Added tests for initial state (no feedback should be shown) and for perfect-score-in-the-past scenario
