Estimation supports the development of higher level mathematical thinking and reasoning; however, has received relatively little research attention. We examined whether making predictions influences the accuracy of numerosity estimates in elementary-aged children, and whether the amount of information available to the estimator determines the accuracy of their subsequent estimate. The study was conceptualized on the basis of five different conditions for estimating, based on the contention that estimates will become more accurate as additional information becomes available to the estimator. To test this idea, the study utilized two tasks to examine the responses of year 2 (grade 2) and year 6 (grade 6) students (n = 110) when using prediction-first and estimation-only strategies. Across both tasks, we found a direct linear relationship between the accuracy of students' estimates and the amount of information available. Moreover, there was some evidence that being given an opportunity to make an initial prediction improved the accuracy of final estimates; however, these conclusions were tempered by task and age. Overall, our findings support a strategy sophistication effect, whereby the amount of information available to the estimator, and the ability to use that information effectively, increases the accuracy of the estimate.