Working Paper: CEPR ID: DP8314
Authors: Claus Bjørn Jørgensen; Sigrid Suetens; Jean-Robert Tyran
Abstract: We investigate the "law of small numbers" using a unique panel data set on lotto gambling. Because we can track individual players over time, we can measure how they react to outcomes of recent lotto drawings. We can therefore test whether they behave as if they believe they can predict lotto numbers based on recent drawings. While most players pick the same set of numbers week after week without regards of numbers drawn or anything else, we find that those who do change, act on average in the way predicted by the law of small numbers as formalized in recent behavioral theory. In particular, on average they move away from numbers that have recently been drawn, as suggested by the "gambler?s fallacy," and move toward numbers that are on streak, i.e. have been drawn several weeks in a row, consistent with the "hot hand fallacy."
Keywords: gamblers fallacy; hot hand fallacy; law of small numbers; representativeness
JEL Codes: D03; D81; D84
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| players' avoidance of recently drawn numbers (C73) | players' number selections (Z22) |
| players' number selections (Z22) | players' betting behavior (L83) |
| players' avoidance of recently drawn numbers (C73) | players' development of the hot hand fallacy (Z22) |
| players' development of the hot hand fallacy (Z22) | players' betting behavior (L83) |
| recent draws (Y10) | players' number selections (Z22) |