How a Thousand Imaginary Futures Saved Us from a Single, Flawed Forecast

I used to hate management presentations that showed a single, precise number for a project’s NPV. “The NPV of this project is $452.3 million.” It sounds so confident, so exact. It’s also almost always wrong. The future is not a single number. It’s a vast, cloudy range of possibilities. Pretending you can predict it with decimal-point accuracy is not just arrogant; it’s dangerous. The most important shift in my approach to financial modeling was when I stopped trying to predict a single future and started preparing for a thousand of them. This is the power of Monte Carlo simulation.

Our first real test of this approach came when we were evaluating a potential acquisition of a mature oil block in Riau, Sumatra. The seller’s model was a classic, deterministic DCF. It used a single, fixed assumption for the next ten years of oil prices (a gentle, upward-sloping curve, of course), a single assumption for the production decline rate, and a single assumption for the operating costs. The result was a single, attractive NPV.

My team and I knew this was a fantasy. We were operating in Indonesia, a country where the only certainty is uncertainty. So, we rebuilt the model from the ground up, but with a crucial difference. Instead of single numbers, we used probability distributions for the key variables:

• Oil Price: We didn’t pretend to know what the price of Brent crude would be. Instead, we used historical volatility and options market data to define a distribution of possible price paths. In some simulations, the price would spike to $120; in others, it would crash to $40.
• Production Decline: Our reservoir engineers didn’t give us one decline curve; they gave us three (P90, P50, P10), which we used to build a distribution of possible production profiles.
• Operating Costs: We looked at the historical cost overruns on similar projects in Indonesia and built a distribution for our OPEX, with a “long tail” of possible high-cost outcomes.

Then, we unleashed the Monte Carlo simulation. Using a simple Excel add-in, we ran the model 10,000 times. Each time, the model randomly picked a value from each of our probability distributions-a different oil price path, a different decline rate, a different OPEX scenario. It was like living through 10,000 different possible futures for this project.

The result was not a single NPV. It was a beautiful, bell-shaped curve showing the full spectrum of possible outcomes. The insights were stunning and changed the course of the deal.

1. The “Base Case” Was a Lie: The seller’s single NPV number, which looked so attractive, was actually on the optimistic side of our distribution. Our analysis showed that there was only a 30% chance of achieving that NPV or higher. The most likely outcome (the peak of our curve) was significantly lower.
2. The Real Risk Was Exposed: The most powerful part of the analysis was the “left tail” of the curve-the worst-case outcomes. Our simulation showed that there was a 15% probability that the project would have a negative NPV. This was a risk that was completely invisible in the seller’s deterministic model. It forced us to ask: “Are we willing to take a 1 in 7 chance of losing money on this deal?”
3. We Found the Key Driver: The simulation also allowed us to see which variable had the biggest impact on the project’s value. By analyzing the correlation between the inputs and the output, we could clearly see that the project’s NPV was far more sensitive to the oil price than to the production decline or the OPEX. This told us that the most important part of our due diligence was not to refine the engineering, but to hedge the commodity price risk.

Armed with this analysis, we went back to the negotiating table. We were no longer arguing about a single number. We were having a sophisticated conversation about risk. We were able to negotiate a lower purchase price, arguing that the seller’s valuation did not adequately account for the downside risk revealed by our Monte Carlo simulation. We also structured the deal with a contingent payment, where the final price would depend on the average oil price over the next two years. This allowed us to share the price risk with the seller.

Monte Carlo simulation is not about predicting the future. It’s about understanding the shape of uncertainty. It’s a tool for humility. It forces you to admit that you don’t know what will happen and to prepare for a wide range of possibilities. In a market as volatile as the Indonesian energy sector, this is not just a good practice; it’s a survival skill.

The Challenge to You

Take your current financial model. Identify the three most important assumptions (e.g., commodity price, production rate, capital cost). Instead of using a single number for each, create a simple triangular distribution: a minimum possible value, a most likely value, and a maximum possible value.

Now, run a Monte Carlo simulation with at least 1,000 iterations. What does the resulting distribution of your NPV or IRR look like? What is the probability that your project will lose money? What does the simulation tell you is the single biggest risk you face?

This analysis will give you a more honest and powerful view of your project than any single-point forecast ever could.

Originally published at https://www.linkedin.com.