Implicit conversion from float -244.15841584158414 to int loses precision