Implicit conversion from float -23.114754098360663 to int loses precision