This was only used for progress report, and it's wrong because: - Pool might in theory be re-used by different tasks - We should not make any decision based on scheduling stats Proper way is to take care of progress by the task itself.