For the example under consideration the expected impact score of the branch A-B-C-D may be computed as below.

Event	Probability	Absolute probability	Magnitude	Significance	Impact value
A	1.0	1.000	-4.0	5	-20.00000
B	0.9	0.900	-3.0	6	-16.20000
C	0.9	0.810	-0.5	7	-2.83500
D	0.2	0.162	-0.1	0.2	-0.00324
Expected impact score of the branch =					-39.03824

In this case the difference in project impact scores arrived at by Westman's method and by the method being proposed by the author is low because the probability of occurrences of impact as well as the magnitude and significance of impact due to event D is very low. Let the following be the observed data in relation to branch 1 (i.e., branch A-F-I-L of Figure 5.3(b).

Impacts	Magnitude	Importance	Probability of Occurrence
	(1-10 interval scale)
F	2	5	A F(0.5)
I	2	9	F	I(0.6)
L	2	10	O L(0.9)

By Rau's method the value of expected branch impact score comes to be
Impact Score = (0.5x0.6x0.9)(2x5+2x9+2x10)=0.27x48 = 12.96
Following Westman's modification the branch impact score may be calculated as below :
Impact Score = 2x5x(0.5)+(2x9x0.5x0.6)+(2x10x0.5x0.6x0.9)
= 5.0+10.8+18 = 33.8
However if the author's proposed scheme is followed the expected impact score of the branch may be given as
Impact Score = 2x5+(0.5)+2x9-(0.6)x(0.6)x(0.5)+2x10x(0.9)x(0.6)x(0.5)
= 5.0+5.40+5.40
= 15.8

As may be observed Rau method under-estimates and the Westman method over-estimates the expected value of impacts of a given branch and in consequence the total impact score of the branches are also wrongly calculated.

However, even with this theoretically correct calculation procedure the replicability of network approaches would continue to be low because it suffers from all the difficulties of ensuring a commensurate scale. Moreover inclusion of probabilistic concept in the format of Sorenson matrix would significantly increase requirement of skill.