Implicit conversion from float -14.145077720207254 to int loses precision