The Danger Of Extrapolation

The trend might be your friend but paper is still flammable. Ignore the non sequitur for a moment and consider this. Does anyone crossing the road expect to get hit by a car? No. But it still happens. But unless the person getting hit is incredibly unlucky, paranoid, or reckless beyond all reason, they'll never project themselves to ever get hit by car anytime in their lifetimes. That doesn't mean it can't happen.

The Uneducated: That seems like a good investment.

Extrapolation is the art of projecting the future given your prior results. Sometimes it works, other times it doesn't. Sometimes the projection is off by a little, other times the projection is off by a lot. Nobody can tell the future. But that doesn't stop people from trying.

As someone who has studied computer science and has taken multiple courses in statistics, I'm very familiar with techniques used to approximate the future. All of them unfortunately require data from the past and not data from the future. To be fair though, getting data from the future is currently impossible, so the past is the best we can do.

But people want to know their future, so we use the past as it is a pretty good heuristic most of the time. For example, if someone has an allergic reaction to shellfish, in the future, they can reliably guess that if they eat shellfish again they will have another allergic reaction, thus they avoid eating shellfish as they can extrapolate the probable result if they went down that path. We use extrapolation everyday to make decisions.

But extrapolation always come associated with the risk of being wrong. The danger of extrapolation is that the observable trend is not indicative of the underlying behavior of the thing in question. For example, if you took daily temperature readings in the northern hemisphere from February to June, the trend would suggest that the temperature would keep going up forever, when clearly we know this is not the case.

But in the prior example, we knew additional information that made that projection look foolish. In a lot of cases, we don't have that additional information. In general, we will never have enough information to predict the future with certainty. The future is dependent on the past, but the past can never be perfectly represented in any model.

But we already know this. There is a reason people buy insurance. Because you never know when you are going to get hit by a car or have a heart attack. The low probability black swan event can always pop up out of nowhere and destroy your expectations, your dreams, and your cryptocurrency portfolios.

So what did we learn? You should stop trusting anything that the technical analysis preaches. Now I realize there are a lot of believers of certain areas of technical analysis out there including the cult of Elliott Waves. But realize all that technical analysts are doing is extrapolating using past information. Given that the entire cryptocurrency space is so young, we don't have a lot of prior data to work with.

So, we really have no idea where this whole market is going. We are simply extrapolating based on prior data and then adding the additional biases of our hopes and dreams. To put it more simply, you can't predict the future with certainty and your emotions tend to make your predictions more inaccurate.

I have always been a fan of hedging and protecting yourself from uncertainly since all predictions are probabilistic in nature. The technical analyst knows this, but often times his audience does not. Unfortunately, sheep are colorblind. They see things in binary vision. Black or white. Good or bad. Up or Down. Bull or Bear. But that means when you are right, you are right. And when you are wrong, it really hurts a lot.

So, let's stop being so reliant on extrapolating candles and spend some time building better models doing some research. Let's take partial positions and hedge to avoid catastrophic loses and still make gains from the technology we believe in.

Or you could go all in. They call that gambling. Putting too much faith in extrapolation (or fate). And more often than not, gamblers end up being losers since the game is probabilistic. Sure, you might have the most chips at the table and are on a winning streak. But play long enough and you'll be sure to see a black swan at the table, and you'll be helpless to escape.

Sources:

Image