Half a century of forecasting science


00:08 Introduction
00:32 Spyros, perhaps you could start by telling us a little more about your background?
01:36 What about the last 50 years of forecasting science?
05:25 Can you take us back to when you started as a young student?
07:19 What does the M5 competition mean to you?
11:09 How did the introduction of computers change the way you were doing things?
12:50 How have the forecasting techniques evolved through the decades?
15:00 How does the approach you take in a test scenario vary from what you would do in the real world?
17:17 How has the perception towards forecasting changed throughout your career?
19:18 Would you say that the perception has shifted so far that people now expect too much from forecasting?
22:19 Do you agree with the fact that sometimes consultants promise too much concerning forecasts?
26:19 Spyros, if you look back at your career, what are you most proud of?

Description

When it comes to forecasting, we often take it for granted that there are already techniques that have been fully tried and tested. Our guest for this episode didn’t have that luxury. For this episode of LokadTV, we welcome Spyros Makridakis, one of the forefathers of forecasting, to discuss its evolution over the past 50 years.

Spyros is a professor at the University of Nicosia (UNIC), Director of the Institute for the Future (IFF), as well as an Emeritus Professor of Decision Sciences at INSEAD where has won the “Best Teacher Award” twice. He is the organizer of the “M-Competitions”, which are named after him (the Makridakis Competitions). In addition, he is the author of multiple articles and books that have sold thousands of copies, such as “Forecasting, Planning and Strategy for the 21st Century”, “Forecasting: Methods and Applications” and “Forecasting Methods for Management”. On top of all this, he was also the founding chief editor of both the Journal of Forecasting and the International Journal of Forecasting.

We talk about how often it’s the simpler forecasting methods that are far more accurate and effective than the more “sophisticated” methods. This is something that the M-Competitions have often highlighted. Spyros tells us more about how forecasting gradually became accepted as a science and the general acceptance that you cannot predict randomness, as well as the need to separate the past from the future by creating a model that is adaptive. At Lokad, we argue that uncertainty is something to be embraced, which is something that Spyros agrees with, but is something that many potential clients have difficulty with. We explain why a forecast that has “99% accuracy” that some vendors claim is simply not realistic and why many consultants promise too much when it comes to forecasts.

To wrap things up, we talk about how to take into account extreme events (such as pandemics) and make forecasts more resilient to such extremities. Spyros also reflects on his career and tells us a little about what he’s learned.