Purpose The objective of this study was to analyze the intra- and interobserver variability of this measurement according to a strict methodology and on a representative sample of the general population, as well as to identify the possible difficulties of measurement in case of patellar or trochlear dysplasia. Methods This observational study involved radiographic analysis by three independent observers of a total of 50 patients who had a loaded patellofemoral X-ray taken with the knee flexed to 45°. An initial reading was taken to measure the angle of the trochlear sulcus, the Merchant angle, and to classify the knees according to a possible trochlear dysplasia and/or patellar dysplasia according to Wiberg. A second measurement was then performed to analyze intraobserver agreement. Interobserver agreement was measured on all radiographic measurements (n = 100). Results The Merchant patellofemoral congruence angle showed good intraobserver concordance ranging from 0.925 (95% CI 0.868–0.957) to 0.942 (95% CI 0.898–0.967), as well as interobserver concordance ranging from 0.795 (95% CI 0.695–0.862) to 0.914 (95% CI 0.872–0.942). Poor results were found in terms of interobserver concordance on the measurement of the Merchant angle in case of stage 3 Wiberg patella ranging from 0.282 (95% CI −0.920 to 0.731) to 0.611 (95% CI 0.226–0.892). Conclusion Congruence angle is one of most commonly used measurements for patellar tracking. However, the convexity of the patellar surface makes it difficult to identify the patellar apex on its intraarticular facet, making the measurement of the Merchant congruence angle unreliable and not very reproducible in cases of stage 3 Wiberg patella. Registration N°IRB 2021/139