Implicit conversion from float -76.59292035398232 to int loses precision