Implicit conversion from float -82.0689655172414 to int loses precision