Implicit conversion from float -22.559241706161146 to int loses precision