Implicit conversion from float -11.121372031662261 to int loses precision