In engineering classes, we were always taught to make things as simple as possible – but no simpler. No matter how much we would like it to be different, properly assessing the value of large capital-intensive projects – and the risks to those values – is usually a complex task.
None of the analyses we provide – risk, decision, portfolio – are new, or anything other than very well-understood. It’s just that they are complex, and require a big step up from the spreadsheet analysis that is customary in the industry. Our tools and techniques are more often found in academia, or in large financial firms, but we believe that they have a home in smaller firms.
As mentioned on our home page, the term “risk” is usually applied only to risk of loss or some bad event happening. But it really means the chance that the future turns out differently than you expect. That can be bad or good. It’s this more comprehensive view of risk that we apply.
Our risk analysis services are based around a technique made possible with the proliferation of cheap, powerful computers. Rather than try to mathematically derive the distribution of the value – a nearly impossible task for any real-world project – we throw computational power at it. The technique is called “Monte Carlo Simulation,” and the concept is simple. Each variable in a formula of interest (in our case, it’s usually the NPV calculation) is replaced with a random number that represents how that variable can vary in the real world. We then run that simulation thousands of times, drawing randomly for those variables.
This method creates the population of future possibilities that we can calculate statistics on: average, standard deviation, percentiles.
It’s tough to make predictions, especially about the future.
– Yogi Berra
Our approach is based on predicting nothing. When we predict, we inevitably inject biases, so we try to predict nothing. For every variable of interest, we work with our clients to make sure that we understand the uncertainty in the variables that they are experts on – subsurface, costs – and we use our expertise to “simulate” future commodity prices. Those simulations use the past performance of prices to generate thousands of future prices, with jumps up in price, jumps down, and semi-random walks that tend back to a long-term average.