Oil & Gas Survey Data as Conditional Probabilities

I was faced with an interesting question over the weekend when my industrial colleagues asked me how to summarize and present oil and gas survey data as a function of whether these data lead to drilling that produced reservoirs with commercial success. These data are critical for both oil and gas (exploration and production) companies and survey companies that provide survey services to the E&P companies.

The data set that my colleagues want is the probability of correctly detecting the commercial reservoirs using the survey data. Being able to state the success rate of identifying commercial reservoirs using a certain survey method is valuable and important to potential clients. The data that they have is the success rate of the drilling campaigns based on the survey final recommendations. These two sets of data are of course related, but they are not the same.

One tricky aspect of solving this problem is that the oil and gas companies that employ different survey methods do not want to reveal their drilling success rate for either practical or confidentiality reasons. In other words, the success rate of identifying commercial reservoirs using a particular survey method – be it seismic, electromagnetic, or gravitational – has to be determined from incomplete data.

These incomplete data do not necessarily lead to complications. There are ways to go around them using probability formulas. In fact, the final formula–as it turns out–allows me to obtain the success rate of identifying commercial reservoirs using only data that depend on other factors, i.e., using only "conditional probability" data.

One interesting result I found, for example, is the success rate of identifying commercial reservoirs depends on the success rate of a survey method to correctly identifying both commercial reservoirs and non-commercial reservoirs. Correctly identifying commercial reservoirs alone does not give a full picture on how good a survey method is. We should also need to know how good the survey method in correctly identifying non-commercial reservoirs as well.

The conditional probability formulas I use to summarize and present the oil and gas survey data can also be used to validate a claim by an oil and gas survey company. This validation process is important for an oil and gas company to decide which survey company it should hire to maintain or increase its production target.