Real Data, Real Uncertainty
Central to any cost risk analysis and model are the uncertainty distributions assigned to point estimates. Ideally, the analyst will have a database of historical cost and technical information that can be used to objectively develop Cost Estimating Relationships (CERs) using approved statistical methods. From that analysis the analyst will be able to objectively define the shape and dispersion of the CER uncertainty distributions. Similarly, with a suitable database or expert opinions to draw upon, the analyst will be able to develop uncertainty distributions for the inputs to the CERs. Unfortunately there is often insufficient time or resources to perform such detailed analysis and there is very little the analyst can use as a ready reference to cross check a detailed analysis should they do have the time, resources and tools to perform it.
There are many authoritative sources for guidance on methods, practices, and reporting requirements to develop a realistic and defendable cost uncertainty analysis. There are precious few sources for the analyst to obtain guidance on the most fundamental building block of uncertainty analysis, that is the shape and size of the uncertainty distributions to be applied to the cost methods and their inputs. AFCAA sponsored a study to fit distributions to common methods and their inputs for a variety of commodities. Over 3000 fits were performed. This presentation identifies the commodities and WBS elements that were examined and:
– shows how cost method uncertainty was measured (rather than assumed)
– presents fits for classic cost method inputs such as weight
– demonstrates how the results can be used in a specific cost model
– discusses trends that were found in the results by commodity
This presentation provides results for multiple commodities of interest to the cost analysis community. The data is presented in a model-neutral manner that facilitates its use in any of the popular simulation software applications. Additionally, while the results are derived from an analysis of real data, the manner in which the results are reported fully protects the source, rendering the results suitable for public domain distribution and use. The full, and therefore sensitive, results were delivered to AFCAA for distribution under the appropriate rules and regulations and are not discussed in this presentation.
Tecolote Research, Inc.
Mr. Smith earned a Bachelor Mechanical Engineering degree from the Canadian Royal Military College and a Master of Science with Distinction in naval architecture from the University College, London, England. He spent a total of 21 years in the Canadian Navy serving in positions such as submarine navigator, operations officer and as a naval architect. He has over 20 years experience leading, executing or contributing to life cycle cost model development and cost uncertainty analysis for a wide variety of military, Coast Guard, NASA and foreign projects. He has been with Tecolote since 1995 and since 2000 has been the General Manager for Tecolote’s Software Products/Services Group, responsible for the development, distribution and support of a variety of web and desktop tools supporting the cost community.
Tecolote Research, Inc.
Mr. McDowell is employed by Tecolote Research, Inc. as the Chief Scientist for the Huntsville Group, where he performs cost research and cost estimating for numerous customers. He has over thirty years of experience in cost estimating, cost analysis and cost modeling. Mr. McDowell received a Bachelor of Industrial Engineering degree from Auburn University in 1979.
Dr. Lew Fichter
Dr. Fichter is employed by Tecolote Research, Inc. as the General Manager for the Huntsville Group, where he manages the activities of 32 professional cost and financial analysts. He has 48 years of experience in cost estimating, cost analysis and cost modeling, and is the recipient of the SCEA Lifetime Achievement Award. Dr. Fichter was a co-author of the AFCAA Cost Risk an Uncertainty Handbook. Dr. Fichter received a Masters in Mathematical Statistics from Rutgers University (1970), and a MBA (1974), and a PhD. in Operations Research (1976) from the University of Alabama.