Implicit conversion from float -124.13301662707838 to int loses precision