there's only 1 k zone exposed in the code, so revised the text to indicate as such
also added a bullet in the takeaway points to highlight that there is value in exploring a model through trial and error, but still acknowledging the important limits we run into pretty quickly
are we cool GLOBALLY with "formal statistical parameter estimation" rather than "Automated"? I mean, it's not that automated in reality right?