The need to establish and modernise the work of national precision level networks has meant it has become important to examine the systematic and random errors that occur at the precision level. While great progress has been made in identifying and reducing the effects of random errors, the detection and inactivation of systematic errors has been less successful. In this study, we evaluated the accuracy of the second class first order Region 3 Precise Levelling Network Project of the Kingdom of Saudi Arabia. Systematic and random errors in the network were tested using Lallemand and Ebong formulae and, then, compared. The results show that the total error of the project was better than the first quality criterion (±1 mm/km) of the second order. The total errors calculated using the Lallemand and Ebong formulae were very similar, although the values calculated using the Ebong formula were generally smaller. Using the Lallemand formula, the random error of the whole R3 network was calculated as η = ±0.66 mm/km and the average systematic error was s = ±0.20 mm/km.