Implicit conversion from float -226.70533642691413 to int loses precision