The ability to make intelligent decisions with limited resources is an essential skill for both humans and artificially intelligent systems. In practical situations where time and computational resources may be scarce, choosing when to perform computation becomes essential to achieving the greatest possible outcome at a limited cost. This is known as metareasoning. A recent study provided evidence that humans are skilled in this regard by showing that chess players make intelligent decisions about when to spend time planning. Additionally, this skill for computational resource management has been suggested as a way to explain overall human decision-making efficiency. To further test this hypothesis on a game with vastly different rules and strategies, we analyzed time management in over 100,000 online Go matches. We found that, as in chess, humans spent more time thinking when there was greater value in computation. This further supports the conclusion that human decision-making efficiency may be achieved through the rational selection of when to spend resources on computation.