Implicit conversion from float -60.506756756756744 to int loses precision