Wednesday, April 11, 2012

USGA Maintains Infallibility

The USGA continues to defend the indefensible.  Appendix E is clearly in error, but the USGA maintains, against all evidence, that it is not.   Below is a discussion of the errors in Appendix E with Scott Hovde, USGA Manager of Course Rating and Handicap Education.

Dougharty (2/1/2012): I see the typos in the web version of Appendix E were corrected as I suggested. You did not, however, change the row headings which I also believe are in error. Let me argue my case again. The probabilities in Appendix E represent the chances of a net differential less than some negative integer. When you make the row heading a range you introduce large errors. For example, the probability for a 5.0 index having a net differential of -8.0 or better is 1 in 20111. The current Appendix E assumes the probability of having a net differential of -8.9 or better is also 1 in 20111. It is actually closer to 1 in 45,000--i.e., a player is 2.2 times more likely to have a net differential of -8.0 or better than he is to have a net differential of -8.9 or better.

Here is what I suggest:

1. Change the row headings back to negative integers.

2. Change the example so it produces a negative integer as was done in the original Appendix E. Let the interested reader worry about interpolating between integer values.

I did receive my new handicap manual today. I was amazed no one caught the typos in Appendix E. I assume a draft was run by various committees, but the members did not take the time to read it. If you ever establish a Copy Editing Committee (CEC), please consider my application.
I am still examining the question of whether the probabilities in Appendix E are reliable estimates. I have one lead on their origin. Knowing that I have until 2016 to complete my task removes a lot of pressure.
Regards,
Larry

Hovde (2/9/2012):  When reviewing the entire Handicap manual for the 2012 update, we looked at the table in Appendix E to determine if it needed some tweaking or updating. The initial thought was to remove the table from the book altogether and only supply an online version once the table was fully reviewed. In consulting with a few of our HRT members (who were involved with the original table), it was decided to keep Appendix E in the book, but expand the rows to ranges, as the formula (and example) listed underneath the table to determine the net differential is taken to the tenth place. The values in each cell represented an average of the probability of net differentials within that range, and not just a single net differential. In reality each 0.1 difference in net differential will have a separate probability, but that table would be far too large to be of use.

In the interest of ease of use by our target audience, the ranges were there to keep them from having to interpolate and just plug in the net differential, and probability was listed in each cell as the divisor, instead of probability itself (so 1200 instead of .0008333…). We also felt probability (in the form of listing the divisor) was more relevant to the typical reader, as 4:1 odds is easy to mistakenly interpret as 1 in 4, as opposed to 1 in 5.

The typos in the manual were unfortunately made at the printer when they formatted the table for final printing, the table we submitted and that was reviewed by all involved had the correct values. Your offer to be part of a Copy Editing Review Committee in the future is noted. While 2016 may be the next version of the Handicap System manual, we do intend on continuing to research this topic (and others) and provide any updated information via the Handicapping section of the USGA website.
Thank you for your clear attention to detail and communicating your concerns with the intent improve the Handicap System. Feel free to pass along any further feedback relating to this topic.
Regards,
Scott

Dougharty (2/13/2012):  Your defense of Appendix E was both disappointing and unpersuasive.  The disappointment stems from seeing a great institution, one responsible for protecting the integrity of the game, blame a printer for an error clearly its own. Someone at the USGA should have signed off on the proof copy of the manual. If no one did, the error is still the organization’s fault and not the printer’s--see Rule 6-6.

You argue the values in each cell present an average probability of net differentials in that range. Your argument suffers from a number of weaknesses:

1. Originally, the cells were represented as the probability of shooting a net differential equal or greater than the number in the left column—let’s call this the integer probability. Now long-time members of the HRT have allegedly come forward and stated the cells represent the average probability of shooting a net differential in “some range” or better. The integer probability and the average probability have different values. Yet the values in the cells (with one exception) have remained the same in the revised Appendix. The HRT members were either wrong in the past or wrong now. (Note: You will now have to inform golf associations the probability of beating your index is no longer 1 in 5. It’s now closer to 1 in 4.)

2. I believe “ranges” were entered because someone mistakenly thought it gave the illusion of precision. Ranges actually make a mess of things. The revised Appendix states the probability of a -7.1 net differential is 1 in 1200. This is not correct. What you now have is the probability of shooting a net differential in the range of -7.0 to -7.9 or better is 1 in 1200. If this sounds clumsy, it is. Making reasonable assumptions, the probability of a -7.1 net differential or better is around 1 in 780. The revised Appendix is in error by approximately 35 percent.

3. If there was a conscious effort to have ranges, I would expect the text of the revised Appendix to reflect that change. It does not. The Appendix reads “The values in the table represent the probability of shooting a net differential EQUAL TO OR BETTER THAN the number (emphasis added) in the left column.” This sentence was not altered from the old Appendix. If there was a careful rewrite of the Appendix as you maintain, the description of the cell values would also have been changed.

4. In the development of the Handicap Reduction Table of section 10, members of the HRT (Knuth, Scheid, and Engel) used integer probabilities in the calculations. It seems unlikely they would go to the effort to compute average probabilities for which there is no obvious use.

To settle this matter, I am forwarding our correspondence to Dean Knuth, former Senior Director of Handicapping at the USGA. If Mr. Knuth believes the revised Appendix E is an accurate reflection of the work of the HRT, then I stand corrected and apologize for my nuisance.

Regards,

Larry

Note: As of this writing (4/11/2012), neither Mr. Knuth nor the USGA have responded.