Implicit conversion from float -11.10132158590308 to int loses precision