Implicit conversion from float -27.92452830188678 to int loses precision