Wrong Tomorrow
April 1st, 2009
Wrong Tomorrow collects predictions made by public figures, be they politicians or pundits, and notes the timeframe in which the prediction is supposed to come true. The idea is that as predictions fall due they'll be tagged as right or wrong, thereby revealing who really knows what they're talking about.
It's a nice idea, but despite the site's claim that each prediction needs to make an empirically testable claim about the world
some of the predictions are too ill-defined to be testable. For example:
author: matt simmons
prediction: "We are three, six, maybe nine months away from an [oil] price shock."
Is there a formal definition of an oil price shock
? If prices suddenly double over the course of December 2009 that would presumably qualify, but what if prices increase gradually over the course of the next three/six/nine months? What if prices go up suddenly, but only by 15%?1
[Via Daring Fireball]
April 3rd, 2009 at 12:32
I'm the guy who runs the site.
My criterion for judging these predictions is what a disinterested layperson would believe them to mean at the time the prediction was made. In the case of an oil shock, there is a helpful definition by the Cleveland Federal Reserve: "a sudden and sharp
increase in the price of oil and its derivatives like gasoline".
There is still ambiguity in what "sudden" and "sharp" mean, but the prediction isn't completely untestable. There are clearly outcomes that would invalidate the prediction (steady drop in oil prices) and confirm it (spike to over $100/barrel). There's also a gray area where the correctness of the prediction depends on one's interpretation.
I think due to the nature of predictions this kind of waffling and fuzziness is inevitable. My hope is enough of the outcomes will be clear to render the issue moot (but I'm open to changing my editorial criteria if problems of interpretation start to bedevil the site).
April 4th, 2009 at 13:41
I'll concede that if the pundits' predictions are vague there's little option but to fall back on commonsense interpretations, but I suspect that over time so many predictions will fall into that grey area that any conclusions will themselves be open to interpretation. I hope I'm wrong, but only time will tell.
Take, for example, this Thomas Friedman prediction from March 2006: "I think we are in the end game. The next six to nine months are going to tell whether we can produce a decent outcome in Iraq."
You've marked that as Wrong. Is it wrong because by March 2006 it was already too late to produce a "decent outcome" in Iraq? Is it wrong because the middle of 2006 turned out not to be so crucial after all? Insofar as US troops are still in Iraq today and the country's government has yet to become a terrorist-supporting hotbed of Islamist extremism, does that not qualify as a "decent outcome" from a certain political perspective? Is it wrong because that period did in fact prove to be crucial, whichever way you view what has happened since? The bottom line is that any given reader's view as to what exactly amounts to a "decent outcome" will colour how they view Friedman's prediction.