I have personally a dislike for unsubstantiated modifications to any rating system. This means any adjustment for an old boat society (dead boat society) rating that is not specified as a general rule.

For example under the current texel system the Dart 18, Prindle 16 and Hobie 16 get their ratings upped by 1 point because somebody in the past decided that that was necessary. Why these boats and not others, what was the reason for this mod that was never specified in public ?

I can agree to such modifications to outdated designs to compensate for their relative backwards movement because the newer boats arguably have more refined features. However the selection proces needs to be general and verifiable.

Surely a case can be made that asymmetrically hulled designs like the P15, P16, P18, H14 and H16 suffer from not having boards or skegs. The idea behind the asymmetric hulls was an interesting one but we now know that it produced worse performing boats to symmetrically hulled ones with skegs or boards.

Therefor I propose that a list of 5 features (or lack of such features) is specified and that any design satisfying at least 3 of these features gets 1 or 2 points added to its rating.

I'm thinking off.

-1- Cat has assymetrical hulls without boards or skegs
-2- Cat has a mast unsupported by diamond wires or extra shrouds
-3- Cat has a pinhead mainsail
-4- Cat doesn't have a downhaul system nor a mast rotation system that can be continiously adjusted
-5- Cat doesn't have a smooth (rounded) transition from its sides to the deck (deck lips as on H16 and P16)

All these features pretty much add extra drag to the platform or limit sail control to such an extend that the rig is producing drive less efficiently than more modern cat designs.

This list effectively gives :

P15, P16, P18
H14, H16
Dart 15, Dart 18
nacra 5.0 (not the newer nacra 500) and Nacra 4.5 (not the newer Nacra 4.5

all an extra allowance on the ratings. I'm still debating wether this should be 1 or 2 points.

I do not favour or propose to have a statistical alogarithme produce offsets on calculated ratings. This is not verifiable nor easy in use. Also you get continious swings around a base rating depending on freak occurances in race data. Also this is sort of a black art and I'm not willing to saddle any rating committe official with the endless bickering that this so easily can cause.

Please note that statistical analysis may sound dependable and easy enough but often it isn't.

Example : A rating committee member when a race has been held under sufficiently constant conditions. Steeple chase 2004 contains very undependable data because of that freak low tide. A particular good race for the M20 in NL in 2003 was found to have contained a narrow wind band some 2 km out of shore that the M20's found together with one I-20 while the others tacked upwind closer to shore. Incidents like these very much skew data that very easily can cause an errornous swing in ratings. Checking all this is a [censored] load of work, I know because I don it. It involved number crunching and possibly phoning up participants to find out why the data does what it does. If anything this should be done on the background and no automatic link should be made to the rating systems.

Again when a large block of data over various class and events suggests that a general modification is off the a modification should be made, but never when such a thing happens in an individual case with a very small amount of data.

Simply put It is also very weird when say an F18 shows some exceptional behaviour that is not reflected in the behaviour of other comparable catamarans. Therefor all things are better incorporated as general rules. This may leave som inequality in the system BUT by far in most cases these are neglectable in magnitude and to small to sacrifice stability and simplicity for. Remember the less then 1/10 th of the end result threshold.

Nor am I sure such statistical systems can be made to work effectively. I wouldn't know how and I know some stuff about statistical and probability math.


Wouter