Why it is Better to Measure Bit Instead of Word Error Probabilities

In this paper we compare two measures frequently used when assessing the service quality of a digital transmission system: the word error probability (WEP) and the bit error probability (BEP). These two measures are compared by their accuracy and the complexity to obtain them using a sample mean estimator. The results are suitable for simulated systems (by means of Monte Carlo simulations) as well as the application of error probabilities when measuring the QoS in existing systems. It is shown that a soft measurement of BEP and WEP (based on reliability values instead of error events) is more likely to be accurate. The soft measurement of the BEP exhibits the highest accuracy in conjunction with the lowest complexity.