Implicit conversion from float -24.33070866141732 to int loses precision