Sep 25, 2012, 11:41 AM
Post #6 of 7
From John Clauser and Bobbi Tosse
Re: [dobbsdavis] HPR Dual Scoring Analysis, Rolex Big Boat Series
[In reply to]
Log-In to Post/Reply
NorCal PHRF ratings provide better handicapping for the 2012 Rolex Big Boat Series than do either HPR or IRC ratings.
Recently, discontent with the IRC rating rule has spawned the HPR rating rule. Dual scoring under both HPR and IRC of a very small subset of the 2012 Rolex Big Boat Series (Div C, with only four boat types) was offered as a possible test of the new rule. The results were reported recently in Scuttlebutt by Dobbs Davis. He noted a “dramatic compression (20 – 25%) of the corrected time deltas in HPR compared to IRC”. He also reports a “less dramatic compression of the corrected time deltas” for “four HPR-style boats in the IRC A class”.
It should be noted that the NorCal PHRF system was used prior to 2000 for handicapping the StFYC Big Boat Series. It was therein rebranded as the StFYC Handicap System (to allow minor adjustments to it by StFYC). Following a squabble between US Sailing and RORC over IRC, IRM and IMS usage for international events and the Admiral’s Cup, the StFYC Handicap System was replaced by IRC for BBS handicapping, and has been used thereafter.
Sadly, the popularity of IRC on the Bay has significantly and continuously declined since then, as has the number of BBS handicap entries. Moreover, the StFYC is the only club on San Francisco Bay to offer IRC handicapped regattas. In 2012, it offered 3 IRC events -- the Rolex BBS drew 24 IRC boats, the Phyllis Kleinman Swiftsure Regatta drew 8 IRC boats, and the Aldo Allesio Regatta drew only 3 IRC entries whereupon its IRC class was canceled for lack of interest.
On the other hand, PHRF events thrive on the Bay. For example, the 2011-2012 Corinthian YC’s 4-race mid-winter series drew at least 35 BBS-qualified entries. The authors’ limited survey of the owners of a few boats who eschewed entering the 2012 BBS revealed that they felt that IRC gave them grossly unfair handicaps, whereupon they believed that entering was pointless.
With the declining popularity of IRC in mind (with respect to NorCal PHRF), it is perhaps worthwhile to consider what the results of the 2012 Rolex BBS would have been, had it been instead scored using NorCal PHRF (i.e. the old StFYC Handicap System), and to compare the corrected time deltas for IRC, PHRF and HPR. As with Dobbs Davis’s study, we shall use the statistical spread of corrected time deltas as a “figure of merit” for the quality of the handicap system – a narrow spread indicates a fair system, while a wide spread indicates an unfair system.
It is thus the purpose of this note to present the results of this comparison. For improved generality and statistical accuracy, we include all IRC BBS entrants in this study, and not just the very small number of HPR boats considered by Dobbs Davis
Table 1 (BBS_IRC_spreads.pdf) shows the IRC-handicapped results for each boat in each of the seven races of the series. To account for the fact that different divisions sailed different courses in each race, and to allow a cross comparison of the different divisions, each race result presented in the Table is the associated boat’s corrected time divided by the associated race length. Then, for each race and each division, the associated unbiased standard deviation is calculated as a measure of that division’s spread of corrected time deltas for that race. Finally, (in the last column) for each division, the average standard deviation is calculated as a measure of the division’s spread of corrected time deltas for the seven race series.
Table 2 (BBS_PHRF_ spreads.pdf) shows the same results obtained using the NorCal PHRF ratings. Notes to the table indicate assumptions, adjustments, and/or corrections regarding the PHRF ratings therein used.
Table 3 (BBS_HPR_ spreads.pdf) shows the same results obtained using HPR ratings, i.e. Dobbs Davis’s calculation here allowing for a variation of race lengths, as discussed above.
Several conclusions may be drawn from a casual perusal of the results presented in these tables.
- A comparison of the HPR results of Table 3 with the Div. C results of Tables 1 and 2 shows that the compression of corrected time deltas in HPR compared to IRC (as noted by Dobbs Davis) is almost exactly the same as that observed under PHRF handicapping compared to IRC. That is, HPR and PHRF give equivalently improved results for Div. C (about a 20% reduction of the spread).
- For Div. A as a whole, PHRF provides a whopping 51% reduction of the corrected time spreads, relative to IRC. (Note that Dobbs Davis reports a “less dramatic reduction” of the corrected time delta for his Div. A 4-boat subset than that which he observed for Div C. (15-20%).
- For Div. D, PHRF provides a 5% reduction of the corrected time spreads, relative to IRC. Curiously, for Div. B, PHRF provides a 20% increase of the corrected time spread. However, Div. B is the smallest of the four divisions (only 5 boats), whereupon its results are presumably less statistically significant. On average, for all four divisions, PHRF provides about a 14% decrease in the corrected time spreads, whereupon it may be judged as the fairest among the three rating systems considered.
- Neither HPR nor PHRF provided different division winners with respect to IRC. However, under PHRF, Vesper’s Div. A win is now decided via a tiebreaker with Swiftsure. As might be expected, both HPR and PHRF, however, do shuffle some of the subsequent positions relative to those of IRC.
All three rating systems (IRC, PHRF and HPR) offer a single number rating. Another noted advantage touted for HPR over IRC is that HPR is a measurement rule, while IRC involves subjective judgments in the assignment of its ratings. A noted disadvantage (to many) of HPR is that it is self-admittedly strongly type forming. If a single-number-rating measurement rule that is not type-forming is deemed appropriate, perhaps ORR-16 knots WL might be considered as appropriate for typical BBS conditions. Note also that Transpac and Pacific Cup have both spawned a significant number of available ORR certificates for west coast boats in the BBS size range.
A noted advantage of HPR over IRC is that HPR is not a secret rule, as is IRC. IRC’s total lack of transparency fosters a mistrust of the rule’s uniform application. To be sure, one may purchase IRC secrets at about $120 per secret for a 40’ boat, but no purchase of a boat’s base rating decision criterion is offered. By contrast, both ORR and NorCal PHRF are far more transparent in their assignment of ratings. Last year’s IMS/ORR VPP is available for purchase, so that its whole rating process is available for scrutiny. Also, to the credit of the NorCal PHRF committee, its meeting minutes are open for perusal on-line.
In summary, it appears that NorCal PHRF provides improved handicapping for Rolex Big Boat Series events. PHRF appears to be every bit as good as HPR, but has the advantage that it can be applied to the whole fleet, rather than to a very limited subset of the fleet. It is also far more popular locally, and undoubtedly offers increased attendance of handicap boats in the event. Alternatively, if a single-number-rating measurement rule is deemed appropriate, perhaps ORR-16 knots WL might be considered as appropriate for typical BBS conditions.
John Clauser and Bobbi Tosse, 2012 Rolex BBS participants