Dartcast vs. Forecast

By
Elijah Rotholtz
– July 19, 2011Posted in: Technology

 

In Seattle, weather forecasts are notoriously ill-fated. In other words, they’re plain old wrong. This past week, I started a series of tests comparing the weather forecast to the weather that actually occurred, which I then compared to my own forecast that I created via bombarding a dart board with magnetic projectiles. After one week’s worth of tests, I found that a one-day forecast (a forecast retrieved in the early morning, for the rest of the day), has a 71.4% chance of being correct. A two-day (A day in advanced. e.g. from Wednesday morning for Thursday) forecast has 33.333% of being correct. A three day forecast has a mere 20% of being accurate, after that the likelihood’s go down to an atrocious 0% accuracy.

After the first week of tests, I have decided that only seven data points is not enough to conclusively determine the accuracy of weather forecasts. I have also concluded that restricting my data to one weather station (The Weather Channel), was not a wide enough experiment, for the next two weeks I will be comparing three different weather stations (The Weather Channel, NOAA, and Komo5) for a longer fourteen day period.

I’m also considering contacting the weather stations and requesting past forecasts, my only concern is that the weather stations will not be as forthcoming as one might hope. For an accurate test, I require both the incorrect, and correct forecasts.

If you want to send in your local forecasts and actual forecasts, I would be very grateful for your help.

What do you think? Tell us in the comments section below!

P.S. I’m also currently in the process of making an electric car. More updates on that soon!

Tags: , , , , , , , , , , , , , , , , , , , , , ,

About Elijah Rotholtz