Implicit conversion from float -11.598837209302317 to int loses precision