Implicit conversion from float -11.660377358490564 to int loses precision