Using a less deceptive variant of the false evidence ploy, interrogators often use the bluff tactic, whereby they pretend to have evidence to be tested without further claiming that it necessarily implicates the suspect. Three experiments were conducted to assess the impact of the bluff on confession rates. Using the Kassin and Kiechel (Psychol Sci 7:125-128, 1996) computer crash paradigm, Experiment 1 indicated that bluffing increases false confessions comparable to the effect produced by the presentation of false evidence. Experiment 2 replicated the bluff effect and provided self-reports indicating that innocent participants saw the bluff as a promise of future exoneration which, paradoxically, made it easier to confess. Using a variant of the Russano et al. (Psychol Sci 16:481-486, 2005) cheating paradigm, Experiment 3 replicated the bluff effect on innocent suspects once again, though a ceiling effect was obtained in the guilty condition. Results suggest that the phenomenology of innocence can lead innocents to confess even in response to relatively benign interrogation tactics.