Quick reference
Key formula / structure
Run 1000+ simulations with random inputs from probability distributions → analyze output distribution (P50, P80, P95)
Memory aid
Don't guess a number — model the range. Run thousands of 'what ifs' to see all possible outcomes and their probabilities.
TL;DR
Monte Carlo replaces single estimates with probability distributions. Run thousands of simulations to see the range of possible outcomes. Focus on key percentiles (P50, P80, P95) for decision-making. Use tornado charts to identify which variables drive the most uncertainty.
What Is Monte Carlo Simulation?
Instead of estimating a single 'best guess' number (like project cost = $1M), model the uncertainty: cost could range from $800K to $1.4M. Run thousands of random simulations to see the probability of each outcome.
Anyone who considers arithmetical methods of producing random digits is, of course, in a state of sin.
— John von Neumann, co-developer of the Monte Carlo method (1949 talk, published 1951)
Monte Carlo Simulation replaces single-point estimates with probability distributions for uncertain variables. It then runs thousands (or millions) of random simulations, each time sampling different values from the distributions, to produce a probability distribution of possible outcomes. This gives decision-makers a range of outcomes with associated probabilities, rather than a single number that implies false certainty. It's particularly valuable for project cost and schedule estimation, financial modeling, and risk quantification.
Monte Carlo Simulation Process
From input distributions through random sampling to the output probability distribution, showing how uncertainty is modeled quantitatively.
From input distributions through random sampling to the output probability distribution, showing how uncertainty is modeled quantitatively.
Origin & context
Conceived by Stanislaw Ulam at Los Alamos in 1946 and developed with John von Neumann to model neutron diffusion in nuclear weapons research. Nicholas Metropolis named it after the Monte Carlo Casino, a nod to its reliance on randomness.
Core Components
Input Variables with Distributions
The uncertain inputs, each defined as a probability distribution rather than a single value.
Example
Task duration: Triangular distribution (optimistic: 5 days, most likely: 8 days, pessimistic: 15 days).
Random Sampling
The simulation randomly samples values from each input distribution for each iteration.
Example
In iteration 1, task duration = 7 days; iteration 2 = 12 days; iteration 3 = 6 days (randomly drawn).
Model/Formula
The mathematical model that combines inputs to calculate the output of interest.
Example
Total project cost = Sum of (task duration × daily rate) for all tasks, plus fixed costs.
Output Distribution
The resulting probability distribution of outcomes from all iterations.
Example
After 10,000 simulations: P50 (50% likely) = $1.05M, P80 = $1.2M, P95 = $1.4M.
When to Use Monte Carlo Simulation
Project cost and schedule estimation
Problem it solves: Projects consistently overrun budgets and timelines based on single-point estimates.
Real-world application
A construction company models project costs with Monte Carlo, setting contingency at the P80 level instead of using a single estimate.
Financial risk modeling
Problem it solves: Investment decisions rely on best-case scenarios without understanding downside risk.
Real-world application
An investment firm models portfolio returns under thousands of market scenarios to understand Value at Risk (VaR).
Business case analysis
Problem it solves: Business cases present a single NPV that implies false precision.
Real-world application
A product team models their business case with Monte Carlo, showing there's a 70% probability of positive NPV rather than presenting a single number.
You don't need exact probability distributions. Even rough estimates (best case, worst case, most likely) as triangular distributions provide far more insight than a single 'most likely' estimate.
How to Apply Monte Carlo Simulation: Step by Step
Before you start
- →A model or formula connecting inputs to outputs
- →Estimates of uncertainty ranges for key input variables
- →Monte Carlo simulation software or spreadsheet add-in
Build the deterministic model
Create a spreadsheet or mathematical model that calculates the output from input variables.
Tips
- ✓Start simple — even a basic model with key variables provides insights
Common mistakes
- ✗Making the model too complex before proving the concept
Define input distributions
Replace single-point estimates with probability distributions for uncertain variables.
Tips
- ✓Use triangular distributions (min, most likely, max) as a starting point
- ✓Focus on the variables with the most uncertainty and impact
Common mistakes
- ✗Trying to define precise distributions when rough ranges are sufficient
Run the simulation
Execute thousands of iterations, each randomly sampling from the input distributions.
Tips
- ✓Run at least 1,000 iterations for stable results; 10,000 for precision
- ✓Check that results stabilize (convergence)
Common mistakes
- ✗Running too few iterations for reliable results
Analyze and communicate results
Interpret the output distribution and communicate key percentiles and insights to decision-makers.
Tips
- ✓Present P50, P80, and P95 values
- ✓Show tornado charts to identify which inputs drive the most variability
Common mistakes
- ✗Presenting raw simulation data instead of actionable insights
Value & Outcomes
Primary benefit
Replaces false certainty of single-point estimates with probability-based insights for better decision-making under uncertainty.
Additional benefits
- ✓Quantifies risk in monetary or time terms
- ✓Identifies which uncertain variables drive the most risk
- ✓Enables probability-based contingency setting
What you'll learn
- →How to model uncertainty quantitatively
- →How to interpret probability distributions of outcomes
- →How to communicate risk in terms decision-makers understand
Typical outcomes
Best Practices
📋 Preparation
- •Identify the 5-10 most uncertain and impactful input variables
- •Gather historical data to inform probability distributions
🚀 Execution
- •Start simple and add complexity only as needed
- •Run sensitivity analysis (tornado charts) to focus on what matters
- •Validate the model with historical data if available
🔄 Follow-Up
- •Compare actual outcomes to the simulation distribution to calibrate future models
- •Update distributions as new information becomes available
💎 Pro Tips
- •Tornado charts (sensitivity analysis) are often more valuable than the simulation itself — they show where to focus risk management efforts
- •Present results as 'There's an 80% probability the project will cost between $X and $Y' rather than 'The project will cost $Z'
Limitations & Pitfalls
Results are only as good as the input assumptions — garbage in, garbage out
Mitigation: Use sensitivity analysis to identify which assumptions matter most, and focus effort on getting those right
Can provide false precision if distributions are poorly estimated
Mitigation: Be transparent about assumption uncertainty; present results as ranges, not precise numbers
Requires specialized tools and statistical knowledge
Mitigation: Use user-friendly tools (Excel add-ins) and start with simple models; expertise can grow over time
Where this comes fromStanislaw Ulam & John von Neumann introduced this (1946).
More Management Frameworks
Other frameworks in the Stratrix library.
Corporate Governance Framework
A system of rules, practices, and processes by which an organization is directed and controlled, balancing the interests of stakeholders including shareholders, management, customers, suppliers, and the community.
Risk & GovernanceISO 31000
An international standard providing principles, a framework, and a process for managing risk, applicable to any organization regardless of size, sector, or type of risk.
Risk & GovernanceThree Lines Model
A governance model that structures an organization's risk management and assurance activities into three distinct lines: management controls, risk oversight functions, and independent assurance.
Risk & GovernanceCompliance Management Framework
A structured system for ensuring an organization meets all applicable legal, regulatory, and ethical requirements through policies, training, monitoring, and enforcement.
Risk & GovernanceRAID Log
A project management tool that tracks four critical categories — Risks, Assumptions, Issues, and Dependencies — to ensure nothing falls through the cracks during project delivery.
Risk & GovernanceCOSO ERM Framework
A comprehensive enterprise risk management framework that integrates risk management with strategy and performance, helping organizations identify, assess, and manage risks across the entire enterprise.
More from Stratrix
Test a decision, free
Set a live decision against the companies that made it before you.
AcquisitionsThe Distance TestDivestituresThe Divestiture TestDiversificationThe Diversification TestMake or buyThe Make-or-Buy TestAll Access
Every Casebook, three Strategy Packs a seat each year and the Analyst, for teams that face these decisions all year.
$1,999 a year · $499 a seatWhat is included →New analyses, by email
The defining moves in business, checked against the record, as they publish.
See Monte Carlo Simulation in the wild.
Follow Monte Carlo Simulation across the companies and lenses where it actually shaped the strategy — the decision, the rationale, and how it played out.