Implicit conversion from float -11.846590909090907 to int loses precision