Over the weekend, I completed the WxChallenge, a nationwide academic weather forecasting competition. I finished 134th nationwide for the year, out of 1424 full-year competitors including undergrads, grad students, and faculty, as well as pro weather forecasters (I am “mrg287”). I made the top 9%, and while I just missed making the ‘playoffs’, it was still a lot of fun and a great educational experience as part of the Penn State Online Weather Certificate Program.
The WXChallenge involved forecasting for a number of cities nationwide over the course of the academic year. Each city is focused on for two weeks. Each weekday, forecasters had to make forecasts of high temperature; low temperature; wind speed; and accumulated precipitation. Points were given for errors between your forecast and what actually happened. Forecasters were then ranked based on this score, so those who were closest to correct were ranked ahead of those who ‘busted’ – blew a forecast. After two weeks in a City, you’d move on to another one. The schedule for 2013-14 saw me forecasting from the southeast to the northwest and everywhere in between.
The Challenge forced me to get into a routine and to apply the knowledge I had gained in the Penn State program to actual, real-world forecasts. By moving around cities, I was forced out of my comfort zone into areas I might not ever have considered otherwise (Astoria, OR? Burbank, CA?). It also forced me to take into account game theory. The key to success wasn’t being 100% accurate – no one can do that. Instead, to win, you have to be “less wrong” than everyone else. Thus, there was a certain element of psychology to it.
For example, in a City which was expecting precipitation, there was really no way to precisely forecast an amount to 1/100th of an inch, as the rules require. Instead, we had to figure out (a) what was a reasonable amount to forecast; (b) what would the majority of the rest of the forecasters predict; and (c) given that some error was inherent, should I err high or low?
It was a great experience overall and I highly encourage anyone who has a serious interest in weather to try it out.