Implicit conversion from float -55.030120481927696 to int loses precision