The illusion of control is an important feature of both problematic and nonproblematic gambling behavior. Crucially, this construct is incorporated in most cognitive models of problem gambling, and is also central in numerous approaches to gambling disorder treatment (e.g. psychological interventions using cognitive restructuring to mitigate the illusion of control). In this preregistered study, we tried to replicate the illusion-of-control effect, as defined and investigated in the seminal work by Langer and Roth, in an online context. Using the same trial procedure and a similar cover story as the original study, we presented three groups of healthy participants (N = 289; crowdsourced sample) with three different sequences of wins and losses in a coin-tossing task. Consistent with the original study, we found 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, although the effect sizes were small to medium (biggest Hedge’s g_av = 0.49) compared to the original study which yielded larger effect sizes (biggest ηp2 = 0.14). Thus, we replicated the findings in an online context, although the effect size was smaller than expected.