Implicit conversion from float -27.415730337078656 to int loses precision