Implicit conversion from float -92.03883495145632 to int loses precision