Implicit conversion from float -11.015228426395936 to int loses precision