Chasing losses and wins: A drift-diffusion decomposition of sequential gambling decisions

Loss-chasing and win-chasing, the tendency to continue and/or intensify betting after losses and wins, constitute key behavioral pathways to problematic gambling. However, the decision processes that give rise to these behaviors remain poorly specified, and the literature lacks a task capable of isolating these cognitive mechanisms. To address these questions, in the present study, two hundred UK-based participants (50.5% men; Mage = 40.33) completed a novel online task that assesses loss-chasing and win-chasing within a single paradigm. On each trial, participants first won or lost some money and then decided whether to play a double-or-nothing gamble (‘chase’) or quit. We analyzed their behavior through an evidence accumulation model to characterize their decision process. Pre-registered confirmatory analyses showed that loss-chasing and win-chasing displayed distinct patterns: win-chasing decreased with larger win amounts, whereas loss-chasing was not reliably influenced by loss amounts. The evidence accumulation model revealed that people showed valence-dependent shifts in their starting point, amount-dependent changes in their drift rate, and increased boundary separation (i.e., cautiousness) for higher stakes. Interestingly, chasing decisions, the drift rate and the boundary separation parameters further correlated with participants’ Problem Gambling Severity Index. People’s decisions to chase an outcome thus appear to arise from both biased starting points and altered value-based evidence accumulation, integrating impulsive/compulsive and value-based accounts of chasing behavior. Our combined experimental and computational approach provides an integrative framework for understanding people’s decision to continue or stop gambling, a decision that lies at the core of excessive gambling and its associated harms.