roulette wheel

Probabilistic Forecast Libraries

Decision modelers and forecasters increasingly realize the advantage of probabilistic forecasts -- that explicitly represent the inevitable uncertainties in the form of probability distributions. Many important quantities, such as GDP and the price of oil, are correlated or dependent.  Lumina offers libraries of probabilistic forecasts for groups of such dependent quantities, ready to incorporate in your models. 

These uncertain forecasts are represented as random samples using the DIST™ standard, a compact XML-based representation.  Each individual random sample is a Stochastic Information Packet  (SIP). A set of SIPs are bundled into a SLURP (Stochastic Library Unit with Relationship Preserved).   Inidividual SIPs can be downloaded and read into Analytica or any spreadsheet application that supports the DIST representation, including  Frontline Systems' Risk Solver and XLSim for use with Crystal Ball or @Risk.  An entire SLURP can be downloaded and read into Analytica, which converts it into an array of samples, ready for use in your model.

Watch a brief recording (5:14 mins) of Max Henrion, CEO of Lumina,  talking about how Analytica supports the DIST standard

AEO 2010 25-year Energy Forecasts

This SLURP includes projections over 25 years from 2010 to 2035 for 21 quantities, including US GDP,  US production, imports, consumption, and price of oil, gas, coal, and electricity, along with energy consumption by sector, energy intensity and CO2 emissions.  The median values are based on US Department of Energy's Annual Energy Outlook (AEO) 2010. Lumina estimated the uncertainty about these forecasts from the actual percent error distributions from comparing the past 25 years of AEO forecasts against actual values.  These distributions reproduce the historical serial correlation over time and cross-correlation among the 20 quantities.