Implicit conversion from float -77.09844559585494 to int loses precision