Answers are grounded in this paper and cite their sources.
Xiang et al. (2025) conclude that the gambler's fallacy does not originate in biased beliefs. Reanalyzing their data we arrive at the opposite conclusion. We show, more generally, that their main analysis, examining whether the median belief is biased, is nondiagnostic of the presence of the gambler's fallacy, while our test of the share of probability estimates for the streak ending above/below 50% is diagnostic. We also report results from a new study where at least 80% of gambler's fallacy predictions originated in biased beliefs.
No takes yet. Share an insight, caveat, or question.
Banki et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: