If you are too pessimistic at Tender Stage you may not win the contract?
If you are too optimistic during the Contract Stage you may not complete on time?
DS

Pessimistic or most practicable

I cant imagine the effort of generating such schedules with 35,000 actitivities for one projects. The effective way, maybe, is to present schedule to the client for approval without discussing optimistic, pessimistic or most probable dates. The dates should show only attainment of contractual dates Upon approval,the schedule is the contract schedule. All project activities and performance will be compared to the contract schedule. The contract schedule is the baseline schedule without any thoughts that such schedule is optimistic, pessimistic or most probable.

effectivness of Schedule risk planning is proportional to the detailing of the schedule.
First Schedule/Cost Risk Analysis (Monte-Carlo simulations) is an effective way of finding out prob of the job completion on a date or so.
To make sure of time/productivity estimates, say for eg 150-200 activities (proposal stage), with proper logic/durations/distributions, find out the most sensitive activities (top 10%), attack them, try to narrow down their spreads & so on and so forth. In this way try to get an optimal solution (which in practice is like a mirage).
Now "relationships", FS, FF, SS all are used at some point of time because of nature of activities & also level of details available (Civil/Strl works are quite easy compared to Piping works). These lags also have their own duration distributions. So this increases the degree of freedom.

In Strict sense unless you have FS relationship, it is not feasible to get a good solution using risks. FS relation means too much details (Is that level worth in a schedule???).
Pertmaster, Crystal Ball & all risk analysis s/wares are effective to the limit of understanding of the scheduler. It is the inputs given to S/ware which governs the solution & this inturn depends on the planner’s understanding of job & its dependencies.
Finally, experience/knowledge of the job is the key to any solution. S/ware expedites the solution.
Regards
Dattatreya

here’s my 2 cents...

Complete 3 point estimating
optimistic is based on the most probable best time eg driving to work and getting all green lights
most likely is based on the most probably time eg driving to work and getting some red and some gree lights
pessimistic is based on the most probable worst time eg driving to work and getting all red lights.

Now...the event of getting a flat type is a risk and entered into our risk register, because it has less than 50% chance of occuring.

When it comes to the schedule, we start with most likely estimates, we apply the PERT forumala to the critical path (and near to critical path) activities and update those durations to include some ’schedule risk’ into the critical path.

We also baseline rework into the schedule...so in the software section you would see req’s analysis > prelim des > detailed design > code & unit test - c&ut rework - dt&e rework

right wrong or indifferent...it doesn’t really matter, because before the ink has dried on the contract the schedule changes *grin*

Usual wording:
How long does it usually takes?
Then: what will be your estimate if everything will go smoothly - without rains and interruptions?
Then: and what is your longest estimate - if the risks that may be happen (rains, interruptions and other troubles) will really occur at the time the work is performed.
Besides events you may estimate resource productivity (skills) - if the resources will have best productivity, usual productivity and worst productivity.
You shall also estimate probable risk events like failures in electicity supplies or rework. These events may take place in pessimistic scenario.
You will have three schedules that may contain different events (rework only in pessimistic schedule). Comparing finish dates you will notice that optimistic date is closer to most probable than pessimistic.

Vladimir

www.spiderproject.ru/index_e.htm

No, Jaco.
Optimistic schedule is calculated basing on optimistic duration and cost estimates and certain events.
Most probable is based on most probable estimates and risk events that have probabilities exceeding 50%.
Pessimistic schedule is calculated basing on pessimistic estimates and all probable risk events.
Probability curve will have maximum at the most probable schedule results but it is asimmetrical. So if you want to calculate which target dates and costs will have 50% probability then these dates and costs usually will be between most probable and pessimistic.
In Demo version of Spider Project that may be downloaded from http://www.spiderproject.ru/spider_e.php you may find an example of the probability calculations for the sample project including trend analysis. Check it.

Regards,
Vladimir

http://www.spiderproject.ru/index_e.htm

Actually there is not such thing as a baseline schedule if risk management approach is involved.
There are target dates and target costs for the whole project and major milestones. They may be established different ways.
1. Project Planner creates optimistic, most probable and pessimistic schedules. Then desired probability of meeting target dates and costs is determined and desired dates and costs of the project and major milestone are calculated. These dates are used in the contract negotiations. When target dates and costs are proposed project planner calculates probabilities of these targets achievement. If they are satisfactory then the contract may be signed and these target dates and costs determine what is used for performance estimating (baseline though not schedule). During project performance project planner estimates what is happening with the probabilities to meet targets and suggests correstive actions when success probability trends are negative.

2. When target dates and costs are determined initially then project planner solves another task. What resources are necessary to perform the project on time and within budget with the reasonable probability. If this is impossible then project planner shall inform project manager about dates and costs that are real. If the probable project results will be known ASAP then project manager has more chances to defend reasonable project schedule and budget changes.

Usually most probable schedule has low probability of success (around 30%) and thus should not be used as a baseline.

Vladimir

www.spiderproject.ru/index_e.htm

Mmmmmm.....whatever happened to accuracy?

Cheers,
Stuart

www.rosmartin.com