Implicit conversion from float -9.243466299862447 to int loses precision