Purpose: In accordance with the requirements of section 3.17 of the Regulations on the System of Maintenance and Repair of Freight Cars Admitted to Operation on General-Use Railway Tracks in International Traffic (STOIR), when assigning a service life for a car different from that specified in the document, it is necessary to conduct resource tests of the car until destruction. But since there is a reserve in the structure, then with the same regulatory spectrum of impacts, the estimated number of cycles until destruction exceeds the number of cycles during resource tests by 2-10 times, depending on the selected coefficient of the fatigue resistance of the structure 1.15-1.8. Thus, if the results obtained by calculation are reliable, then the resource tests until destruction will take place with the number of cycles and damage 2-10 times greater than defined by GOST 33788. Resource tests have a high degree of reliability, as they are carried out on natural samples and under real operational loads. The exception may be testing on prototypes; in such cases, it makes sense to introduce a margin for the ‘non-naturality’ of the sample and its loading. Methods: Ways to reduce the safety factor for fatigue resistance have been explored. An analysis of regulatory documents, the requirements of which lead to exceeding the calculated number of cycles until failure over the number of cycles in resource tests by up to 10 times, has been conducted. Results: A series of actions, including innovative approaches, has been proposed to address the issue arising during the testing of loaded freight cars on the resource during collisions. Practical significance: The implementation of the recommended series of actions will eliminate the exceedance of the calculated number of cycles until destruction over the number of cycles in resource tests by up to 10 times. This will significantly reduce the time and financial investments in the process of conducting these tests.