Better Off Without It

Woah, unthinkable! U.S. GDP estimates were too high by 20%…that we know of. Incredibly, the number has been revised downward substantially. This fits nicely with my world view of statisticians being low-intellect monkeys who are too lazy to do actual investigative work. Mind you, I’m a student of mathematics, so I have no reservation in trashing some of the practices.

I have a nasty hunch on how this happened, and it all has to do with “trend analysis” and “time series.” I’ve watched the proponents of this crap first hand; I’ve sat through their lectures, read their books, looked at some of their work; and it is always the same error.

Statistics, in its truest, purest form, is simply a variation on counting. Statistics can be used to give a picture of current starting points, which is subject to error. When used appropriately, statistics can tell us how things look now…NOT what they are going to look like. The results of this exercise in linear relationships can be fruitful. However, trying to use statistics to smooth out complex behavior like that found in non-linear systems (stock markets, ecosystems, competition relationships, etc…) is a fools errand. The same problems inherent with analyzing those systems to begin with are still very much present when you try to, say, run regression over top of them…

You see this common misconception regularly, with people trying to use trends on statistics to predict things from population growth to temperatures to the cost of crude oil. It’s lazy.

Where a dedicated student would attempt to understand the behavior of the system – how things interact and what determines those interactions, using statistics only to calibrate and measure – a pure statistician takes some data, throws a mundane formula at it, and then declares all sorts of stupidity. The greek symbols have gotten prettier, but the logic and reasoning haven’t gotten better.

It makes you wonder why we let these estimates get reported at all? They aren’t useful. In fact, they build expectations and create wild mispricing. Would it have been so bad to wait an extra month and find out that GDP was only 2%, and not 2.5%? What about ADP’s huge mishap earlier this year with new jobs?

Are you all so desperate for an edge you’re willing to subject yourself to such questionable practices?

If you really want an edge, take a step back and stop taking these estimates for gospel. They’re so manipulated, so poorly conceived, so incredibly sensitive to error, and ironically so intensively followed, that one of the best things you can do to improve your performance is to tune them all out. Wait for the finalized results, and only pay attention to that.

The rash and immature chase mirages. The wise follow their nose to the water.

If you enjoy the content at iBankCoin, please follow us on Twitter