Implicit conversion from float -61.0180995475113 to int loses precision