The highest chess rating in the world is not a man, it is not a computer, it is a man + machine combination. Gary Kasparov (rated one of the best chess players of all time) created Advanced Chess which pairs man + machine because he believes it is the pinnacle of chess excellence. Peter Thiel, in his book “Zero to One” spends an entire chapter explaining the combinatorial excellence of man + machine over either man or machine. He went as far as to found a company, Palantir, whose genesis dates back to PayPal’s hybrid man and machine approach to fraud detection.
Excellence in decision making, in our current era, requires the combinatorial power of man and machine. However, in portfolio management today, human portfolio managers generally use some combination of heuristics, instinct, and crude Excel files to determine position size. To be truly world class though, a machine comparison is required to create the interplay between man and machine. The machine should be an engine that takes stock-specific inputs from the research process and combines them with other factors that the portfolio manager deems relevant to position sizing and calculates an optimal size. The machine-based optimal size gives the portfolio manager a point of comparison to their own heuristic-based position size. This controls for two major deficiencies in man-only decision making: 1) emotion can dominate when man is left unchecked and 2) simple errors of omission or tabulation are eliminated.
I’ll give quick examples of each to drive the point home. Recall the market nadir in early 2009. Emotions were super-charged and defensive posturing was the norm. Many Alpha Theory clients were struggling with the defensive emotion themselves but Alpha Theory was flashing buy. Clients applied very conservative assumptions and still saw green across the portfolio. One of the greatest buying opportunities of their career may have been missed without the machine as a backstop to emotional decision making. In situations of high emotion (i.e. things are going really well or things are going very poorly) humans have a tendency to make reactionary decisions that in hindsight, they recognize as poor. A machine-based partner provides an unemotional representation of what you have said ex-ante you would like to do and plays the “voice of reason” in emotional times.
Simple errors and omissions are more common than many of us would believe. Portfolio managers are, in general, intellectual elites. But no human can beat machines in pure calculation and repeated procedures. As an example, I was speaking to a client that thanked us for adding the Cost of Borrow feature. He described a short position that had the best probability-weighted return in his portfolio so he had a large position. He knew the cost of borrow was steep, but still heavily weighted the idea. After turning on the Cost of Borrow function in Alpha Theory, he saw that the probability-weighted return fell and made the short his 12th best idea. He knew all the numbers, he knew the math, but by making assumption explicit it made them objective so they could be judged and measured.
Machines are not a crutch. They are not a replacement. They are the unemotional, logic-driven partner that provides constant advice and feedback to hone our decision process. Machines are only getting smarter and learning to work symbiotically with them will be one of the hallmarks of great managers in the 21st Century.