Implicit conversion from float -61.24653739612188 to int loses precision