| Probability Management extends traditional
information management into the domain of probability distributions. The
goal is to mitigate the Flaw of Averages, a class of systematic errors
which occur when single numbers are substituted for uncertainties.
Two foundational articles on Probability Management were published by
OR/MS
Today in 2006. You can download the original articles in PDF form
(Part
I, Part II)
or view them in HTML (Part
I, Part
II).
Probability Management is based on three core principles designed to
improve the analysis and communication of uncertainty and risk:
- Probability distributions should be organizationally certified to
ensure consistency and accuracy.
- Distributions should be modular, to allow cross-functional consolidation
of risk models. The DIST (DIstribution STring)
data type is a field-tested, compact and additive representation for
distributions.
- Graphical analysis of probability distributions and simulations is
essential to understand, validate and communicate model inputs and outputs.
More about Probability Management >> |
 |
News |
|
DIST 1.0 specification released
The DIST™ (Distribution String) data type encapsulates thousands
of Monte Carlo trials in a single cell, enabling modular risk models
that can be consolidated across divisional boundaries. The DIST 1.0
standard was developed in cooperation with AnalyCorp Inc, Frontline
Systems, Oracle Corp, SAS Institute and Vanguard Software Corp. More
about DIST >>
Risk
visualization seminars scheduled for Q1, 2009
Dr. Sam Savage will deliver a series of half-day seminars on
Risk Visualization in cooperation with SAS. The seminar will include
a discussion of the tools and practices of Probability Management,
in particular, interactive simulation, the validation of stochastic
data, and the presentation of probabilistic results to managers
who are still suffering post traumatic stress syndrome from their
college statistics course.
San Francisco, April 23, 2009
>>
Dallas, May 27, 2009
>>
Cary, June 19, 2009
>>
|
|
