The acquisition of intuition, which guides problem solving by pruning unpromising strategies, is essential to the development of expertise in any domain. Problem-solving intuition may be viewed as analogous to search heuristics in artificial intelligence. One prediction inspired by this analogy is that practicing on subproblems and relaxed problems (versions of a problem with fewer constraints on the goal state and on the possible moves, respectively) may enhance the development of intuition for the full problem. Using the n-puzzle, we found that practice on relaxed problems did promote intuition compared to practice on the full problem, but impaired performance on solving the full problem. More detailed analyses suggest that practice on relaxed problems may discourage planning and encourage reliance on intuition. Planning is slower but more likely to produce optimal solutions if given enough time, whereas relying on intuition is faster but may lead to suboptimal solutions.