Pundits pose as news reporters

Schweber: Just because you can explain something and come up with an alternate calculation mode doesn't mean your theory is correct.

I've pretty much refrained from tuning in to the news in the mainstream media, but not because so much of it is bad. Instead, it's the timeline perspective that infects too much of the so-called reporting that I can no longer stand.

Call me old-fashioned, but I thought the role of reporters and journalists was to bring us the news, and by "news" I mean reports on what has recently happened or is happening now. Yet I find most news reports these days are really predictions about the future: the market today is poised to go up or down, they tell me with my breakfast. The same pundits who a few months ago said that oil could easily soon go to $200 per barrel are now saying, without any shame or humility, that it could easily go down to $50 per barrel. Thanks for wasting my time, I say.

Most of these talking heads are no better than the lady with turban and crystal ball who hangs out a sign that reads "fortunes told here" except they wear more expensive clothes. Few people take her seriously; so why should we give these often-wrong pundits any credence?

Einstein's discovery
The science and engineering worlds, in contrast, have quite different and increasingly strict guidelines for predictions. I'll use my favorite subject to showcase the works of Albert Einstein. In 1905, he published his first paper on relativity and at the same time a seminal paper using thermodynamics to explain Brownian motion. At the end of the paper, he took well-established Brownian-motion data and used his theoretical conclusions to calculate a value for Avogadro's number that agreed closely with this popular parameter's value.

Certainly, it is impressive. Yet this predication is only a necessary, but not sufficient condition of validity: Just because you can explain something and come up with an alternate calculation mode doesn't mean your theory is correct.

At the next level is an explanation for something that is not well understood or still puzzling. Einstein used his general theory of relativity to account, with great precision for the discrepancy of the precession of the perihelion of Mercury, which had been carefully tracked for years. Even after astronomers had taken into account all the second- and third-order interactions between Mercury and other planets, there was still a small, but bothersome difference of 43 arc seconds per century between the observed rotation of that planet orbit's closest approach to the sun, and what calculations said it should be. Einstein himself remarked that he was very pleased that he was able to account for the gap, because of his general theory.

Prediction with basis
Again, having an explanation that fits the data doesn't mean your theory is correct. A real test of it is to predict with accuracy something that no one has even seen or imagined. Einstein's greatest triumph was his prediction that gravity would bend light, and that the observed position of stars would be shifted from their usual locations as their light passed closely alongside the sun, for example. His prediction was confirmed by photos taken during a solar eclipse in 1919. For many scientists who were skeptical of the general theory of relativity, his "far-out" prediction and its confirmation were the final piece of evidence they needed to be convinced.

This series of confirmation steps applies to a more mundane world of circuit, system and software debugging, which is the hardest of all engineering disciplines. One of the first levels of debugging is to form an explanation for any observed behavior, and then verify it. However, the true test for a subtle problem is to be able to hypothesize that if an issue is the case of the problem, then there should also be another, yet unobserved, consequence and it should be verified. When you can do that, you can understand the cause of the problem.

- Bill Schweber
Site Editor, Planet Analog