Monday, March 18, 2019

Get out of the binary world - Focus on probabilities, baby!


One of the key problems with decision-making is that it is often simplified into either/or choices. "Yes/no", "Go/No-Go", is how we often focus our attention and make decisions. Life is easy when problems are framed as either black or white. For example, the Fed will either tighten or not tighten. Employment will either increase or decrease. The stock market will either rise or fall. These are phrased, in the end, as binary actions. Seldom will you hear a market pundit provide anything other than a binary choice problem. Forecasting is often viewed as being so hard that getting just the direction right may be more than enough to be successful. Unfortunately, framing uncertain forecasts as a binary problem is both near-sighted and flawed. 

Thinking in a binary world does not allow for a richness of details and choice in forecasting. Thinking about forecasting in terms of probability is more important. Don't frame the problem as, "I think the Fed will be on hold." Think or believe in something like, "The probability of a Fed "no change" at its next meeting is 70 percent." Placing the forecast in terms of probabilities changes what action can and should be taken. This is even more critical when looking at questions that can have a range of possibilities. You could say, "The change of employment growth being below 180,000 and the market expectations is X% and above 180,000, (1-X)%. Probability estimates can be done for a range of employment numbers to form a distribution of forecasts. This is harder but it allows for more investment insight.  

The response to a forecast that is a flip of a coin is very different from a belief that the chance is 70% favorable. In the first case, it may not be worth taking a bet. In the second case, the size of the bet may be large because the odds are favorable. By thinking in probabilities, the size of the bet will be more effectively managed. This applies even to what may seem like a yes or no question. It takes time and effort to think outside a binary decision world; however, once a pattern of thinking is established, the process becomes easier.

Many argue that being right more than 50 percent is not the measure of a good trader. Good traders can have a success rate below 50% and still make money. That is absolutely the case. Good risk management can offset forecasting errors. Holding winners and cutting losers can allow for lower forecasting skill, but increasing forecasting skill will only improve and not detract from performance.  The best way to improve this skill is to focus on the odds of forecasting. As you receive new information, there is a shift in the probabilities, not just a shift to either yes or no. Stop binary thinking and focus on the odds.  

No comments: