Implicit conversion from float -11.3776337115073 to int loses precision