Chaos is normal in modelling Covid-19

aaa butterfly effect

It is normal in chaos to crave certainty. This is why many of us are demanding comfort through a magical statistical model that presents a rock-solid prediction of our Covid-19 future. Unfortunately, as any meteorologist will tell us, that’s as impossible as predicting the exact rainfall on a given Sunday in September. But we can still be prepared.

Perhaps the problem in this time of nano-second communication is that we have not stopped to understand the terminology. Words like prediction, projection, model, and forecast are all jumbled together in the non-scientific, layperson’s mind. To most of us they mean “What does the future hold?”

Meteorologist Edward Lorenz can help us understand. In 1961 he proved that when there are many, many variables at play it is impossible to accurately predict the weather over the longer term. He’s the guy behind the science of chaos theory who coined the phrase “sensitive dependence on initial conditions” or, more commonly referred to by the phrase “the Butterfly Effect,” a metaphor saying that the flap of a butterfly’s wings can reverberate into a thunderstorm.

What this means is that when you have hundreds or thousands of different interconnected variables at play even a small change in one can ripple through the entire network causing unpredictable change.

Covid-19 brings us many butterflies of change. Scientists calculating the future must incorporate such multiple data sets as Covid-19 presumptive cases, actual deaths, population demographics including age, gender, health conditions, where people live (urban, rural), usual population-movement patterns (work, shopping, gatherings, travel), the current supply of a wide range of medical equipment from masks to ventilators, the availability of hospital beds, the availability of medical professionals and laboratory scientists and technicians. Then add the global availability of supplies, equipment and personnel all dictated to by political decisions plus multiple variables I haven’t mentioned.

Every one of the above-mentioned variables is not frozen. Each one is shifting and moving and twisting about. This means yesterday’s facts, or even what happened in the last hour, is not likely what it is now. Trying to figure out how all those randomly moving  pieces will look in the future is the job of scientists who do statistical modelling. Imagine counting exactly how many mosquitoes are buzzing around you on a warm summer night.

However, it is very important for us to know that a model is not a prediction or a forecast. A model offers projections of what could happen if some or all of the currently identified interconnected variables like public behavior, supplies, equipment and other factors stayed the same or changed by a range of calculations based on today’s best evidence.

It is quite easy to make a prediction if what we’re looking at doesn’t change. All Poodles are dogs. If you see a Poodle in one year the probability is extremely high that it will be a dog. But not all dogs are Poodles. Asking how many Poodles there will be in one year gets more complicated because of all the things that can influence the outcome. And so it goes with Covid-19.

As pointed out by Dr. Mike MacCracken of the U.S. Global Change Research Program, it is very important to understand the difference in terminology. In 2001 MacCracken wrote an outstanding, simplified explanation of modelling terms that can be applied today. The article, Prediction versus Projection – Forecast versus Possibility, should be a must-read for all of us demanding a “model” of Covid=19 so we know exactly when life returns to normal  (https://sciencepolicy.colorado.edu/zine/archives/1-29/26/guest.htmlhttps://sciencepolicy.colorado.edu/zine/archives/1-29/26/guest.html).

However, in this period of chaos there is good news. It is true that modelling and projections (not predictions) involve a lot of “what if” scenarios with limited and uncertain control. However, we desperately seek control and power to lessen our fears. Here’s what MacCracken writes, “For a decision maker, a projection is an indication of a possibility, and normally of one that could be influenced by the actions of the decision maker.”

Each one of us is that powerful decision-maker in the chaos of Covid-19. The simple action of frequent hand-washing and social distancing is a very controllable variable although some may dismiss the ongoing message as mundane. But think of it this way. Our collective “simple” actions are proving to be the flapping butterfly wings leading to a reverberation of positive change in the future of all our health.

 

 

 

 

 

 

One thought on “Chaos is normal in modelling Covid-19

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s