The whole philosophy of using machine learning and artificial intelligence in investing is that machines can make trades quickly and without being burdened by emotions.
It’s the last bit that I’ll talk about today. Emotions — like fear and greed — skew the decision-making processes of the investor, causing him or her to but too high or sell to low based purely on feelings. A good example: you would probably pay more for food if you’re hungry than if you just had a big dinner.
Therefore — and this conclusion is inescapable — machines that do not have emotions can trade better.
Except, this logic is totally wrong. Well, not wrong. But it forgets the most important variable. The market and the economy runs on emotions. Even if every investor was an emotion-less cyborg, the rest of the economy, which underlies all stocks and commodities, runs on desires and emotions.
A successful investment system can not possibly rule out emotions. In fact, the most successful investment system would totally understand emotions, but not ruled by them. This system would understand how desire for something quickly turns to the desire to not lose something — fear. Likewise, it would be able to sense when the momentum of fear turns into a buying opportunity. It could sense that the greed in the real estate market was becoming too frothy, or that the fear of a recession was starting to bottom out.
In short, the best investment system understands emotions. It just isn’t ruled by them.
Can a machine understand emotion?
Not the current ones. However, it could probably see patterns of behaviors that are more predictive. The key term is “more predictive.” If we recognize emotions as one of the main propellants that fuel the market, we also have to understand that emotions can often be unpredictable in how they find expression. A scared investor may buy more stock, not less. A greedy investor might sell stock, not buy.
What do you think? How does emotion govern our markets?