Implicit conversion from float -52.280945757997216 to int loses precision