Implicit conversion from float -24.304812834224606 to int loses precision