The Hurricane Gustav game has barely been closed and we’re already opening Hurricane Hanna. Fortunately, Hanna looks to be weaker than Gustav. It will be a fun forecast, though, since the angle it approaches the coast at makes it very easy to miss the landfall location (and thus time) by large values. Meanwhile, if you have suggestions on how to improve the scoring equation, post them here!
Contest link: http://funnelfiasco.com/cgi-bin/hurricane.cgi?cmd=view&year=2008&name=hanna
How have you decided the weighting of each of the 3 elements of the forecast? Time seems to be the most weighted. If you choose to keep the equation, perhaps you should indicate what the equation is BEFORE folks enter forecasts so they know which category to emphasize most. Otherwise, I think all 3 categories should be equally weighted.
Good point. For subsequent forecasts I’ll add the equation to the script so it gets printed at forecast time.
As for the weighting, it is partially because the units are so incredibly different in their effect (e.g. a 100 minute error won’t affect people’s plans, but 100 miles might, and 100 mph definitely will). So the idea isn’t so much to weight them differently, but to weight them equally by compensating for the various units. Admittedly, it’s all very arbitrary, which is why I’ll happily accept ideas for improving the equation.