There are dozens of formulas in literature and standards for residual strength calculation of the pressurized ductile pipes with axial rectangular defect. Their accuracy is usually characterized by the so-called “model error” notion. The latter is established as the averaged deviance of experimental data from the calculated results. In fact, the model error determined in such way is partly related to the validity and range of the chosen experimental data. The idea of our work is to introduce the notion of “targeted model error.” It is expected to characterize the accuracy of given formula for the particular defect only with given ratio of its dimensions, which is named as the “point of interest.” So, every experiment can give the contribution to the targeted model error according to its weight on the point of interest. These weights are subjectively established based on the difference between the dimensionless residual strengths calculated for the point of interest and for the real defect dimensions of the considered experiment. Practical examples of determination of the targeted model errors are performed for the ratios of the defect depth to wall thickness equal to 0.5, 0.6, 0.7, and 0.8 for three well-known formulas—ASME, PCORRC, and O-formula.