hazard.  That is, clubs would not have an incentive to guard against risk because they would be protected from the consequences.  Second, Fay’s proposal would require the USGA to approve and monitor expenses.  If the USGA is to ensure host club is not put in dire financial straits, it must make sure the expenses are both reasonable (e.g., no strippers for the after-party) and legitimate (i.e., no charging the tournament account for steak when hot dogs are served).  This is essentially a forensic accounting exercise the USGA should avoid.

If the article’s thesis is correct and more USGA financial support is required, there should be a shortage of clubs vying to host tournaments.  This is not the case as the article concludes:

…Despite the watered down media exposure and fan interest, and overall lack of attention, top clubs continue to line up to host the USGA’s non-remunerative championships…

In essence, there does not appear to be a problem with the USGA’s current level of financial support.   This raises the question of why the article was written.  Was it to be a paean to the 2015 Mid-Amateur (Comments such as “even the Russian judge would give it a 9.9,” and”Sammy Schmitz made a stunning hole-in-one” were not essential to the article’s objective and should have been removed in editing)?  Was it to lay the blame on the USGA rather than on tournament director Mulvoy for the budget overrun?  Was it to cajole the USGA into increasing the subsidy to a future tournament where members of the Links editorial staff have an interest?  The answers to these questions are not known.  What is known is the article does not meet the high editorial standards associated with Links

Wednesday, May 10, 2017

Southern California Golf Association Errs Badly in Explanation of the Slope System

From time to time the USGA and regional golf associations publish articles explaining the USGA Slope System.  More often than not, the articles get it right.  On too many occasions, however, the articles demonstrate a lack of understanding of the Slope System.  Such articles tend to confuse rather than inform their readership.  The latest example comes from Fore Magazine, the publication of the Southern California Golf Association (SCGA).[1]  The article is entitled “Slippery Slope: Comparing Course Difficulty” and was written by Doug Sullivan, the Director of Course Rating for the SCGA.  The article contains three major errors.  First, the suggested criterion for measuring course difficulty is not adequate for many situations.  Second, the article incorrectly describes how the Slope Rating is determined.  Third, the article’s explanation of why a player gets more strokes is not consistent with the theory behind the Slope System. Each error is reviewed in turn
Error No. 1 – Misleading advice on measuring course difficulty
Sullivan follows in the footsteps of several other articles that claim the Course Rating is the dominant criterion for ranking courses by Course Difficulty.[2]  Sullivan is more adamant, however, and claims the Course Rating should be the sole criterion is measuring course difficulty.  He writes:

… Slope Ratings should not be used to compare the difficulty of different golf courses.  Instead a Course Rating is generally a better indicator in determining the difficulty of one course compared to another. 

To make the case for the Course Rating as the measure of difficulty, Sullivan creates a straw man.  He describes two courses.  Course A has a Course Rating of 68.5 and a Slope Rating of 130.  Course B has a Course Rating of 71.8 and a Slope Rating of 120.  Sullivan asks which course is more difficult.  Sullivan argues your answer should be “Course B because Course B has a higher Course Rating.”   Sullivan never defines a “difficulty.”  For the sake of argument, difficulty is defined here as the score a player needs to play to his handicap (i.e., Diff = Course Rating + Course Handicap).   Because of Sullivan’s fortuitous selection of ratings, a player’s Diff will always be higher on Course B.  Sullivan used this faux example as definitive proof the Slope Rating should not be used to compare the difficulty of courses.[3]
Sullivan concludes:

So the next time someone asks you which course you think is more difficult, compare the Course Ratings, not the Slope Ratings.  Not only will your answer be more accurate, but your friends will be more than a little impressed with your knowledge of SLOPE.    

Sullivan is wrong.  While the Course Rating may be the dominant factor in measuring difficulty, it is not the only factor.[4]  If the definition of difficulty shown above is accepted, then:

                            Diff = Course Rating + Index∙Slope Rating/113

If there is a large difference in Course Ratings (e.g., 3 strokes), the course with the highest Course Rating will also have the highest Diff.   If the difference in Course Ratings is small, then Diff will be influenced by the Slope Rating and the player’s Index.   To demonstrate how Sullivan is wrong, two fantasy courses are constructed.  Course A has a Course Rating of 70.0 and a Slope Rating of 69.0.  Course B has a Course Rating of 70.5 and a Slope Rating of 155.  The vast majority of players would score higher (i.e., have a higher Diff) on Course B since it has both a higher Course Rating and Slope Rating.  But what if a player’s Handicap Index was +4.0?  Course A would have a Diff of 67.6 while Course B would have a Diff of 65.0. This player would find Course A more difficult.  In essence, Course Difficulty for an individual player cannot be determined by simply comparing Course Ratings as Sullivan suggests.

Since Sullivan did not acknowledge the importance of Index and Slope Rating in determining difficulty, his knowledge of SLOPE should not impress his friends.  Moreover, though left unstated in Sullivan’s article, course difficulty will vary by a player’s characteristics (hooker, slicer, short hitter, etc.).  If you really want to determine which course is more difficult, you should play them and draw your own conclusion.  To try to determine the most difficult course solely by USGA Ratings that are prone to error will only yield answers of little reliability and even less consequence.

Error 2 – Incorrectly describing how the Slope Rating is determined
Sullivan demonstrates a lack of understanding behind the derivation of the Slope Rating when he writes:

This is the type of graph (shown below) that is created to compare scores of golfers of different abilities.  As you can see, scores increase as a player’s Course Handicap increases.  A line is created to connect as many scores as possible.  The steeper the line, the higher the Slope Rating.  The flatter the line, the lower the Slope Rating.  The concept of SLOPE simply refers to the slope of this line – that bit of trivia should help you make a few bucks on your next friendly bet.

Sullivan is simply wrong.  Since the advent of the Slope System, the slope of the line relating average scores to handicap is the same for all courses.  After all, that was the purpose of the Slope System. To be correct the horizontal axis in Sullivan’s graph should have been labeled Handicap Index and not Handicap.[5]   If you followed Sullivan’s advice and placed a bet—you lost!

Sullivan makes a technical error when he writes the “line (shown in the graph) is created to connect as many scores as possible.”  Connecting as many scores as possible is not the appropriate criterion for determining the slope.  The line is typically determined  by linear regression techniques where connecting the scores is definitely not a requirement.

Error No. 3 – Confusing the reader on how Course Handicaps are determined
In trying to explain how and why the Slope System assigns additional strokes, Sullivan writes:

In simple terms, SLOPE is designed to make sure a golfer receives more strokes when playing a more difficult course and fewer strokes when playing an easier golf course compared with the USGA Course Rating.

This sentence is so poorly constructed it is hard to discern what Sullivan means.  Since Sullivan has just proclaimed the Course Rating is the measure of difficulty, he seems to be saying a player should receive more strokes at the course with the higher Course Rating.  This would be incorrect.  The handicap player does not receive more strokes because the Course Rating is higher and the course more “difficult.”  Handicap strokes are given as a function of the difference between the Bogey Rating and the Course Rating at a course.  Even in Sullivan’s own example, most players will get more strokes at the course Sullivan has declared to be the easiest (Course A).

So why do articles such as Sullivan’s continue to be published?  It’s probably because neither the SCGA in this case nor the reader are sticklers about accuracy.  The SCGA is looking to fill its magazine and assume the Director of Course Rating must know what he is talking about.   And if he doesn’t, it doesn’t really matter.  SCGA members can be divided into three groups.  The first will pass on any article about the Slope System so the image of the SCGA will not be harmed.  The second group will assume the article must be accurate since it came from the experts at the SCGA.  The third group that recognizes the article as nonsense is a very small subgroup of the SCGA membership (i.e., approximately 1) that can easily be dismissed.  

[1]Sullivan, Doug, “Slippery Slope: Comparing Course Difficulty,” Southern California Golf Association’s Fore Magazine, Spring 2017, p. 94.
[2]Cowan, Jim. “An Explanation of Slope,” Northern California Golf Association website,  Metropolitan Golf Association, “How do the Course Rating and Slope numbers affect my Handicap Index?”
[3] It is not clear why the authors of these articles have chosen peculiar Course and Slope Ratings to make their point.  Jim Cowan of the Northern California Golf Association, for example, used a course with a Course Rating of 72.8 and a Slope Rating of 114.  Since Course Ratings are highly correlated with Slope Ratings, it is very likely such a course does not exist.  Real courses could have been selected to demonstrate how difficulty is measured, but no author has taken that path.    
[4] Cowan, loc. cit.
[5] Stroud, R.C. and L.J. Riccio,  “Mathematical Underpinnings of the slope handicap,” Science and Golf, E & FN Spon, London, 1990, pp. 129-140.  Stroud shows the Perfect Valley Handicap as the label for the horizontal axis.  The Perfect Valley Handicap is a player’s Handicap Index. 

Wednesday, April 12, 2017

Another Inequitable Tournament and So It Goes...

This blog has documented many cases where a misunderstanding of the USGA Handicap Manual has led to inequitable results. The latest example comes from a tournament at a club in Southern California.  The tournament consisted of four different stroke play competitions of nine holes each: 1) Four-ball, 2) Total score of partners, 3) Scramble, and 4) Pinehurst. 
The Tournament Committee imposed a limit of 8 strokes between the handicaps of the partners.  The handicap of the higher handicapped player was reduced until the 8-stroke limit was reached.  This, however, is not what the USGA recommends:[1]
It is recommended that the Committee considers it a condition of four-ball stroke play competitions that the Course Handicap (after allowance) of the members of a side may not differ by more than eight strokes. A side with a large difference has an advantage over a side with a small Course Handicap difference. If a difference of more than eight strokes cannot be avoided, it is suggested that an additional 10 percent reduction be applied to the Course Handicap of each member of the advantaged side.[2]
Moreover, this recommendation is only for four-ball stroke play.  Its application to the other forms of competition, as was done in this tournament, is an egregious error.  To see the size of the error, assume Player A has an Index of 3.5 and Player B has an Index of 14.8.  The Table below presents the handicaps that were used in the Tournaments and those that would have been used if USGA guidelines had been followed. The USGA handicap calculations are shown in the Appendix.
Tournament and USGA Handicaps

Player A (3.5 Index)
Player B (14.8 Index)
Four Ball
Total Score

In Four-ball, Player B would play as a 7-handicap under USGA guidelines instead of a 6-handicap in the 9-hole competition. The competition used a modified Stableford scoring so that a net par was worth 2 points and a net birdie would be worth 4 points.[3]  It is likely the loss due to the Tournament handicap is in the 1 to 2 point range. (Note: The difference in Tournament handicaps after the allowance is 7-strokes and not the 8-strokes as recommended by the USGA.)
In the Total Score event, Player B would play as an 8-handicap rather than as a 6-handicap that was assigned by the Committee.  The loss due to the Tournament handicap is probably in the 2 to 4 point range.  If Player B bogeyed each hole where he had an additional stroke, he would gain 2 points.  If he made par on the two holes he would add 8 points instead of 4 for a net gain of 4 points.
The Scramble competition would not be affected since both the Tournament and USGA handicaps are the same.  Similarly, the handicaps are equal in a nine-hole Pinehurst competition.  The Tournament Pinehurst handicap of 3.5 is rounded to 4.0, the same as the USGA handicap.
The number of points lost by this team due to the Tournament Committee’s handicaps is in the range of 3 to 6 points.   Would this have affected the outcome?  Probably, since no team that had its handicap reduced by the 8-stroke rule came in the money.  The bigger problem, however, is the Tournament Committee that failed to follow USGA guidelines.   Players have an expectation a tournament will be run fairly.  In this case, that expectation was not met.

USGA Handicaps
Four-BallPlayer A’s course handicap is 4. His handicap after the 90 percent allowance is still 4 (4x.9= 3.6 rounded up to 4.0).  Reducing his handicap by an additional 10 percent still leaves the player at a 4-handicap (4  - .1 x 4 = 3.6 rounded to 4.0).  Player B’s course handicap is 16.  His handicap after the allowance is 14.  After an additional 10 percent reduction, his handicap is 13 (14 - .1 x 14 = 12.6 which is rounded to 13).
Total Score of Partners – The USGA recommends players be assigned their full handicap.  Player A would be a 4-handicap and Player B a 16-handicap.
Scramble – The USGA recommends the team handicap should be 35 percent of Player A’s handicap and 15 percent of Player B’s handicap.  Player A would have a 1-handicap (.35 x 4 = 1.4 rounded to 1).  Player B would have a 2-handicap (.15 x 16 =2.4 rounded to 2.0).
Pinehurst –The USGA recommends the team handicap should be 60 percent of Player A’s handicap and 40 percent of Player B’s handicap. Player A’s handicap would have a 2-handicap (.6 x 4 = 2.4 rounded to 2).  Player B would have a 6-handicap (16 x .4 = 6.4 rounded to 6).

[1] The eight stroke limit stems from research done by Francis Scheid published in Golf Digest in June 1971.  Scheid never studied actual tournaments, but used scorecards from his home club to simulate matches.   In 1971 there was no Slope System and the bonus for excellence was .85 rather than .96 as it is today.  Nevertheless, the  8-stroke limit has been imposed in many four-ball events even though it has never been proven to lead to increased equity in studies of actual competitions. 
[2] USGA Handicap System, Sec. 9.4bii.  The USGA does not recommend the 8-stroke limit for Four-ball match play.  The USGA’s reasoning does not seem consistent.  If a large difference in handicap leads to low scores in Four-ball stroke play, it would seem that a large difference in handicap would also lead to low scores in Four-ball match play—i.e., the team with a large difference would always have an advantage.  The USGA has never explained why the 8-stroke limit should only apply to Four-ball stroke play.
[3] Modified Stableford scoring adds an element of serendipity in deciding the winner.  Under modified Stableford scoring, two players with the same handicap and gross score can have different point totals in the Four-ball competition.  It is not clear if the Tournament Committee purposefully wanted to add an element of chance to the scoring or simply made a mistake in selecting the modified Stableford scoring system.