Implicit conversion from float -13.46260387811634 to int loses precision