Implicit conversion from float -177.85714285714283 to int loses precision