Implicit conversion from float -57.76119402985074 to int loses precision