This workbook demonstrates how to calculate the value of reducing uncertainty for two types of distributions. The first tab models a normal distribution by slicing outcomes into thousands of pieces, while the second tab shows a uniform distribution where all outcomes are equally likely. Enter your own values to see how the expected value of perfect information (EVPI) changes under each approach.

This workbook contains examples for the “one-for-one substitution” model described in Chapter 5. There are three tabs. The first tab provides a risk register that demonstrates how subjective expert input can be translated into fully quantitative risk analysis, replacing traditional heat maps with probabilistic modeling. The second tab shows the Monte Carlo Simulation and how to plot the “loss exceedance curve” (LEC) as shown in figure 5.6. Finally, there is a tab showing how multiple portfolios can be combined to make an aggregate LEC.

This workbook presents the calculated probabilities of various outcomes on calibration tests. Although these charts are not included in the published version of the book, they illustrate concepts discussed in Chapter 6. Specifically, they model how calibration responses would be distributed under the assumption that all test-takers are perfectly calibrated. These results provide a theoretical benchmark against which actual calibration test data can be compared.

 This resource provides additional calibration questions and answers not found in the book to help build your intuitions. 

Get the REAL calibration training here! HDR provides asynchronous training with self-paced videos so you can practice with proven methods, all while seeing your real-time calibration results on the calibrator dashboard. The data can even be used to optimize estimates for real-life problems with Team Calibrator!

This resource provides a practical introduction to using AI for estimation, with examples of generating confidence intervals through chatbots and integrating prompts directly into Excel.

This workbook uses a Monte Carlo Gantt chart to show how uncertainties in project schedules add up. It visually displays the start and end times of tasks, then simulates how delays and dependencies between tasks affect overall project duration. The tool also calculates the criticality of each task and its expected contribution to total duration. A built-in risk register feeds into the simulation, allowing you to see how potential risk events impact project timelines.

This spreadsheet introduces Bayesian probability through simple, interactive examples. Using the analogy of drawing marbles from two urns, it shows how new evidence updates prior beliefs to form more accurate predictions. One example applies this logic to testing whether AI performs better than humans, while another estimates the chance a project will be canceled if it is over budget.

This workbook provides interactive tools for working with the t-distribution and confidence intervals based on random samples. The first tab shows how to use the t-statistic to estimate the confidence interval of a population mean from sample data. The second tab demonstrates how confidence intervals change across different populations and sample sizes, with dynamic graphs that illustrate how the interval narrows as more data are included. Users can adjust inputs and select among populations to explore the behavior of sample means under uncertainty.

This workbook helps you estimate uncertainty about a proportion (for example, what percentage of customers prefer a product, or how often projects get canceled). It uses a beta distribution, which lets you start with a prior belief and then update it as new data come in. Simply enter your prior estimates and new observations, and the charts and calculations will show how your confidence changes.

This spreadsheet demonstrates how to estimate the size of a population when you can’t count or observe all of it directly. By entering two sample sizes and the overlap between them (for example, the number of tagged fish recaptured), you can see how the 90% confidence interval for the population size is calculated using the beta distribution.

This workbook introduces different types of regression models in Excel using project management examples. One sheet demonstrates simple linear regression, using the number of tasks in a project phase to predict completion time. Another shows how to build dynamic multiple regression models with the LINEST() function, letting you include or exclude variables on the fly. A third sheet illustrates logistic regression, estimating how budget relates to the probability of project cancellation using Solver. The final sheet explores nonlinear and interaction effects, using LOGEST() for log-scale fits and LINEST() to test how combinations of variables can strongly predict outcomes.

This workbook lets you test which distribution best fits your data when the true shape is unknown. Enter candidate distributions and sample observations, and the spreadsheet automatically updates probabilities and displays dynamic graphs of the candidate distributions, probability-weighted averages, observed outcomes, and how probabilities update as more data are added.

This workbook introduces two methods for modeling probability distributions when data is limited or irregular. One sheet demonstrates the Keelin MetaLog distribution, which can be fitted to percentiles or expert judgments, and the other shows how to build a kernel density estimate (KDE) that uses data directly without assuming a fixed shape. Both tools let you explore uncertainty and create flexible probability models for project and risk analysis.

This sheet outlines an example method for analyzing controlled experiments. It demonstrates a z-test for comparing test and control groups and requires sample sizes larger than 30. Enter your own values or try the provided examples to see how results change.

This workbook demonstrates the Lens Method, a model for comparing expert judgment to actual outcomes in order to improve forecasting. By entering likelihood estimates alongside observed results, the tool evaluates calibration, consistency, and accuracy of predictions. Users can explore how well different experts or models align with real data and identify ways to adjust future estimates for better decision-making.

This workbook shows how to build and use the machine learning method, Classification and Regression Trees (CART), entirely in Excel. Instead of relying on an automated algorithm, you step through the process manually—testing candidate splits, recording choices, and seeing how each decision affects predictions. It provides a simplified, transparent way to understand how CART works and how decision trees can be applied to project data.

This workbook models a technology regret analysis to help you evaluate the costs and benefits of adopting or delaying a new technology. It uses Monte Carlo simulation and scenario inputs to estimate when adoption is optimal, measure potential regret from acting too early or too late, and explore how uncertainty in costs and benefits affects the decision.

This resource demonstrates how Artificial Intelligence can be combined with expert judgment to produce a fully quantitative risk analysis. The workbook translates likelihood and impact estimates into probabilistic models suitable for simulation. Users can generate AI-based starting points, adjust them with their own expertise, and produce outputs such as expected and simulated losses, risk reduction from controls, and loss exceedance curves.

This workbook includes the calculations for Appendix 3: Selected Distributions. It includes several useful random distributions that apply to a variety of estimations that may be included in a model for project management. For instance, the Binary distribution produces a simulated value of “1” or “0”, which can be used as “event occurred” or “event didn’t occur” in simulating risk. Other distributions are more appropriate for impacts of a risk event. These distributions can also be used to simulate costs and benefits.

This paper, authored by Doug Hubbard, explains the pseudorandom number generator developed and applied at Hubbard Decision Research. It describes both the underlying algorithm and its implementation, emphasizing how the generator ensures statistical validity and practical reliability in simulation modeling.

This white paper introduces the principles and applications of Applied Information Economics (AIE), a decision analysis framework that combines quantitative modeling with practical business insights. While it is written with IT managers in mind, the methods address broader organizational challenges involving risk, uncertainty, and resource allocation.