Forecasting in the past, present, and future: David Orrell at TEDxParkKultury

Summary

The speaker discusses the limitations of mathematical models in predicting the future, using examples from weather forecasting, financial markets, and biology. They highlight the challenges of emergent properties, feedback loops, and the inherent unpredictability of complex systems. The talk suggests a shift from mechanistic models to a more organic, network-based approach, emphasizing scenario forecasting and system health over precise predictions.

Takeaways

• 🧩 The speaker emphasizes the human desire to predict the future, drawing a parallel to Wayne Gretzky's quote about playing where the puck will be, not just where it is.
• 🚫 The speaker highlights the limits of prediction, citing Yogi Berra's famous line about the difficulty of making predictions, especially about the future.
• 🌡️ Weather and climate forecasting is discussed as an example of how models have not significantly reduced uncertainty over time, despite technological advancements.
• 💊 The human genome project is presented as an example where understanding the 'book of life' has not led to the ability to predict drug effects or the spread of diseases like avian flu.
• 🔮 The speaker suggests that the problems in prediction are related to our worldview and how we approach thinking about the future.
• 🪐 A brief history of prediction is given, starting with the ancient Greeks who used circular models for predicting celestial movements, despite their flawed assumptions.
• 📉 The speaker discusses the limitations of economic forecasting, pointing out that markets do not follow the predicted stable equilibrium and are subject to boom-bust cycles.
• 🌐 The concept of the 'butterfly effect' is questioned, with the speaker proposing that the unpredictability of complex systems is more due to the inherent complexity and emergent properties than small disturbances.
• 💻 The speaker notes that despite advances in computing power, models for weather, economics, and health have not improved as much as expected in terms of predictive accuracy.
• 🌿 A shift in perspective is suggested, moving from viewing the world as a machine to seeing it as a living organism with complex, interconnected systems that are inherently unpredictable.
• 🌳 The speaker concludes by advocating for scenario forecasting as a method to prepare for an uncertain future, rather than attempting to pinpoint exact outcomes.

Q & A

What is the main theme of the speaker's talk?

The main theme of the speaker's talk is the use of mathematical models to predict the future and how these models affect the future.

What quote from Wayne Gretzky does the speaker use to illustrate the desire to predict the future?

The speaker uses the quote 'a good player plays where the puck is, a great player plays where the puck is going to be' to illustrate the desire to predict the future.

What are the three examples the speaker uses to discuss the limits of prediction?

The three examples used are weather and climate forecasting, the financial crisis, and the challenges in biology such as predicting the effects of drugs and the spread of diseases like avian flu or swine flu.

What was the initial uncertainty range in climate change predictions from over 30 years ago?

The initial uncertainty range in climate change predictions was between 1.5 to 4.5 degrees centigrade.

Why does the speaker mention the butterfly effect in relation to weather forecasting?

The speaker mentions the butterfly effect to explain the inaccuracy of weather forecasts, suggesting that small disturbances can lead to significant changes in weather patterns.

What was the ancient Greek approach to predictive models and what were its limitations?

The ancient Greeks developed predictive models of the cosmos based on the assumption that everything moved around the Earth in circles. The limitations were that these assumptions were incorrect, leading to models that were not accurate.

How did Isaac Newton's law of gravity change the way predictive models were developed?

Isaac Newton's law of gravity replaced the circular models with equations, moving away from the idea that everything moved in perfect circles and towards a more accurate understanding of planetary motion.

What are the two major problems the speaker identifies in modeling complex systems like the atmosphere?

The two major problems identified are emergent properties, where local effects lead to complex outcomes that cannot be reduced to simple laws, and the dominance of nesting and opposing positive and negative feedback loops.

What is the efficient market hypothesis and why did it become a popular explanation for forecast failures?

The efficient market hypothesis is a theory that suggests price changes are random perturbations to an optimal steady state, making markets unpredictable. It became popular because it provided a convenient way to calculate risk based on the normal distribution, despite its failure to predict major economic events.

What is the speaker's suggestion for a new approach to forecasting in light of the limitations of current models?

The speaker suggests moving from seeing the world as a machine to seeing it as a living organism, using new mathematical techniques like network theory and agent-based models to better understand and prepare for the future.