Implicit conversion from float -288.35820895522386 to int loses precision