Implicit conversion from float -10.866141732283467 to int loses precision