Implicit conversion from float -11.067073170731717 to int loses precision