This is true generally in pharmacology, but the plasma concentration of the androgen esters (prodrugs) isn't a good measure of "active" testosterone in your sys

This is true generally in pharmacology, but the plasma concentration of the androgen esters (prodrugs) isn't a good measure of "active" testosterone in your system. Thus the graph is a bit of an exaggeration. Basically, as long as the slopes are the same (rate of breakdown ≈ rate of prodrug conversion), you'd get ~same plasma levels. Ideally they'd break down at a constant rate (they don't, but close enough) leading to steady levels despite levels of the prodrug steadily diminishing. This makes once a week dosing of testosterone cypionate (for example) viable for many, but not all, as you've discovered. Dosing less but more often helps minimize the uncertainty/variability in breakdown/conversion, thus being more reliable at the cost of convenience.