The LinStat models developed to date for Portsmouth Number equivalents and corresponding time correction factors have proved to work well. I am conscious though that many involved, willingly or unwillingly, in setting handicap numbers would still appreciate some sort of list to refer to. Gaps or uncertainty in user supplied boat data is often an issue. So, a project is underway to produce a database of sailing cruisers, identified by a range of key data, along with predicted PNs from a number of equally well performing LinStat models. The list will be confined to single keel boats initially and the aim is to have a first edition published on this website in time for the (hoped-for) 2021 sailing season. It has to be remembered, though, that no such list will ever contain every possible boat, which is one of the reasons for all the work in producing straightforward statistical formulae for handicap numbers.
In response to the desire of some users to include overall length as well as waterline length in their calculations, I have looked at producing an acceptable alternative to Model 10A. There are inevitably pros and cons in doing this, but one possible option is discussed in the new Article 2.5A
No single handicap formula will 100% capture all boats in all conditions, so inevitably ways to tweak formulae will be sought. There are several approaches to this, but care must be taken not to damage the statistical integrity of the formula. For example, it is inappropriate to replace LWL (waterline length) with some other definition of length without a re-assessment of the model coefficients. I will shortly publish an article on model variation, with particular reference to Model 10A (see Article 2.5).
There is a continuing problem with getting sufficient reliable performance data for bilge keel yachts on which to base refinements of my Portsmouth Number equivalent formula for these craft. So, I have decided to look at the problem from a different angle. Given that the formulae for fin keel boats continue to work well, my current quest is to look at designing an appropriate adjustment that will enable the same formula, in essence, to be used for all keel types. A significant advantage of this approach is the possibility of extending it to the TCF models described in Articles 2.6 and 2.9. I will report on progress in due course.
My thanks to the eagle-eyed reader who spotted a typographical error in Article 2.11. The suggested high-tech sail adjustment should of course be negative not positive for a PN type handicap. The correction has been made.
A new article has been added here, 2.11, on the subject of handicap adjustments. This covers some of the reasons and approaches to varying handicap numbers. Of course all systems need some review on a regular basis, but whatever the system to arrive at a ‘base’ number, there will be additional factors, not necessarily already accounted for, that might need some minor allowance.
As another sailing season draws to a close, it is time for more reflection on the performance of the LinStat models. As explained in a new article, the suite of models, though looking different from each other, all give similar results. Further, they can be applied equally well to different handicap systems.
A real practical issue is the accuracy of input data, as referenced in Article 2.9. More detailed illustration is given in the new article, 2.10. A further article will discuss a number of factors that those producing handicap listings need to think about in the context of handicap adjustments/allowances.
The relationship between the two measures of genoa overlap discussed in the second section of Article 2.9 has been amended slightly.
My current search for model improvements (however small) has focussed in part on assessing how a number of statistically equivalent models behave when applied to estimating handicap numbers. In particular, the sensitivity of results when the input data has some inaccuracy can point to the type of model components contributing to making a model robust.
Clearly, the effectiveness of any model relies on the accuracy of input data. Article 2.5 states the accuracy needed in the required ‘brochure’ measurements. To date, the most efficient models require extra information about foresail area and this may involve user measurements and estimation. Newly added here is Article 2.9 which discusses this issue, and articles 2.5 onwards have links with 2.9.
The most recent articles to be added here are 2.8 and 3.2. New readers should refer to the last posting for a description of how the articles are organised. Work is ongoing to increase the amount of reliable boat data available for model development and to search for any worthwhile model improvements.