Implicit conversion from float -16.055209260908285 to int loses precision