Implicit conversion from float -18.180300500834733 to int loses precision