Implicit conversion from float -22.288135593220346 to int loses precision