Implicit conversion from float -11.49812734082397 to int loses precision