Please use this identifier to cite or link to this item:
Title: When More is Better-Design Principles for Prediction Markets in Defense Acquisition Cost Forecasting
Authors: Taroon Aggarwal
Ricardo Valerdi
Matthew Potoski
Keywords: Prediction Markets
Cost Estimation
Prediction Market Model
Issue Date: 31-May-2012
Publisher: Acquisition Research Program
Citation: Published--Unlimited Distribution
Series/Report no.: Prediction Markets
Abstract: This paper discusses the applicability of prediction markets in Defense Acquisition projects, specifically in estimating their cost and schedule. Several temporal and political factors can sometimes limit the effectiveness of traditional methods of project tracking and cost estimation, which may be overcome by using prediction markets. A prediction market provides an environment for traders to buy and sell contracts whose values are tied to uncertain future events. Efficient prediction markets have been shown to outperform available polls and other forecasting mechanisms. There are various prediction markets based on different models and algorithms. Our focus is not to analyze these models, but to identify the design principles of implementing a proven prediction market model in a defense acquisition project. Some pilot studies have been carried out that provide insight into the behavior of the market participants. We found increased involvement of participants and greater interest in the projects to be major benefits. The areas that need to be considered in the design and implementation of markets are related to the participants (like, which traders to include), the information to be collected or the stocks, the marketplace to be used and the incentive structure to keep the participants motivated to trade.
Description: Cost Estimating / Grant-funded Research
Appears in Collections:Sponsored Acquisition Research & Technical Reports

Files in This Item:
File SizeFormat 
MIT-CE-12-172.pdf725.15 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.