Implicit conversion from float -79.91869918699186 to int loses precision