Unexpected but Incidental Positive Outcomes Predict Real-World Gambling

A. Ross Otto, Stephen M. Fleming, Paul W. Glimcher

Research output: Contribution to journalArticlepeer-review


Positive mood can affect a person’s tendency to gamble, possibly because positive mood fosters unrealistic optimism. At the same time, unexpected positive outcomes, often called prediction errors, influence mood. However, a linkage between positive prediction errors—the difference between expected and obtained outcomes—and consequent risk taking has yet to be demonstrated. Using a large data set of New York City lottery gambling and a model inspired by computational accounts of reward learning, we found that people gamble more when incidental outcomes in the environment (e.g., local sporting events and sunshine) are better than expected. When local sports teams performed better than expected, or a sunny day followed a streak of cloudy days, residents gambled more. The observed relationship between prediction errors and gambling was ubiquitous across the city’s socioeconomically diverse neighborhoods and was specific to sports and weather events occurring locally in New York City. Our results suggest that unexpected but incidental positive outcomes influence risk taking.

Original languageEnglish (US)
Pages (from-to)299-311
Number of pages13
JournalPsychological Science
Issue number3
StatePublished - Mar 1 2016


  • decision making
  • emotions
  • open data
  • risk taking

ASJC Scopus subject areas

  • General Psychology


Dive into the research topics of 'Unexpected but Incidental Positive Outcomes Predict Real-World Gambling'. Together they form a unique fingerprint.

Cite this