Know a good forecaster? – The Hindu

Some predicted the outcome of the State election in Tamil Nadu. Likewise, there are some whose predictions of market crashes turn right. The query is: Are some always good predictors of future events? Here, we discuss what you must be mindful of before you decide that an individual or a firm is a good predictor or a good forecaster.

Consistency factor

Statistically, when many professionals and firms engage in forecasting, some of them are bound to be right. After all, there are only three ways a market can move — up, down or stay flat. So, what matters is not about whether an individual’s or a firm’s view was right this time. Look for consistency in their predictive (forecasting) abilities. This is crucial because both luck and skill play a role in any outcome. Prediction (forecast) can be correct once or twice because of luck. But getting the outcome right consistently needs skill, in addition to luck. So, you should check how many times in the past an individual’s or a firm’s prediction (forecast) turned wrong.

Moderate confirmation bias

This is important to moderate confirmation bias (refer to Wason Selection Task to know more about this bias). This refers to the systematic error we make when we conclude based on our preexisting beliefs. For instance, we may conclude that if a professional correctly forecast the market outcome in the past, then he/she must be a good predictor or forecaster. So, we tend to act on their next prediction (forecast) or at least believe their view will come true, even if there were evidence to the contrary. Market is driven by human behaviour — individuals’ interpretation of news and events. And human behaviour is driven by biases — errors we commit during a decision process. When several biases inter-react, we may not necessarily react the same way in the future when faced with a similar situation as in the past.

Conclusion

It is moot if AI can consistently predict human behaviour accurately. While AI can capture market trends faster, you may still have to take a meaningful decision to act on the AI-generated reports. So, biases could still matter and impact the market. It is easier to attribute an individual’s correct prediction to skill. The point is both luck and skill are important. Should you not consider these factors before deciding whether a professional is a good predictor (forecaster)?

(The author offers training programmes for individuals to manage their personal investments)