FAST Trader
Forecast Accuracy: Swift Transmission
Metra Fast Trader delivers the most accurate temperature forecast with the current market moving weather forecast data.
FAST Trader plots the GFS and ECMWF operational and ensemble runs against Metra’s more accurate deterministic forecast, the previous run(s) and 10 year climatology. All forecasts and the climatology are aggregated to produce a single temperature forecast number for the UK, based on demand weighted LDZ sites. This information is delivered to subscribers at high speed as the raw data is published.
Why have we done this?
FAST Trader has been designed to offer you the ability to trade delta between Metra’s best forecast and the raw, ECMWF and GFS, market moving data; as soon as the raw data is published. While the raw model data drives perception, Metra’s forecast is much more accurate. Metra’s lower bias and lower Mean Absolute Error allow you to take stronger positions, improve your gains and better judge how the market will move when temperature is a key driver.
How good is Metra’s forecast?
The graph below shows the performance of Metra aggregated deterministic forecast for the UK LDZ sites between November 12 and December 11 2009. Consider ECMWF’s 2 degree cold bias at day 10 compared with Metra’s negligible cold bias. Metra’s Mean Absolute Error (MAE) is consistently 1 degree lower than the ECMWF ensemble mean throughout the entire forecast horizon. FAST Trader comes with an ongoing skill score so you will know the skill of Metra’s forecast in all weather regimes.
What makes the Metra Forecast so accurate?
Our temperature forecast combines medium range forecasts from global forecast systems, with short range mesoscale forecasts from high resolution models. We then apply our proprietary methodology to produce a seamless 15 day forecast with dependable accuracy through the step changes in forecast resolution. It calibrates huge amounts of model data using a short time series of historical data and uses cutting edge statistical techniques, to extract, combine and place the correct emphasis on the different strengths of the best models. A running skill score within the product will further demonstrate Metra’s lower error and bias against ECMWF & GFS.