Implicit conversion from float -24.26294820717132 to int loses precision