Implicit conversion from float -113.81679389312978 to int loses precision