We prove that a subtle but substantial bias exists in a common measure of the conditional dependence of present outcomes on streaks of past outcomes in sequential data. The magnitude of this streak selection bias generally decreases as the sequence gets longer, but increases in streak length, and remains substantial for a range of sequence lengths often used in empirical work. We observe that the canonical study in the influential hot hand fallacy literature, along with replications, are vulnerable to the bias. Upon correcting for the bias we find that the long-standing conclusions of the canonical study are reversed.

The canonical hot hand fallacy result was recently reversed, based largely on a single statistic, and a data set that was underpowered for individual-level testing. Here we perform a more robust analysis, testing whether hot hand performance exists across (i) data sets: four different controlled shooting experiments, (ii) time: multiple sessions per individual spread across a six month gap, and (iii) various (improved) approaches to statistical testing. We find strong evidence of hot hand performance, both across data sets and within individuals across time. Moreover, in a study of beliefs, we find that expert observers can successfully predict which shooters get the hottest.

We provide experimental evidence that subjects blame others on the basis of events they are not responsible for. In our experiment an agent chooses between a lottery and a safe asset; payment from the chosen option goes to a principal, who then decides how much to allocate between the agent and a third party. We observe widespread blame: regardless of their choice, agents are blamed by principals for the outcome of the lottery, an event they are not responsible for. We provide an explanation of this apparently irrational behavior with a delegated-expertise principal-agent model, the subjects’ salient perturbation of the environment.