Let's use LOWEST_SCORE_THRESHOLD instead of magic 1000f when converting the resolution of a floating-point number to integer for comparison with zero.