PlayStation's core crew spent weeks in Manhattan's Chelsea Hotel waiting for the perfect forecast.
A perfect forecast for wind, if it represented 20 percent of the power supply, would save $1.6 billion to $4.1 billion a year, according to several studies.
But that science, he emphasizes, will never offer a perfect forecast; the atmosphere, with complex links to the sun and oceans, will long remain a source of surprises.
A perfect forecast would require not only a perfect model, but also perfect knowledge of wind, temperature, humidity and other conditions everywhere around the world at one moment of time.
Some have argued that the above figure should have a downward trend because of improvements in weather forecasting whereas we interpret the experience of the 2009 Victorian fires in which 173 people lost their lives as suggesting that even with near perfect forecasts there is always the possibility of large numbers of fatalities.
BACK in the spring, James Hamilton provided the economics world with one of the more interesting results of the crisis and recession that if you took a macroeconomic model and plugged in observed oil prices through the middle of 2008, you got a near perfect forecast for output in the third quarter of 2008, nevermind the financial crisis.
Thanks to computer models, the odds on a freakish storm or earthquake are better known, epidemiologists are more successful at tracking diseases, and even man-made crises such as banking debacles and stockmarket crashes can be catalogued and studied to produce better (though, as we shall see, still far from perfect) forecasts.
Dodging and craning, peering around the silhouetted sighing, smiling, and hairy Harnacks, we caught occasional glimpses of the classic doings onstage, but came home talking mostly about the nearer and steamier events of the evening — a perfect forecast of a happy marriage, which lasted, according to Ms. Calisher's obituary yesterday, for another fifty years.
A perfect forecast would have a skill score of 1.
If we could make perfect forecasts, then y f (t) = y(t).
IC provides a relative measure of forecast accuracy in terms of deviation from the perfect forecast (Henderson 2010).
