Prozentor  >  Forecasting  >  Electricity Prices

Electricity Price Forecasting

Our EnergyForecaster tool provides up-to-the-day projections for the energy and natural gas market

In addition to all major energy exchanges and Dow Jones indices we also forecast over-the-counter (OTC) electricity and gas market prices. For the VWD Market Manager Energy product distributed by vwd group we supply point forecasts for 105 workdays, spanning a 67% confidence interval.

Currently we are providing forecasting for the following time series:

• Germany EEX (total 39 time series)
  • EEX Phelix Baseload Index
  • EEX Phelix Peakload Index
  • All EEX hourly spot prices (24 time series)
  • All EEX block prices (11 time series)
  • EEX future baseload front month
  • EEX future peakload front month
• Germany vwd OTC electricity prices (total 5 time series)
  • OTC baseload day-ahead
  • OTC peakload day-ahead
  • OTC baseload weekend
  • OTC baseload week-ahead
  • OTC peakload week-ahead

Why use this kind of forecasting?

Our electricity price forecasting can be employed to obtain financial advantages in electricity buying and selling. It also helps portfolio managers anticipate near-term changes in electricity prices, enhancing trading tactics. These electricity price forecasts are used by energy suppliers and traders in active portfolio and risk management.

What is special about this type of forecasting?

In forecasting electricity prices we employ scientific methods, yet users do not have to have statistics expertise. The models employed are specially configured for particularities specific to electricity price time series.

Who uses this kind of forecasting?

Electricity forecasting is primarily used by portfolio managers, analysts, dealers and brokers active in the energy sector. Up-to-the-date forecasts allow precision projections of electricity and natural gas prices.

Procedures and methods

EnergyForecaster supplies point forecasts for the vwd market manager energy product, spanning a 67% confidence interval. The only input data processed are historical prices for the respective electricity time series and four additional informational time series as exogenous regressors (ICE natural gas futures front month, ICE Brent futures front month, ICE gasoil futures front month and vwd temperature index).

1. The electricity prices to be forecast are first transformed before generating the actual forecast. These prior transformations are either filters, such as differences or non-linear functions like logarithms and Box-Cox transformation.
2. An ARIMAX model is adapted to the correspondingly transformed time series. This class of model makes it possible to factor in the other values like oil prices, gas prices and temperature to improve forecast quality.
3. Short, medium and long-range forecasts are prepared and combined together into a single four-step forecast in optimal fashion by means of step scope blending.
4. A forecast for the upcoming months can thus be produced for every hourly block of the energy market in daily resolution.