Implicit conversion from float -117.07317073170731 to int loses precision