Algorithmic Trading on the Electricity Market

What the big players already do

Large prop and energy companies (IMC, Citadel-level desks and specialised energy firms) already run algorithms for:

Speed is a key factor. The guest explicitly says he trades manually while others already run algorithms and react faster.

Where a human / AI agent still has an edge

  1. Context and market regime

    Bots are good at micro-imbalances but weaker at recognising when the market has entered a “forced flow” regime (a large player must buy or sell at any price).

  2. Weather + news + generation

    Combining weather forecasts, actual generation and news (outages, strikes, political decisions) is still better handled by a human or a well-designed agent.

  3. Event-driven opportunities

    Football matches, major events, abnormal demand — separate models can be built here.

  4. Negative prices as a distinct regime

    Strategies specifically designed around the appearance and disappearance of negative prices.

How to use AI (from the video)

Practical directions for agents