Last week, John challenged me to comment on the WSJ Numbers Guy column on the accuracy (or inaccuracy) of weather forecasts. In his columns, the Numbers Guy writes:
It turns out that, despite the city-by-city variations, the national average accuracy ratings for the five major providers are all pretty close.
and
Forecasts are considered accurate if they come within three degrees of predicting the high and low temperatures for a given day.
In John's words "If you don’t post something on this I’ll be very disappointed." Never being one to disappoint, here we go.
Gregg Easterbrook (Brookings Institute) writes a somewhat irreverant football column every week for ESPN.com's Page 2 (called Tuesday Morning Quarterback). This week he uses a profoundly simple football analogy to perfectly sum up my views on economic and weather forecasts:
Now, for my off-price generic forecasts. First, I predict that every NFL team will end the 2006 season with the same record as it did in 2005. Obviously this won't be right, but will it be closer than the countless pseudo-scientific forecasts floating around? I bet if you analyzed the last, say, 20 years, endlessly predicting every team would finish with the same record as the previous season would do you better than actually thinking about your prediction.
So here's my view on economics forecasts using modified Easterbrookese:
Now, for my off-price generic forecasts. First, I predict that every [quarter's GDP growth rate will be the same as the previousl quarter's real growth rate--or whatever economic indicator you want to predict]. Obviously this won't be right, but will it be closer than the countless pseudo-scientific forecasts floating around? I bet if you analyzed the last, say, 20 years, endlessly predicting every [quarter] would finish with the same [growth rate] as the previous [quarter] would do you better than actually thinking about your prediction.
And weather forecasts:
Now, for my off-price generic forecasts. First, I predict that [the high temperature tomorrow will be the same as the high temperature yesterday]. Obviously this won't be right, but will it be closer than the countless pseudo-scientific forecasts floating around? I bet if you analyzed the last, say, 20 years, endlessly predicting [daily temperatures would be the same] as the previous [day] would do you better than actually thinking about your prediction.
