Implicit conversion from float -59.16317991631797 to int loses precision