Implicit conversion from float -11.295731707317074 to int loses precision