Implicit conversion from float -193.50515463917526 to int loses precision