In software validation, the procedure that is not a good test is estimating possible outcomes using rough calculations and guesswork (Option D). Proper software validation involves thorough, systematic testing that provides reliable, reproducible results, not estimations based on guesswork. This approach lacks the precision and reliability required in software testing, which seeks to verify that the software performs as expected under various conditions through structured testing methods like dummy test runs, full-scale duplications, and analytical calculations.