Curve building methodology

All forward curves on this website have been constructed with KyCurve, the forward curve building software of KYOS. This advanced analytical tool builds consistent arbitrage-free forward curves. It transforms publicly available commodity price data into forward curves with hourly (HPFC), daily or monthly granularity. The curves can be used directly to value contracts and assets.

All curves are on average equal to the end-of-day price quotations of various energy exchanges, including EEX, ICE, Endex and Powernext. Furthermore, historical day-ahead spot prices are used to shape the prices to an hourly or daily level. All prices are collected and processed overnight.

Methodology for constructing forward curves

The main points of the applied methodology include:

  1. Use all market data

    All curves are based on as many forward price quotations as possible. When available, it uses the day-ahead, weekend-ahead, working-days-next-week, weeks, months, balance-of-month, quarters, seasons, calendars and gas-years. When available, for power, it uses baseload and peakload forward prices.

  2. Arbitrage free

    All curves are arbitrage-free, so on average equal to the prices of all forward products in the market. When there is some inconsistency in the market prices, then the smaller products are leading. For example, when the quarter product is not on average equal to 3 months, then the quarter is discarded.

  3. Start with shaping to a monthly curve

    The forward curves are first shaped to at least a monthly granularity. For example, when only a calendar product is available, then KyCurve breaks the prices down into 12 prices with the correct relationships. This month-shaping is derived from historical price relations between traded forward months over the past year.

  4. From monthly prices to daily curves

    The forward curves are smoothed to a daily granularity using an advanced algorithm. The smoothing algorithm tries to avoid sudden jumps in the continuous curve from one month to the next. The algorithm furthermore avoids illogical waves, a common problem of standard smoothing algorithms.

Hourly Price Forward Curves

For any market, KYOS specialists monitor the curves and make adjustment to the methodology or input parameters where necessary.

For the construction of hourly price forward curves, the following methodology is applied:

  1. Include various daytypes

    The methodology distinguishes between various day-types (working days, Sundays, Saturdays, holidays and bridge days) specific to any country. The day-type and the month determine the appropriate hourly shaping.

    For example: for the German EEX HPFC a German holiday calendar is used, whereas a Dutch holiday calendar is applied to the Dutch Endex HPFC.

  2. From day to hour

    The methodology uses historical spot prices to find representative hourly weights for the future. Based on a unique backtesting methodology, the right set of historical price data is selected.

  3. Peak and offpeak hours

    The model takes into account the correct number of peak and offpeak hours in a month to perform the monthly and hourly shaping.

Purchase KyCurve Curve Building Software

KyCurve is in use by various leading energy market players. It can be licensed and installed locally to be run with your own market data, own parameter settings and own frequency of recalculation. It can be integrated within other systems or the KYOS Analytical Platform.

If you are interested in a KyCurve software license, please contact KYOS at info@kyos.com
or call +31 23 55 10 221.