Truer words have rarely been spoken. FlyOnTheWall has some valuable advice to pundits who want to pronounce on every shift and change, whether it’s politics or markets:

It’s a problem endemic to real-time coverage of campaigns. Many things become clear only in retrospect. When the longue durée is measured in days, not centuries, we often mistake signal noise for significance. Think, if you will, of the coverage of the stock market – you can get updates every 30 minutes, explaining in detail why the market is moving as it does, even though expected levels of volatility alone can explain most of these movements. I can tell you with a high level of confidence why the market moved as it did over a decade, a year, or in many cases even a quarter, but only a fool would claim to know why it moves as it does every half hour. These reports tend to focus on whatever economic indicator or earnings report has just been released as the causal explanation for broad movements. In a similar fashion, shifts in opinion polling tend to be ascribed to the stories that are dominating a given news cycle, and those stories receive much of the attention they do because they are presumed to be capable of shifting opinion. But the laws of probability tell us that some of those apparent shifts are bound to be illusory. Others may be the result of grass-roots organizing or social networking. And some will simply stem from voters becoming better acquainted with the candidates. Sorting the wheat from the chaff is no mean feat, but that humbling truth doesn’t seem to impact coverage.