Implicit conversion from float -243.05921052631578 to int loses precision