Persistence on a task is both beneficial and costly, so it is important to understand how children learn to effectively balance between perseverance and seeking alternatives to reach a goal by monitoring their performance and tracking their progress over time ("adaptive persistence"). Typically developing children (N = 136) ages 3-7 years in the Midwest United States were invited to catch pretend fish at 7 ordered ponds with increasing numbers of fish. Unbeknownst to children, however, the probability of catching fish decreased across successive ponds, making it most rational to briefly "explore" new ponds to learn the payoff structure and then to "exploit" the earlier ponds before their chances ended. A latent class analysis of children's choices suggested three distinct patterns: (a) Explorers (55%), who repeatedly explored novel ponds, despite their lower rewards, (b) Exploiters (24%), who rapidly returned to earlier, more lucrative ponds, and (c) Balancers (21%), who began by exploring novel ponds but later returned to exploit the earlier, more lucrative ponds. Older children showed a greater probability of being classified as Exploiters or Balancers over Explorers. Controlling for age, children with higher executive function and metacognition tended to be classified as Balancers or Exploiters rather than Explorers. These findings suggest that self-regulation is a potential target for interventions aiming to support children's effective and fulfilling decisions.