On Wednesday, February 12, 2014 8:55:21 AM UTC, David Mitchell wrote:
Odd, that's what I've always said too. What I do better than others is to interpret the models to say what they are likely to be correct at any one time. (That's what you've never said, but is what actually happens).



No, I've always said that it is you interpreting the models.

Then we agree. That's what every forecaster does and I am no different. Without models, forecasting at 10 days would be utterly impossible, as it was 50 years ago.

I love the way that you say that matching my record is not difficult, but you haven't actually done it, because you haven't got the time! You must be very good at knowing when the models are consistent to be that confident. Without ever having done this, how could you possibly know!

I get it wrong around 20% of the time. You are right, one learns from one's mistakes and I have honed the technique over time as a result of that and the accuracy has increased from around 75% to over 80% in 7 years. I do forecast when the models are inconsistent, by default. I say that they are not to be trusted to 80% confidence, when I don't forecast, as long as that is 3 days after my last forecast.

One day you will actually try to understand what I do, or not; it won't make much difference, but do allow me to put you right on your misconceptions about what I do! *))