Implicit conversion from float -9.063670411985015 to int loses precision