Explore the challenges and changes in forecasting energy and commodity markets with Matteo Mazzoni, Director of Energy Analytics at ICIS. From traditional methods to modern machine learning, adapting to geopolitical events and market complexity, and the vital role of traders in managing volatility.
Energy forecasting has shifted from simple models to complex scenarios due to market volatility.
External events like COVID-19 and geopolitical tensions have disrupted traditional forecasting models.
Forecasting now focuses on long-term energy transition predictions incorporating policy changes and technology advancements.
Deep dives
Evolution of Energy Forecasting Over Time
Energy forecasting has evolved from simpler, stable models to complex, volatile scenarios. In the past, models relied on regression and basic forecasting tools due to limited data and regional boundaries. However, with energy market liberalization, more sophisticated models using machine learning have emerged to predict short-term dynamics. The expansion of energy trading, more players in the market, and increased data availability have transformed forecasting from basic regression to advanced algorithms.
Impact of External Events on Forecasting
External events like COVID-19 and geopolitical tensions have shattered traditional forecasting models. The pandemic disrupted demand patterns, exposing the limitations of current forecasting tools, especially machine learning algorithms. The invasion of Ukraine and geopolitical uncertainties added new complexities to forecasting, challenging assumptions of stable market conditions. Companies had to revise their forecasting methodologies, emphasizing the importance of adapting to unpredictable global events.
Shift Towards Long-Term Energy Transition Forecasting
Forecasting has shifted towards long-term energy transition predictions, incorporating factors like policy changes, technology advancements, and renewable energy deployment. The focus is now on forecasting how the energy transition will reshape supply chains, impact global economies, and influence energy market dynamics. Models now face the challenge of integrating political factors, technological disruptions, and evolving market landscapes into accurate forecasting.
Complexity of Forecasting in Interconnected Energy Markets
Energy markets have become interconnected and complex, requiring a diverse set of skills for accurate forecasting. With markets no longer bound by geography, forecasting methods must reconcile political decisions, technological disruptions, and shifting demand patterns. The rise of LNG markets, evolution of carbon pricing, and increasing reliance on renewable energy drive the need for advanced forecasting models. Combining machine learning with human expertise and adaptable models is crucial to navigate volatile and uncertain energy markets.
Adapting Forecasting Strategies to Navigate Volatile Energy Landscape
Navigating the volatile energy landscape requires adapting forecasting strategies to incorporate short-term dynamics, long-term trends, and geopolitical uncertainties. Companies are reevaluating their forecasting capabilities by integrating weather analysis, energy transition forecasts, political landscape assessments, and market simulations. Strengthening trading capabilities, leveraging insights from experts, and embracing a diverse skill set are essential to manage risk and capitalize on market opportunities amidst evolving energy market complexities.
At the core of commodity trading is forecasting. The last decade of low prices and low volatility has transformed into a market of high volatility and higher prices. How has that changed the nature of forecasting? What does it mean for tools, models and skillsets. And are markets still forecastable? Our guest is Matteo Mazzoni, Director of Energy Analytics at ICIS, an independent commodity intelligence service that provides data analytics in energy and chemicals.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
