The MMPI and more recently the MMPI-2 have been held to be the clinical standard for assessing both fake-good and fake-bad response styles. In a contrasted-groups design, we compared simulators under fake-good (n = 67) and fake-bad (n = 58) instructions to controls (n = 90) and psychiatric patients (n = 95) under standard instructions. For fake-good profiles, we found that F-K < -12 had a hit rate of 80.6% and was partly supported by earlier research (Austin, 1992). For fake-bad profiles, F > 89 and F-K > 7 were the optimum cutting scores with hit rates of 86.2% and 87.0%, respectively. Unfortunately, previous research fails to confirm these cutting scores and leading proponents of the MMPI-2 substantially disagree on what methods to employ. Therefore, we recommend in cases of suspected malingering that the MMPI-2 be used only for screening purposes.