I wish to ask forum contributors to address the question "How do you decide which activities in a risk schedule are in reality duration-correlated and by what percentage?"
This is a key question for obtaining reliable results from Schedule Risk Analysis (SRA). Without correlation, as you divide activities into smaller parts, provided you keep the proportions of uncertainty on the activity durations, SRA analysis will show the schedule duration risk to be smaller and smaller. However, from practical experience we know this is not true and the reason is correlation.
SRA will assume each of those sub-divided activities is completely independent of the other activities, whereas they are +100% duration correlated, if you have simply divided an activity into smaller pieces in series.
There is a related question that I will ask separately.
This is a key question for obtaining reliable results from Schedule Risk Analysis (SRA). Without correlation, as you divide activities into smaller parts, provided you keep the proportions of uncertainty on the activity durations, SRA analysis will show the schedule duration risk to be smaller and smaller. However, from practical experience we know this is not true and the reason is correlation.
SRA will assume each of those sub-divided activities is completely independent of the other activities, whereas they are +100% duration correlated, if you have simply divided an activity into smaller pieces in series.
There is a related question that I will ask separately.