A mechanism is disclosed for enabling efficient testing of a set of computer code. Untested code portions within a body of code are determined, and testing priorities are assigned to untested code portions according to how frequently the untested code portions are likely to be invoked during normal operation. More frequently invoked untested code portions are ranked higher than less frequently invoked untested code portions. The prioritized data may be used by a testing team to determine which additional tests should be created for effective and efficient testing of the body of code.