Implicit conversion from float -59.28229665071771 to int loses precision