Slashdot Mirror
Web Site Selling "Earthquake Forecasts"
waytoomuchcoffee writes "The San Francisco Chronicle is running a story on geoForecaster.com, a site that offers 'earthquake forecasts,' for a fee. California is looking into claims that the site is practicing geology without a license."
Unfortunately, the reality is increasingly that many in the scientific profession achieve success by attracting public attention, the public often being a poor judge of true innovation. Why? Because if you aren't making wild claims, CNN just doesn't care, and how does a Professor that has made a genuine contribution to their field compete with an idiot that is on CNN every second day?
There are those that have made a career out of telling the media what they want to hear. People who gladly accept publicity even when their self-aggrandization hurts serious research in their field.
For the perfect example, learn a little about the career of Kevin Warwick, the UK's foremost pseudo-scientist.
Science and academia are increasingly a joke. For some time now, it has been more about public image than genuine contribution to the human understanding of the world around them.
I'm skeptical, but it seems to me as if they've made it easy to test the accuracy of their claims. All we need is for one person to subscribe to their service, record all of their predictions for a few weeks, then compare it to public earthquake data from USGS.
Be sure to check that they don't change any of their data after the fact - i.e. that their archive of past forecasts really does match what they predicted. Also, make sure that the "updates" they make to each forecast aren't too dramatic - if the forecast says that there'll be an earthquake here in one week, but tomorrow the forecast says it will actually be 300 miles away from here, then it's a lot less useful as a resource.
Reading through their site, they certainly don't show many of the typical warning signs of a scam. Sure, it would be nice if they published their methodology, but it doesn't really matter. We can test the accuracy of their system as a "black box" without their cooperation, simply by comparing their forecasts to reality.
That said, here are my main concerns:
1. They claim 90% accuracy of earthquakes magnitude 6.5 and higher. Their sample period is three years - how many 6.5+ earthquakes have there been since 2000? Also, does this mean that of all earthquakes that did happen, they predicted them with 90% accuracy, or that of the earthquakes they forecast, they were 90% accurate? With the latter interpretation, they wouldn't be penalized for earthquakes they didn't forecast at all.
2. They give themselves a near-perfect score if they underestimate the magnitude of an earthquake. Is this reasonable? Should they get credit for forecasting a 2.5-3.5 earthquake if a 5.5 hits? Or a 7.5?
3. After the first time they forecast an event (up to a year in advance!) they update their prediction daily. After the predicted time window has passed, do they score themselves based on the most recent prediction, or based on the first prediction? One could imagine that their methodology really does work - but only two days in advance. To make it seem like they can predict much farther in advance, they just make up random predictions and update them daily, changing the closest random prediction two days before a "real" prediction says an event will occur.
Have you ever tried talking to a reporter about something vaguely scientific? I agree with you that there are people in the public eye who make a living out of hyping up the media, but I tend to disagree with a lot of what you've said.
As someone who's had to talk to the media on several occasions about scientific subjects, I can say with some certainty that with very few exceptions, the media does everything they possibly can to sensationalise whatever information you give it. They can and do chop and change whatever you might say to put whatever spin they want to put on it, and there's absolutely nothing you can do about it. Welcome to capitalism, where populist media determines public opinion.
I've seen lots of friends get caught out by this. It's easy to read people quoted in the paper as saying something, and assume they were stupid to say it. In actuality it's much more likely that the reporter's chopped out every second word and rearranged some sentances to get a desired effect as well as completely and absolutely ignoring the 95% of your conversation where you stressed that whatever you said was excessively unlikely. They will have done just enough to have quoted you completely out of context, but stopped mind numbingly short of mis-quoting altogether.