Implicit conversion from float -77.96267496111975 to int loses precision