Implicit conversion from float -141.74157303370785 to int loses precision