Expert predicts 14 named storms in 2007 ("above average activity")

AP on Yahoo ^ | 12/8/06 | AP

[Wormwood]

I have a 20 sided die, can I play in this game too?

No, because your 5% chance of sucecss may throw off the curve for everyone else.

Try this instead:

[Abathar]

Hehehe :-)

[Wormwood]

LOL!

You totally had me.

[COEXERJ145]

[Strategerist]

I'm going to wait to see what the folks from AccuWeather have to say. They got it right for 2006.

Huh? They were the biggest embarassing disaster of all.

[Retired Chemist]

AccuWeather predicted that the storms that formed in 2006 would move up the east coast and not enter the Gulf and that's essentially what happened.

[edsheppa]

Don't know if you all are interested, but Gray's forecasts for 2000-2006 didn't do any better than random forecasts. I got Gray's December forecasts for those years and actual data for 1995-2006 for NHC (I would have got more actual data but for prior years it's more difficult to pull out of the NHC web site). Then, assuming the named storms and hurricanes obey a Poisson distribution, I made a bunch of random forecasts, for each year using the mean of the actual data of prior years.

Gray's forecast for named storms was better ~28% of the time. His forecast for hurricanes was better ~49% of the time. (Better == smaller sum of squared differences)

Obviously the modelling isn't adding much value for Atlantic tropical storm prediction.