Braden Bowie
Product Manager at WhitsonX
Abstract:
The purpose of this talk is to improve automated production forecasting. The intent is to show that decline models shouldn’t be judged by how pretty they look on a single well, but by how consistently right or wrong they are across thousands of wells. This talk will show a way to systematically choose the level of complexity in the decline model, and blind test the results. Production forecasts for the Midland Basin were improved considerably as a result.