Implicit conversion from float -152.8785607196402 to int loses precision