Implicit conversion from float -27.798165137614674 to int loses precision