Implicit conversion from float -29.84402079722703 to int loses precision