Implicit conversion from float -27.024793388429757 to int loses precision