That is common a misconception.

Example : Throw a dice 1.000.000 times and calculate the average value per throw.

That value be very close to 3.5, but I assume everybody would have figured that out for themselves.

Now, does that mean that you are most likely to throw a "3.5" each time you throw the dice ?

Or how small is the error between any real throws and this average "3.5" ? It's error is always AT LEAST 0.5.



Another example : Nobody races the Supercat 17 design in any significant way anywhere. Basically you got one or two geezers pottering around the course on it maybe twice a year, beer cooler close at hand. We can gether their race data for a 1000 years and still not get an accurate rating for the Supercat 17 that way. Now you say, who cares about any Supercat 17's ! Then I say, how are the M18, M20, F14, and a score of other newly launched boats different ? How long would it have taken for the F16's to be rated right, we're still waiting for it in the USA. Maybe 10 years of waiting should do it ?


Some people should read up a little bit on rating history, there are good reasons why yardstick systems were abandonned in favour of measurement based rating systems. There is no point in making the same mistakes all over again.

Wouter