Implicit conversion from float -16.051660516605182 to int loses precision