Implicit conversion from float -113.52112676056339 to int loses precision