An automated process identifies which components that retain their state need to be resettable in a design. The design is analyzed to identify components that retain their state and are non-resettable. A set of simulation tests is run on the design, where each test is known to pass when all components that retain their state are reset at reset. The tests are run with a respective logic value (1 or 0) randomly assigned to each non-resettable component at reset, until a test run fails. The failed test is rerun a specified number of times, each time with a different set of randomly assigned logic values provided to non-resettable components at reset. For each run, statistics are logged for each non-resettable component according to the test results and the logic value provided to the non-resettable component. The process determines which non-resettable components need to be resettable according to the statistics.