Rock riprap is one of the most widely used erosion control methods for protecting embankments, levees, spillways, and instream structures subjected to overtopping flow conditions. At least 21 stone-sizing relationships exist to determine the median stone size of a protective riprap layer based on the results of 96 overtopping, laboratory experiments. Test parameters include median stone size, slope, unit discharge, coefficient of uniformity, and riprap layer thickness. A regression analysis was performed relating the observed median stone size to the predicted median stone size to each of the 21 relationships, yielding a coefficient of determination (R2) and percent error for the full spectrum of data. Zonal (partial spectrum of rock sizes) and complexity analyses were also conducted for each relationship. It was resolved that the Khan and Ahmad, and Chang relationships best aligned with the composite dataset. The predictive expressions by Olivier, Hartung and Scheuerlein, Knauss, Maynord, Abt and Johnson, and Siebel yield a noteworthy second tier of stone-sizing relationships for overtopping conditions.