When gambling, people tend to speed up after losses. This ’post-loss speeding’ is in contrast with ’post-error slowing’, which is often observed in behavioral tasks in experimental psychology. Importantly, participants can control the outcome in most behavioral tasks, but not in gambling tasks. To test whether perceived controllability over the outcome influences response speed after negative outcomes when gambling, we ran two online studies in which we created an illusion of control without changing the nature of the chance-determined gamble. Using the manipulation by Langer and Roth (1975), whose effect is replicated in Part I, we presented three groups of healthy participants (N = 600 per experiment, crowdsourced samples) with three different sequences of outcomes in a coin-tossing task. We replicated that participants presented with more wins at the beginning of a sequence estimated their ability to predict the outcome of a coin-toss higher than participants presented with more losses at the beginning, or those presented with a random sequence. Additionally, participants generally responded more quickly after a loss than after a win. However, the illusion of control did not influence post-loss speeding. This result is not consistent with several theoretical accounts for changes in response speed after sub-optimal outcomes.