ICEAA’s annual Professional Development & Training Workshop brings together professionals from government, industry and academic cost communities for an intensive 3 1/2-day program dedicated to developing the understanding and appreciation of using data-driven estimating and analysis techniques.
2015 Best Papers
2015 Best Paper: Program Management Track
2015 Best Paper Overall
The Psychology of Cost Estimating
Research into human psychology has yielded amazing findings into how we process information and how we use information to make decisions. Cost estimators can use psychology and behavioral economics to improve not only our cost estimates but how we communicate and work with our customers. By understanding how our customer’s think, and more importantly, why they think the way they do, we can have more productive relationships and greater influence. The end result will be better decisions by the decision makers.
Paper: PM07 Paper Prince Psychology Cost Estimating
Presentation: PM07 Presentation Prince Psychology Cost Estimating
2015 Best Paper: Government Processes Track
The Navy Modernization Program:
Estimating the Cost of Upgrading AEGIS Guided Missile Cruisers
The Navy is well into its 20 year, $16B (CY10$) plan to modernize 84 AEGIS warships. This cost estimate covers eleven of the planned ships and accounts for maintenance and upgrades to the ship and the combat systems. The estimate leverages recent contract and shipyard performance data and interviews with engineers, resulting in a detailed study and recommended cost savings initiatives. The methods and data in this estimate will assist ship modernization cost efforts across the fleet for the foreseeable future.
Paper: GP06 Paper Goucher Navy Modernization
Presentation: GP06 Presentation Goucher Navy Modernization
2015 Best Paper: Methods & Models Track
Generalized Degrees of Freedom (GDF)
Minimum-Percentage Error/Zero-Percentage Bias (ZMPE) method is commonly used for multiplicative-error models. But ZMPE users do not adjust degrees of freedom (DF) for constraints included in the regression process. This generates misleading ZMPE CER fit statistics and underestimates the CER error distribution variance. Hence, ZMPE results are incomparable with those derived without constraints. This paper details why DF should be adjusted and proposes a Generalized Degrees of Freedom measure to compute fit statistics for constraint-driven CERs.
Paper: MM06 Paper Hu Generalized Degrees
Presentation: MM06 Presentation Hu Generalized Degrees
2015 Best Paper: Parametrics Track
An Early Production Methodology
Learning slope selection is a critical parameter in manufacturing labor estimates. Incorrect ex ante predictions lead to over- or understatements of projected hours. Existing literature provides little guidance on ex ante selection, particularly when some actual cost data exists but the program is well short of maturity. A methodology is offered using engineered labor standards and legacy performance to establish a basic learning curve slope to which early performance asymptotically recovers over time.
Paper: PA03 Paper Johnstone Improvment Curves
Presentation: PA03 Presentation Johnstone Improvment Curves
2015 Best Paper: Risk Track
Beyond Correlation: Don’t Use the Formula That Killed Wall Street
Risk models in cost estimating almost exclusively rely on correlation as a measure of dependency. Correlation is only one measure (among many) of stochastic dependency. It ignores the effect of tail dependency, when one extreme event affects others. This leads to potential outcomes that do not make sense such as a program with a large schedule overrun but no cost overrun. This method has been widely blamed for the financial crisis a few years ago. The copula method is introduced as a way to overcome these deficiencies.
Paper: RI11 Paper Smart Beyond Correlation
Presentation: RI11 Presentation Smart Beyond Correlation
2015 Best Paper: Software Estimating Track
Early Phase Software Cost and Schedule Estimation Models
Software Estimating Track
Software cost estimates are more useful at early elaboration phase, when source lines of code and Function Points are not yet available. This study introduces effort estimation models using functional requirements as a size predictor and evaluates development schedule using estimated effort based on data from 40 military and IT programs delivered from 2006 to 2014. Statistical results show that estimated functional requirements and peak staff are significant contributors and that estimated or actual effort is a valid predictor of development duration.
Paper: SW12 Paper Rosa Early Phase Software
Presentation: SW12 Presentation Rosa Early Phase Software