Implicit conversion from float -11.605839416058402 to int loses precision