Implicit conversion from float -11.321585903083701 to int loses precision