Implicit conversion from float -11.704918032786878 to int loses precision