Blind-Tested Digital Twins Aim to Tame Refrac Uncertainty

A URTeC 2025 study shows blind-tested refrac forecasts using integrated simulators, pointing to more predictable screening and capital decisions

24 Oct 2025

Refracturing has returned to the shale patch, though without the bravado that once accompanied it. The appeal is obvious: old wells, new techniques and the hope of cheaper barrels. The problem has always been uncertainty. Depletion, altered stresses and well interference can scramble outcomes, making refracs look less like engineering and more like a gamble.

A paper presented at URTeC 2025 in Houston points to a calmer, more methodical path. Titled “Blind Testing Simulator Predictions of Refracturing Performance in the Bakken and the Permian Basin”, it describes an attempt to forecast refrac results before they happen and then check whether the forecasts were right. The authors modelled four refractured wells, three in the Bakken and one in the Midland Basin. They first matched the simulator to production history and diagnostic data before refracturing. Only then did they predict post-refrac performance, without seeing the outcome in advance.

This blind testing matters. Many refrac studies explain results after the fact, which flatters the models and teaches little about real-world decision-making. Here, predictions were exposed to failure. That they broadly tracked actual performance suggests that integrated frac-and-reservoir simulators may be maturing into practical forecasting tools, rather than academic toys.

The paper’s author list is also revealing. Alongside ResFrac are Continental Resources, Devon Energy, ConocoPhillips, Hess and APA. Refracturing, it seems, is no longer being pushed forward by lone operators experimenting on their own acreage. Instead, progress is coming from shared data and joint validation, a slower but sturdier route.

This fits with the wider mood at the conference. World Oil’s coverage of URTeC’s first day noted renewed interest in restimulating older Bakken wells using updated designs. As top-tier acreage is drilled out and inventories thin, squeezing more from existing wells looks increasingly sensible.

None of this makes refracturing safe. Geological risk cannot be modelled away, and refracs will remain selective. But blind-tested workflows offer something valuable: discipline. By forcing predictions to be made, and judged, before capital is spent, they shift refracturing towards repeatable engineering. For operators weighing incremental barrels against the cost of new drilling, that gain in predictability may matter more than any single technical trick.