Implicit conversion from float -27.766497461928935 to int loses precision