Algorithms textbooks typically approach the analysis of algorithm performance as a desk-based pen-and-paper bean-counting exercise that focuses on estimating the number certain operations in the worst case. The goal of this project is to develop a tool to supplement such analysis by automating the collection of performance statistics as the algorithm executes. The tool should parse the algorithm's code, identify the key operations to be included in the count and then instrument the code so that the number of such steps executed during a run of the algorithm is collected. The tool will also manage a suite of test inputs for each algorithm, to permit the user to explore those inputs that give rise to the worst-case performance.