Some of the most influential modern theories of risky monetary decision-making assume that choices result from stable, trait-like preferences, invariant to contextual influences such as recent events. Recent research has challenged this assumption, demonstrating that even when values and probabilities are explicit and known, decisions under risk are contextually sensitive, affected by recent events on multiple timescales, including immediate (previous monetary outcomes), neighborhood (recently encountered values), and global (cumulative earnings relative to dynamic expectations) events. Such temporal context-dependencies are perplexing, because relying on recent events at any timescale is inconsistent with the assumed goal of risky monetary decision-making: to maximize payoff. Identifying this suboptimal behavioral pattern raises the possibility it can be mitigated using behavioral change strategies. We tested whether the effects of temporal context in risk-taking can be attenuated with an intentional cognitive strategy. 124 participants completed two rounds of a contextually structured gambling task with trial-by-trial feedback, the Emotion Regulation Questionnaire, and working memory capacity tasks. Participants were randomly assigned to complete each gambling round with a strategy either emphasizing a natural, uncontrolled decision-making approach or directly instructing participants to ignore context. Instructions to ignore context influenced temporal context effects on the immediate timescale but did not change those on the neighborhood or global timescales. The strategy was not uniformly effective for all individuals. The cognitive strategy eliminated (and even reversed) contextual effects on the immediate timescale for individuals with moderate and high habitual use of reappraisal. However, the cognitive strategy paradoxically strengthened contextual effects on the immediate timescale for individuals with low habitual use of reappraisal. The selective effects of strategic control on contextual influence indicates both intriguing possibilities and limits on the ability to optimize human decision-making, and suggests that people do not simply maximize local utility, but rather that even simple risky decision-making may be fundamentally goal-dependent.