Rethinking Post-Training Quantization: Introducing a Statistical Pre-Calibration Approach