Supplement to Three common statistical missteps we make in reservoir characterization Authors: Frank Male and Jerry Jensen

Here we show, through Monte Carlo experiments, three examples of statistical missteps that we have seen in the reservoir characterization literature.

The mistakes are

1. Applying algebra to linear least squares regression models
2. Improperly de-transforming a log-transformed variable in a linear least square model without accounting for bias
3. Mis-applying R²

All of the examples are in Statistics pitfalls.ipynb. Exposition around the first misstep is available in the notebook Applying algebra to regression.ipynb; the second is discussed in Regression on transformed variables.ipynb. The third misstep is detailed in Misinterpreting R-squared.ipynb.

Anyone can run these examples on Binder or Google Colab by clicking on the buttons above.

The citation is

Male, F. and Jensen, J.L., 2022. Three common statistical missteps we make in reservoir characterization. AAPG Bulletin, 106(11), pp.2149-2161. https://doi.org/10.1306/07202120076

The official version is at the AAPG Bulletin. A preprint is available at EarthArXiV.