Systematic Thinking in Investing

One of the key ideas for investing well, over a long period of time is to handle as much of thinking using structured methods. Leaving things open and to be decided by chance is an easy way for it to get affected by emotions, biases and the chatter in your head. A downside to using systems to invest is it is very easy to fool yourself that you have found a silver bullet.

“To a man with only a hammer, every problem looks like a nail” – Charlie Munger.

Systematic thought processes can be achieved in many ways. A well-known approach is having a toolbox of mental models. Another popular way is to use algorithms.

An algorithm is a well-defined set of instructions that can be executed to achieve a solution. Algorithms help make the thinking process more deterministic. These algorithms can be based on different heuristics. Screening is a common practice to filter out companies that satisfy different conditions. Usually conditions on various financial ratios. Filters in itself is an algorithm.

Checklists are probably the most simplest version of an algorithm. They are a very “analog”, pen-and-paper way of putting a system in place.

In the context of investing, systems like checklists and algorithms work better when they are not defined down to the last decimal point. There needs to be a variable part to the system that can change depending on the context and company in question. This is where investing turns into something of an art. Knowing when to use what thought process to value a company.

In some cases, the system in place won’t have a variable part in it. And the complete process can be boiled down to a well-defined algorithm. Once the system reaches that state of maturity, computers can take over and do a much better job then us. Algorithmic trading and investing allows to push the boundaries from 2 perspectives. Trades can happen in a whole new time domain. Instead of timing in terms of time of day, trades can be orchestrated with a resolution of milliseconds. Secondly, the limit of quantitative analysis can be pushed considerably if we let computers take care of the complete process.