Modeling Efficacy Report
There are so many ways one can validate a model. In power modeling, the bottom line comes out to be how your model translates the various inputs (commodity prices, outages, load, hydro conditions, etc…) into power prices. We cannot know for certain the inputs we place into the model will be the outcome of the future. Developing a range of simulations allows that uncertainty to be quantified. The next and final step is to make actionable items from all the work. This actionable item ultimately proves the value and efficacy of the model.
In this report, we take our default Power Market Analysis Near-term (PMA-NT) and put it through the test of making sure it is able to translate the risk and produce profitable actionable trades from April to December 2014. Though you may not be into trading and rather want to understand your asset or your procurement risk, by proving this process it will surely prove to be useful for your needs. Every day the model produces results and a screener can be run. All this is stored so you can always go back and validate it. Unlike other consulting companies all calibrations and historical performances are saved and clients have FULL access to them.
The test we used was to produce a list of recommended trades using a screener at the end of the month for a monthly power contract three months out (e.g. April analysis for July Contract). The screener takes the forward curve of closing and compares it to the all the runs. The screener is customizable and shows the calculated gains and potential losses from the simulations. The screener that was used for this report was to find trades having an upside of 3X compared to its loss. Finally all positions were closed mid-month prior to expiration.
The results were beyond expectations – 43% return from April to December 2014 without any additional analysis. For each month the model was able to screen the entire N. American power hubs and find trading strategies that produce positive returns. There were no months producing negative returns. The screener was sufficient to reduce high risk trades yet produce some options each month. A full monthly analysis and listing of each trade strategy is available in this PDF.
There is still much room to make the model results even better. All the trades were weighted equally. Further analysis of history and using the 60+ simulations could have indicated better trades likely re-weighting or eliminating some trades. In addition, a spread analysis could have been done to reduce some of the risk. Employing current market intelligence in terms of liquidity and market directions could improve results.
We challenge your modelers / consultants to do this post analysis. Do not add any additional analysis but use just the model results and see if your model can really quantify risk to produce actionable items. Please accept our challenge and may the best model win.
Your Modeling Energy Analyst,
David
David K. Bellman
Founder/Principal
All Energy Consulting LLC- “Adding insights to the energy markets for your success.”
614-356-0484
[email protected]
@AECDKB
