Implicit conversion from float -11.357142857142847 to int loses precision