Two paths for power traders today.
Algorithmic energy trading is evolving along two routes, the traditional algo route following equities, and foreign exchange markets, and the fundamentals-based route.
This divergence presents a crucial decision point. Your decision will be determined by your perception of energy as a tradeable asset. Do you perceive the value of power as FIAT, i.e. its underlying value is determined by the value others place on it? Or do you view your power trade as a fundamental commodity, whose underlying value is based on physical market conditions?
If you believe short-term power is more FIAT, then you fall into the first category and are looking for an algorithmic trading solution that focuses on the orders that others make.
If you believe that ‘good’ human traders have the most credible view on market direction – then you fall into the second category and believe algos should do the same.
At Brady Technologies, we advocate trading strategies that encompass a directional market view as delivery approaches. While reacting to market orders is a viable strategy (akin to the first path), we do not believe adopting this approach alone leads to the most successful outcome. When it comes to embracing algorithmic trading in the traditional energy trading landscape, we believe it is important to establish your own direction, that is to lead – not follow.
What is the value of power?
As you approach delivery of your power position, your incentive to trade power is simple – you want to avoid the costs of not trading. Typically, this is avoiding the costs of imbalance. For the remainder of this blog, we will reference the GB market, however arguments are equally valid in any commercially balanced energy system.
In GB, if you do not trade*i you would receive the imbalance price (locally known as cash out). Thus, your human trader is marking their success on how well they beat every other trade (mark-to-market), as well as how well they beat cash out (mark-to-cash out).
The imbalance price itself is based on a bimodal distribution with the mode determined by the system direction. If there is more generation than consumption, the price is determined by the marginal MWh that the National Grid must sell through its balancing mechanism. If there is more consumption than generation, the price is determined by the marginal MWh the Transmission System Operator buys.
Thus, whilst it is difficult to predict the value of power– the value is determined by real-world actions by batteries and power stations changing their load, customers switching off and interconnectors being rescheduled.
At Brady, we believe you need to be able to lead the market to where you think fair value is and that you cannot simply follow.
Explore a world of possibilities with PowerDesk Edge
At Brady, we are developing a pioneering algorithmic trading solution for short-term power markets. PowerDesk Edge is highly capable of initiating orders, i.e. placing passive orders on screen hoping to move them fast enough to gain small incremental benefits (as described in category 1 in the first section of this blog.) But PowerDesk Edge also allows you to incorporate the latest AI (Artificial Intelligence) and ML (Machine Learning) into predicting direction and winning in category 2 as well.
Eager to learn more? Meet us at E-world 2024 in Essen between the 20th and 22nd February, where we will be showcasing our latest work in this field. Pre-book your meeting here.
Coming soon…
In forthcoming blogs, we will introduce a variety of concepts to help you enhance your algorithmic energy trading strategies:
DEAL – Trade now or trade later. If you are buying and you think the price is too high, how long should you wait and what should change your algo’s view? PowerDesk Edge’s Decision to Execute or Act Later methodology simplifies the actions.
SAGE – However complex your algo gets, its interface your core position management solution is as simple as Quantity Direction and Timing. Our Simple and Generic Execution retains complexity in the algorithm, not the interface to the market.
VITALS – PowerDesk Edge is python-based system, allowing you to introduce any of your own time series and streams that you believe could impact price. These Variables Influencing Trends and Analytical Systems mean you decide which inputs matter to you.
MERLIN – Brady’s governance structure consists of: Market-Insight, Embedded-Backtest, Live Papertrade, Integrated-Position-Management and Execution, Net-Profit-Review. You have full transparency on how this is built into our algo ecosystem.
PYCLAST – PowerDesk Edge is based on a series of bespoke python classes. This means that your quant can implement simple regression to advanced machine learning techniques. They are equipped and ready to go with our Python Classes Libraries and Sub Classes Templates.
