[email protected]

Experts’ panels doing better job predicting virus growth than data models.
20 Apr 2020

Experts Beating Data Models in Predicting Virus Growth

Predicting virus growth tells us much about the pluses and minuses of data-based decisions. That’s because panels of virus experts consistently do better than the best computational models at predicting virus growth.

Predicting Virus Growth

To be clear, many experts have their own models. Yet, they still integrate their experience and analysis into a final prediction. Moreover, by tapping the wisdom of crowds their predictions are integrated with other experts. This yields a final prediction.

Predicting virus growth via data models works in the same way but without the experts. Data modeling experts weigh the collective wisdom of the various data models. They merge this wisdom into one data model or prediction.

Predicting virus growth shows us the limitations of basing decisions solely on data.

Examining the success in predicting the growth of viruses highlights ways to best integrate data into our own predictions and decisions.

Problems Predicting Virus Growth

The problem predicting virus growth comes because viruses don’t always behave the same way. This is especially true of mutations of existing viruses and true of new ones like the coronavirus.

Whether it’s the historical data behind data models or the experience behind experts, newness raises uncertainty. Generally, the newer and more unique any event is, the harder it is to predict. Viruses are no different.

Making Predictions and Decisions

In the end, predicting virus growth sheds light on how to make better predictions and decisions in our businesses and lives. That light says that unless the future will tap the same trends as history did, data will need expert human input. Data-driven decisions alone won’t yield good predictions and decisions.

Now, those who think simply, in black-white terms will see this as reason for overriding data. That’s false. Again, think of the virus. Experts use their own models and overlay their expertise.

In fact, the biggest obstacles to data are human biases. This is especially true if the data conflicts with people’s views. It’s true when stories dominate thought processes. Stories beat data in the space of influencing humans. Stories form some of the greatest defenders of the status quo.

Thus in making predictions and decisions, neither ignoring the data nor relying totally on them works best unless the future will run on the same tracks as the past. Remember the coronavirus. Yes, it’s a virus. It’s like other viruses in some ways. It’s also different. The more different an event is the more likely experts’ experience, analysis and intuition become.

Leave a Reply

Powered by Paranoid Hosting™. 'Cause you never know...