prediction percentages once again very good.
Posted by
jimw on 12/7/2018, 10:54 am
Reality of statistics and how it has worked on my hurricane predictions just released stats for 2018 and comparing it to my averages.
#1)since 2003 taking between 140 to 156 cities & islands then narrowing it down to 20 or roughly 13% on average.
#2)Out of that 13% narrowing it down to top 5 averaging nearly 50% accuracy = practically a needle in a haystack. All 20 combined near 40% impacted by named storms usually within 60 to 80 miles.
#3)After a horrendous stretch in 2013/14 using Gray/Klotzbach's number of named storms since 2003 I decided to go it alone. Since 2014 I have been off by an average of ONE named storm.
#4)Since adding the NAO North Atlantic Oscillation my accuracy has increased 20% on top 5 each year this year at 100%.
#5)8 of 16 years have 60% or higher of top 5 impacted.
#6) This past season I left out cities on the FL peninsula & Texas coast neither were impacted.
Meanwhile agencies continue predicting GARBAGE as each season approaches using huge stretches of coastline with 40% to 60% chances & ranges of named storms such as 6 to 10 from NOAA the industry can do better than this. And lets obsessively talk about the ACE which means little to nothing for the average person living on the shores of the Atlantic basin.
hurricanecity.com/predictions
|
78
In this thread:
prediction percentages once again very good. - jimw, 12/7/2018, 10:54 am Post A Reply
This thread has been archived and can no longer receive replies.