Posted on September 30, 2020 by Lisa Ward

by Marc Bond, Senior Associate

COGNITIVE BIAS

Success and value creation in the oil and gas industry have not been particularly good, as evidenced by its relative performance. This is widely recognized both in the global markets and within the industry itself. There are certainly many reasons that may explain why the oil and gas industry has not done well over the years, and one area that has received a lot of traction in recent years is the concept of cognitive bias.

Cognitive biases are predictableconsistent, and repeatable mental errors in our thinking and processing of information that can, and often do, lead to illogical or irrational judgments or decisions.

Surprisingly, the notion of cognitive bias has not been around for many years. It was first proposed by Amos Tversky and Daniel Kahneman in 1974 in an article in Science Magazine (Tversky and Kahneman, 1974). Since then, there have been numerous publications and research studies on the various cognitive biases and how they impact our judgments and decisions.

The book that introduced the concept of cognitive biases and their influence on the decisions of the general population was the seminal publication by Nobel Prize-winning psychologist Daniel Kahneman – Thinking, Fast and Slow (Kahneman, 2011). It is interesting to note that Dr. Kahneman won the Nobel Prize in 2002 not in Psychology, but rather in Economics. Why? Because traditional economic theory assumes that we are rational creatures when we make decisions or choices, and yet research and observations continually show that we do not.

There are many different cognitive biases (see Wikipedia), but there are a few that play a significant role within the oil and gas industry. These biases can act individually or in combination, leading us to poor judgments and decisions.

HOW BIASES MAY BE REPRESENTED IN THE OIL & GAS INDUSTRY

For example, imagine an exploration team is assessing a prospective area that is available for license bids. In the analysis of the data, the focus is on a very productive analog to describe the play. Nearby there has been a successful well recently drilled; and although it is acknowledged to be in a different play, the team is very excited about the hydrocarbon potential of their new play.

Some existing, older wells suggest that the new play may not work, but it is felt by the technical team that those wells were either old or poorly completed, and hence could be dismissed as valid data points. Given the uncertainty, the prospects and leads developed should have a very wide range of resource potential. However, given the team’s confidence in the seismic amplitudes, the range of GRVs estimated is quite narrow.

The team is also optimistic about the play potential and presents the opportunity to management in very favorable terms. If the company were to bid on and be awarded the license by the government, the team would be quite excited; and of course, success is often rewarded. The company ended up bidding on the license with a commitment of several firm wells. Upon further data collection and data analysis, a new team re-assessed the hydrocarbon potential and it is now believed to be limited; and yet there still is a large commitment to fulfill.

What happened to cause this result?  Was the original team overconfident in their expectations? Did they think that because they understood their commercial analog, they understood the perspective?  Were they so focused on the nearby successes? Was the data that was dismissed highly relevant? Were other alternatives and models not considered, which might have suggested that the resource size could be small?

Although the above narrative may appear to be contrived and one’s reaction to the scenario would probably be “I would never do that”, each of the justifications and decisions made are possibilities and all of them are rooted in forms of cognitive bias. You likely have recognized all or part of the scenario from your own experience. Further, these biases can work together in a complementary fashion, reinforcing the biased assessment, and making one “blind” to other possibilities.

Cognitive biases and their negative impact do not just present themselves during the exploration phase. There are numerous similar real-world scenarios observed in appraisal, development, production, and project planning projects.

STRATEGIES TO MANAGE

The bottom line is that these cognitive errors lead to poor decisions regarding work to undertake, issues to focus on, and whether to forge ahead or exit a project.  This makes it important to identify them and lessen their impact. Unfortunately, awareness alone is not sufficient. These biases are inherent in our judgments and decision-making and serve the purpose of helping us make rapid judgments based on intuition and experience. In our everyday life, they work generally well. Unfortunately, particularly in complex and uncertain environments such as the oil and gas industry, they can lead us to poor choices.

Hence, it is important to understand first what the biases are, why they occur, and how they can influence our assessments. This will then help us identify when our own, or our colleague’s judgments, assessments, and decisions may be affected by these cognitive biases. We then need to learn mitigation strategies. Given that these cognitive biases are normal and serve a purpose, the goal cannot be to remove them but rather to recognize the biases and then apply mitigation strategies to lessen their impact.

As noted above, there has been a lot of research on the biases, yet there is little published on actual, practical mitigation strategies. Hence, to help our industry, my colleague Creties Jenkins and I  have developed a course entitled Mitigating Bias, Blindness, and Illusion in E&P Decision Making course, where we go into further detail regarding these vitally important mitigation strategies. We use petroleum industry case studies and real-world mitigation exercises to reinforce the recognition of the biases. Finally, we show how to employ the mitigations to ensure any assessments or decisions are as unbiased as possible.

REFERENCES

Kahneman, Daniel, 2011, Thinking, Fast and Slow, Penguin Books, 499p.

Tversky, Amos and Kahneman, Daniel, 1974, Judgment Under Uncertainty: Heuristics and Biases, Science, vol. 185, no. 4157, pp. 1124-1131.

Wikipedia, List of Cognitive Biases, https://en.wikipedia.org/wiki/List_of_cognitive_biases

Posted on September 2, 2020 by Lisa Ward

I’m Doug Weaver, and I’m a partner with Rose and Associates residing in Houston, Texas. I joined Rose a little over three years ago after retiring from a 39-year career with Chevron. I’ve spent well over half of my career in exploration as a petroleum engineer.

I’m often asked, “Why would an engineer be so interested in exploration”? There are many reasons, but let me pose one of my usual responses – “if you think it’s difficult to generate resource estimates with all the data you’d want – try doing it with none”.

I hope to continue this blog well into the future and get into some of the services engineers provide for exploration teams. But in this first session, let me convey an observation on a topic that will be pervasive in future notes – Engineers and Geoscientists approach problems differently.

As I was scheduling my final semester of undergrad, I met with my advisor to get his feedback on one last technical course. Though my major was geotechnical engineering, I was a bit surprised when he suggested an advanced course in geology. Being that my advisor was one of the top geotechnical engineers in the world, I took his advice and enrolled in Geomorphology. The class consisted of about twenty geologists – and me. A good background for a future engineer in exploration!

All my engineering, math, and science classes had followed a very familiar cadence. Three hourly exams and a final. No reading, no reports, just understanding equations and concepts and solving problems with that knowledge on a test. Solve problems with math.

In the geomorphology class, we were posed with the problem of figuring out where a glacier had stopped and created a moraine. We collected data in the field. We then went back to the lab, plotting and interpreting this data. To my surprise, I was able to plot the exact location where the glacier had stopped. No formulas, just data collection and interpretation.

I’m fairly sure that Professor Hendron not only intended for me to learn about geomorphology but also to give me the experience of this alternate approach to solving a problem.

From what I’ve observed, this typifies the way most engineers and geologists solve problems (of course, I’m typecasting us all). Engineers start with a systematic workflow leading to a precise answer, while geoscientists use a more fluid, interpretive approach. Which leads us to the best answer? Both methods – when used together. The issues we face in exploration will certainly not allow the precise answer an engineer would normally want. In exploration, engineers need to embrace the uncertainty present in every aspect of their calculations. But, at some point, we need to quantify our analysis. We can’t make effective decisions if we can’t quantify and rank the investment options for our companies. And that becomes the primary role of the engineer in exploration – to quantify opportunities.

Back to our glacial moraine. Suppose I’m a Midwestern gravel company looking for mining opportunities. It’s great that I’ve identified my moraine and a potential quarry, but what does that imply from an investment perspective? How does this deposit compare to others I might exploit? What’s the quality of the sand and gravel within the deposit? Are others more accessible?

Switching hats from geologist to engineer, my task is now to answer these questions. I now understand that I will never know the exact size of the deposit, as it is uncertain. I’ll have to rely on samples collected to build a representation of the nature of the deposit, realizing the samples reflect a tiny portion of the total moraine. This data will inform me about the range of possible sizes of this deposit. I’ll want to investigate other deposits in the area to support the analysis of the samples I’ve collected in my own deposit and investigate how they were developed to get some idea of how to best evaluate the costs and timing of the extraction process. Finally, I somehow have to transform my moraine map and all these answers into a range of economic metrics, primarily Net Present Value, or if risk is present, Expected Value.

That’s where we’ll pick up next time, interrogating the Expected Value equation. Thanks for reading!