It's worth asking - is the oil futures market broken? Or - perhaps the oil futures market is not broken and instead is increasingly influenced by drivers which are difficult to detect using conventional techniques.
The best examples are the simplest examples.
In Front of Our Noses
Dogs have approximately 300 million receptors in their noses, or 50 times more than humans. In addition, the part of a dog's brain that is responsible to analyzing smells is about 40 times greater than ours.
So when a human cannot detect what a dog can smell, it would be prudent to say that the real problem isn't a broken sense of smell.
“Between June 2014 and January 2016, WTI front month contract decreased from $110 / barrel to $28 / barrel. Almost every expert indicated that the oil prices hit the bottom - before oil prices hit the bottom"
Despite human experts continuing to suggest that the oil prices hit the bottom, TerraManta machine learning software generated a much more accurate price forecast on December 24, 2015 showing the WTI front month contract was expected to reach $28 / barrel - 2 weeks in advance.
TerraManta machine learning software generated more than 30K price forecasts to-date. And both the number of forecasts and accuracy of forecast are always growing.
Analytics and Forecasting of the Oil Market Need to Follow Suit
On February 20, 2018, commodity research team at Citi Group - led by Edward L. Morse - indicated in the research newsletter ...
“Couple this with a record level of speculative positioning on crude futures that is fast-moving and becoming increasingly responsive to data, trends and momentum and it makes for a market constantly on the move—analytics and forecasting of the oil market need to follow suit."
That's exactly right. Current techniques to analyze and forecast oil markets continue to be less and less effective.
So What's Broken: Oil Futures Market or Analytics?
Humans can only see the visible portion of the total electromagnetic spectrum. Less than 1% of all light that reaches us is in the visible portion of electromagnetic spectrum.
If humans decided to interpret the inability to see x-rays as a broken vision, x-ray machines which are immensely valuable probably would not be created.
But they did not.
Just because something cannot be seen, it's not broken. Different approach and techniques are needed instead.
TerraManta has been pioneering and successfully applied machine learning to forecasting commodity futures since 2015.