Weather strategies for algotrading

from analogies to feed-in fluctuations

The majority of the intraday market follows the same weather forecast

On the intraday market, there are some significant deviations from the previous day’s forecasts, especially for wind power volumes.
On the one hand, this is because the inaccuracy of weather models can’t be reduced much further and, on the other hand, because of the spatial concentration of installed capacity. Particularly intense deviations occur in ramps, for example, during a cold front, hour-by-hour forecasts can be of by as much as 8 GW. For this reason, a significant share of the volatility in the intraday market is a result of changing weather forecasts or inaccurate estimates of the availability of fluctuating renewables.

The dominant position of the market-leading weather forecast has a major impact on plant and network management and thus the entire market. While many market participants follow this forecast literally, innovative trading companies and direct marketers are using alternative approaches to take advantage of this market inefficiency.

Considering the intraday weather trend even for the day-ahead auction

Energy Weather has addressed this market-inhibiting issue over the last two years and has developed an alternative: the Renewables Analogy Model (RAM). It bypasses the described marked inefficiency and makes taking an earlier position possible by analyzing similar weather conditions from the past. To this end, day-ahead weather forecasts are compared with similar patterns from previous years, and the intraday deviations that occurred afterwards are examined for significance. Result: even before the day-ahead auction the probable feed-in development of the next day’s intraday trading is obtained.

From a single source: Weather-based trading algorithms

The use of algorithms in the marketing of renewables per se generates a competitive advantage due to the speed of execution of a trading strategy. This advantage, however, exists primarily over manual traders and slow algorithms. And algorithmic trading alone is not enough if you want to compete with the best of the best in Europe. Equally relevant is the underlying model for taking and adjusting positions.

This results in two main ways to stay one step ahead of the competition:
1) Intelligence of the trading strategy
2) Speed

VisoTech and Energy Weather have bundled their know-how to develop the fastest and most intelligent trading algorithms.

The result is the first fully integrated service available on the market to combine even faster trading algorithms with weather-based short-term trading recommendations.
The charm of the early positioning is on the one hand in the liquidity of the day ahead auction, on the other hand in the subsequent possible continuous review of the assumption made. Position taking in the day-ahead auction can be fully automated via the existing ETS interface to EPEX SPOT.

The trading algorithm cycles the existing position against the latest weather models to determine either a further confidence gain or a conflict with the signal from the day-ahead analogies. Both signals are tradable immediately and well ahead of the final pricing. In addition, forecast measures for feed-in management as well as current measured values of wind and PV systems are included in the trading strategy. Thus, this is a well-rounded trading approach taking into account all price-relevant weather information, with the special feature of being one step ahead of the majority opinion formed on the market.

With a majority of market participants already using algorithms to market their positions, we continually invested in the performance of algorithms and the underlying framework. We are proud to be the fastest in the market and therefore, together with our customers, optimize in the range of milliseconds. To minimize latency to EPEX SPOT, the service is deployed in a data center in Frankfurt.