Icon Class Association

I guess not, people just want to argue the toss about PY some more....again

Originally posted by jeffers I guess not, people just want to argue the toss about PY some more....again

Well if all was well there wouldn't be an argument, but it aint and the argument has gone, in my time here, from moaning about the odd lipstick bandit, to questioning the statistical validity of the entire system, now that computery is involved.

Originally posted by Blue One We already have a fact based system.

Just because it is based on facts doesn't mean it's accurate.

Women who own a Prada handbag live 8 years longer than average. Does a Prada bag make you live longer? Of course not - if you are rich enough to own one you probably eat well, are image conscious and have good health care.

Preganancy in the UK is officially 40 weeks (280 days). In France it is 41 weeks (287 days). We have oodles of data for that, and we still can't agree on a number!

A Swedish study of 427,581 healthy 'full term' births established the mean gestation to be 281 days, the median to be 282 days and the mode (most common value) to be 283 days. Only 70% of births happen with 10 days of the mean.

Another recent survey of 125 women in th US found the avereage to be 268 days.

So, does pregnancy last 268, 280, 281, 282, 283 or 287 days?

This is a trivial question compared to the PY of a dinghy - there is loads and loads of data and lots of well funded research and yet the answer is still unclear.

If you are basing a PY on a sample of 1000 (and the Icon is substantially less than that), there is no chance of being able to confidently and accurately pick a number that is 'right'. My guess is that if you are within 5% (50 points) you are doing well if you are basing purely on race results received.   

That's partly my point Jimbo - 969 is within the error that you might expect. Possibly on the low side, but well with a sensible range.

But I actually don't think you can be very much more confident with most other classes. The sample sizes are so small, and the underlying detail so wide ranging that the knowing what is signal and what is noise is very hard.
If the icon handicap was 97 when perhaps it should be 99 or 100, that doesn't seem so bad, pyschologically. Bring back 3 digits I say!