A Model Is Just a Model

I'm eventually going to follow up on my second entry on this blog, about complex systems and our inability to fully understand them, and in many sad cases in the financial markets, safeguard against outliers.

For now, though, I'm going to lead with a quote that I came across on the Long Term Capital Management (LTCM) Wikipedia page. It makes the point I wanted to initially make in just a couple of sentences, as good quotes often do (I was taking way longer to get my point across so I scrapped that in favor of this quote). It has to do with models (in this case, financial models, but it really applies broadly to anyone who relies on models as a part of their practice--whatever that practice may be). Here's the quote:

"Economist Eugene Fama found in his research that stocks were bound to have extreme outliers. Furthermore, he believed that, because they are subject to discontinuous price changes, real-life markets are inherently more risky than models."

The case of LTCM should serve as a cautionary tale for failing to understand that you don't know what you don't know. Often times, very smart people are more susceptible to this failure of understanding, and the guys at LTCM were some of the smartest guys in the room--any room.

But this isn't just about financial modeling. 

When it comes to programming for strength and conditioning, the predicted results (based on what the model suggests) are often different than the actual results, sometimes on the downside and sometimes on the upside. This doesn't necessarily mean the program wasn't good. It just means the model wasn't robust and there were factors that either weren't accounted for (that should have been) or there were factors that couldn't have been known. 

But going down the rabbit hole of trying to find ever-more-robust models (or programs) is a fool's errand, in many cases. I see this all the time with newer coaches. They're obsessed with finding the perfect program and often make things needlessly complicated. The best programs match goals with processes, and are flexible enough to make adjustments along the way. You also have to be patient and let the program play out. Look for trends but don't make wholesale changes to the program hastily. And always keep long-term development in mind.