Constraints
Schedule constraints have a significant
effect on risk analysis results. They should be used sparingly, and only
when the constraint reflects reality. Constraints to be particularly aware
of are: • Must start on and Must finish on – preceding delays will not
delay the task, and preceding time savings will not bring it earlier •
Start on or after (SNET) and Finish on or after (FNET) – preceding time
savings will not bring the task earlier Consider removing these
constraints and replacing them with logic (e.g. Finish-to-Start links)
instead. Other types of constraints are less significant because they do
not influence the tasks' dates, only their floats. For example, you can
use a Finish on or before constraint to indicate a desired completion date
of a task – this will not force the task to finish on that day, but the
shortfall will be indicated in the task's float.
Open-ended tasks (Does not include ignored
links)
Options selected:Predecessors,
Successors
For a schedule risk analysis to be
meaningful, it is important that tasks' dates are set by logic (e.g.
Finish-to-Start links) rather than constraints. This is so that the risk
analysis will recognize the knock-on effect of delays. An open-ended task
is one that does not have at least one predecessor and one successor – it
indicates a possible lack of logic. Consider closing open-ended tasks: •
If a task has no predecessor, try to find some other tasks which could
potentially delay it. Leave it as open-ended if it is the project start
milestone. • If a task has no successors, try to find some other tasks
which it could potentially delay. Leave it as open-ended if it is a
project finish or reporting milestone.
Out
of sequence updates ("broken logic")
The logic in a plan can be broken when tasks
have started or finished before their predecessors. It is recommended that
any broken logic is removed or corrected to ensure the project schedules
as expected. For example, if task A has a Finish-to-Start link to task B,
but B has been started (by giving it an Actual Start date), this is broken
logic. It is not clear whether B's remaining work should wait for task A
to finish, or start straight away. Consider fixing the broken logic by
either removing the link or removing the actual dates. Also consider using
the retained logic / progress override options on the Scheduling tab of
the Plan | Options dialog box.
Lags
longer than 100 units
Option selected: Display lags greater then
100 Days
A lag is a gap in the logic between two tasks
– a delay between the dates of two tasks that are linked together. Lags
cannot have risk or uncertainty. In reality it is likely that the lag
represents either work or a delay, whose duration is uncertain. This is
particularly significant for long lags. Consider replacing the lag with a
task, so that uncertainty and risks can be assessed against it. Use the
Convert Lags to Tasks tool when a project contains a large number of long
lags.
None found.
Negative lags ("leads")
A negative lag is an overlap in the logic
between two tasks – often it is used to represent a task starting earlier,
with sufficient time allow some other work to happen. Lags cannot have
risk or uncertainty. In reality it is likely that the negative lag
represents an necessary overlap, whose duration is uncertain. Consider
replacing a negative lag with another kind of link that does not need the
lag. For example: • Replace a negative lag on a Finish-to-Start link with
a positive lag on a Start-to-Start link. • Split the tasks so that the
overlap is explicitly represented by a task.
Positive lags on Finish-to-Start
links
A lag is a gap in the logic between two tasks
– a delay after one task finishes before the next one starts. Lags cannot
have risk or uncertainty. In reality it is likely that the lag represents
either work or a delay, whose duration is uncertain. Consider replacing
the lag with a task, so that uncertainty and risks can be assessed against
it. Use the Convert Lags to Tasks tool when a project contains a large
number of long lags.
Start-to-Finish links
Start-to-Finish links are used deliberately
very rarely, because they have the unusual effect that the successor
happens before the predecessor. Consider whether this logic might be a
mistake, especially if it is between tasks that are not milestones.
None found.
Lags
between tasks with different calendars
Options selected:FS
A lag is a gap in the logic between two tasks
– a delay between the dates of two tasks that are linked together. When
the two tasks have different calendars, it is not clear which calendar the
task will use – whether it is the preceding task's or the succeeding
task's calendar. Consider replacing the lag with a task, so that its
calendar can be explicitly defined. Use the Convert Lags to Tasks tool
when a project contains a large number of long lags.
None found.
Links to / from summary tasks
Many people prefer not to put predecessors or
successors against summary tasks, because other project management tools
(such as Primavera P6) do not support them. Consider removing logic on a
summary task, by using a milestone to represent the start or finish of all
the tasks in the summary heading, and putting the link on the milestone
instead of the summary task.
None found.
Duration uncertainty distribution
shape
Tasks identified where: Maximum - Most Likely
duration divided by the Most Likely - Minimum duration is greater than 2.
The validation is only applied to the following distributions:
The check also tells you whether the numbers
entered for Min, Most Likely and Max create a valid distribution.
A task's duration can have a "skewed"
three-point estimate, which means it is not symmetrical. Usually
three-point estimates are skewed on the "pessimistic" side, where the
maximum is further away from the most likely than the minimum. The ratio
of (Maximum - Most Likely) to (Most Likely - Minimum) is used to measure
this skew. When the skew is significant (for example minimum = 1, most
likely = 10, maximum = 100), this may be an indication that there are
low-probability events that cause this pessimism. Consider using the risk
register to represent these risks events, and reduce the amount of skew on
the task's uncertainty.
None found. | 