Implicit conversion from float -260.1945525291829 to int loses precision