Analyzing expert decision-making in geosteering: Statistical insights from a large-scale controlled experiment

Geosteering is a sequential decision-making process used in the oil and gas industry which adjusts and controls the drilling trajectory of a well in real time, aimed at maximizing values derived from hydrocarbon production operations. For layered geological formations, Stratigraphy-Based Steering (SBS) has emerged as a popular approach to generate decision-supporting information to guide steering horizontal wells. This method involves the interpretation of log data measure while drilling, the development of a geomodel around the wellbore based on the log interpretation, and the use of the geomodel to guide well placement decisions. However, the main challenge in geosteering is that it is often not approached as a structured decision-making process. Consequently, essential decision quality elements—such as defining clear objectives and their trade-offs, alternatives, and properly quantifying uncertainties—are often missing. This issue causes a lack of unique and standard guidelines for geosteering practices.

This paper presents an analysis of data collected from 349 participants of a controlled geosteering experiment – the Rogii Geosteering World Cup (GWC) 2021. The data consists of log interpretations and geosteering decisions made by the participants, acting as geosteerers, for two wells representing conventional and unconventional drilling operations. More than 10,000 snapshots were recorded, consisting of interpretations of log data for each participant's well and corresponding decisions, every 2 min. These snapshots form a comprehensive database that is useful and valuable to provide insights into the decision-making process of the geosteerers and learning for improving geosteering decision-making. The dataset utilized in this study is openly accessible and published alongside the paper.

The novelties and key contributions of this paper are (1) a statistical analysis of recorded data to investigate causation and correlation between geosteering decisions and the quality of well placements, (2) revealing the factors that contribute to good geosteering decisions and well placements and (3) evaluating the extent to which good well placements are the result of interpretation and decision-making skills versus luck. By conducting a comprehensive statistical analysis of the recorded data, this study provides insights into the geosteering decision-making process and identifies key factors that are likely to contribute to favorable outcomes.