Implicit conversion from float -63.9406779661017 to int loses precision