Guild of Project Controls: Compendium | Roles | Assessment | Certifications | Membership

Tips on using this forum..

(1) Explain your problem, don't simply post "This isn't working". What were you doing when you faced the problem? What have you tried to resolve - did you look for a solution using "Search" ? Has it happened just once or several times?

(2) It's also good to get feedback when a solution is found, return to the original post to explain how it was resolved so that more people can also use the results.

Practical questions on working with optimistic, expected and pessimistic schedule

15 replies [Last post]
Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Hi, everbody.

I am just looking at Spiders risk analysis section.

As I understand, the idea is to have 3 schedules: optimistic, expected and pessimistic.

The parctical question is on how to maintail them all 3 in parallel.

I can imagine how I will start it: create Probable schedule, then save it with siffix_pes and and change durations to make it possimicrics and then save with the suffic _opt and make it optimistic.

The question is on how to maintain all 3 schedules going forward:

How to apply progress to al 3 of them?

How to make changes to all 3 of them (e.g. add resource assignment, change dependency)?

What is the expected approach here? Is it really expected, that user would replicate every single schedule update to 3 schedules?

 

Another questions:

Can optimistic, expected and pessimistic scenarios have different set of activities or list of activities must be the same?

 

Replies

Yes, Monte Carlo provides more information on project behavior and this is why we suggest to use both methods in parallel.

These methods provide the same information on success probability trends and this is the main information required for making informed management decisions. But for large models Monte Carlo requires a lot of time, just imagine calculating 10 000 times resource constrained schedules and budgets for projects (portfolios) consisting of tens of thousands or hundreds thousands activities. Three scenarios calculations are much faster (3 schedules only). So we recommend to use MC from time to time to verify Three scenarios results and to get additional information on such things like criticality indexes, scatter diagram, tornado chart but use Three scenarios for day by day management calculating success probability trends with each project update.

Besides, as I wrote earlier, we recommend to use Optimistic scenario for management of project workforce and subcontractors and manage project estimating project buffers consumption by analyzing success probability trends.

Rafael Davila
User offline. Last seen 11 hours 39 min ago. Offline
Joined: 1 Mar 2004
Posts: 5229

By the way, with 3 scenarios approach you will not get criticality index while Monte Carlo will.

Each method has its advantages. 

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Vladimir, thanks. Understood.

I can see,  that Spider has a lot of sophisticated functionality built in (not that I did not know this before) 

"Normal" duration is not linked to MC simulation.

The user may want to create the schedule where "normal" duration is between expected and pessimistic, or optimistic depending on the schedule purpose and it will not change MC results. Using formulas it is easy to make normal duration equal to optimistic, most probable and pessimistic or any combination like (expected + pessimistic)/2. The same with costs and other parameters.

Remember that we recommend to use optimistic estimates for workforce management, but set targets using risk simulation and selecting reasonably achievable values (usually with 70-80% probabilities to be met.

Contract values of some parameters may be larger than pessimistic on some activities. Additional reserves may be useful, isn't it?

Some managers prefer to set super optimistic plans that we do not approve but do not ban.

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Vladimir, another thing:

So, now in schedule I have 4 estimates:

Duration Remaining - Opt

Duration Remaining - Exp

Duration Remaining - Pes

and just "normal" Duration Remaining

 

How does in case of monte-Carlo normal Duration Remaining is taken into account? I also found, that I can enter it to be outside of optimistic - pessimistic range. Which is a bit strange

 

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Vladimir, another thing:

So, now in schedule I have 4 estimates:

Duration Remaining - Opt

Duration Remaining - Exp

Duration Remaining - Pes

and just "normal" Duration Remaining

 

How does in case of monte-Carlo normal Duration Remaining is taken into account? I also found, that I can enter it to be outside of optimistic - pessimistic range. Which is a bit strange

 

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

If you want to add ranges select the column that you want to show and tick Monte Carlo at lower right corner of the dialog box.

Vladimir, thanks, I found them!

Monte Carlo simulates both risk events and uncertainties.

If some parameter (activity duration, cost, resource productivity, etc.) is not certain it is necessary to define the range (optimistic, most probable and pessimistic values). If parameter is certain it is not needed.

If risk event may happen and change future plans define “trigger” activity, its probability and probabilistic branches that may follow risk event (trigger) and their probabilities (with 100% probability of allbranches).

Monte Carlo model may be created withot creating three scenarios - just enter ranges in MC fields in the project model and create triggers for modelling risk events. If you want to add ranges select the column that you want to show and tick Monte Carlo at lower right corner of the dialog box.

If Monte Carlo model is created Spider Project can create three scenarios from this model automatically and vice versa – create Monte Carlo model if three scenarios were defined (triggers shall be added manually). There are menu items for this. There are examples in Demo for modelling risk events (project SD) and parallel three scenarios and Monte Carlo (road construction).

Don't forget to add risk responses if some conditions are not met. We use special activities called switches when one branch shall be selected instead of another if some condition is not met. Look at SD project for an example.

Rafael Davila
User offline. Last seen 11 hours 39 min ago. Offline
Joined: 1 Mar 2004
Posts: 5229

Monte Carlo vs 3 Scenarios Approach:

  • Pessimistic might be a schedule that considers the possibility of a Hurricane, this require a different set of activities absent in optimistic schedule and perhaps a different work sequence.
  • If a project is running late, management does not sit idly by, but they do something!

https://www.pmi.org/learning/library/monte-carlo-simulations-project-networks-mislead-2537

  • Correlation (Risk Drivers) between activities, Probabilistic Branching, Probabilistic Calendars and a few I might be missing are hard and perhaps in some cases impossible to model in Monte Carlo.  
  • Running Monte Carlo in very large projects as well on a portfolio consisting of several large interconnected projects takes much computer time.

All of the above can be easily tackled by 3 scenarios approach.

Maybe the claim that Monte Carlo is better is just a perception.

Rafael Davila
User offline. Last seen 11 hours 39 min ago. Offline
Joined: 1 Mar 2004
Posts: 5229

To show them select the required fields.

opt-exp-pess

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Incorrect, you can use Monte Carlo without 3 Scenarios, but you have to enter the 3 estimates for each parameter manually. In a resource and cost loaded schedule, by setting the 3 estimates for resource productivities (productivity type activities) or 3 estimates for activity durations (duration type activities) you can run Monte Carlo.

 

Bogdan, thanks. Which fields are these (e.g. for 3 estimates of duration)? Can't find them

Bogdan Leonte
User offline. Last seen 2 weeks 6 days ago. Offline
Joined: 18 Aug 2012
Posts: 284
  • Is it correct, that Monte Carlo does not work without three scenarious

Incorrect, you can use Monte Carlo without 3 Scenarios, but you have to enter the 3 estimates for each parameter manually. In a resource and cost loaded schedule, by setting the 3 estimates for resource productivities (productivity type activities) or 3 estimates for activity durations (duration type activities) you can run Monte Carlo.

Also Spider allows to create 3 Scenarios from Monte Carlo by saving 3 different versions of the project and then setting the parameters from one of the 3 estimates in Monte Carlo.

Evgeny Z.
User offline. Last seen 35 weeks 5 days ago. Offline
Joined: 13 Jan 2008
Posts: 442
Groups: None

Bogdan, Vladimir, thanks.

questions:

RE Spider Project includes two methods of risk analysis. One is Monte Carlo and another one - Three scenarios.

Is it correct, that Monte Carlo does not work without three scenarious, because it still needs to know pessimistics, most likley and pessemistic estimations for every task? the only way I found to communicate to Monte Carlo this iformation is to do "Import Monte Carlo Data from three scenarious"

 

 RE: Three versions may include different activities but we recommend to use the same set in all three.

If some activity is not present in all 3 versions, how then Nonte Carlo algorithm works?

Monte-carlo algorith needs to know pessimistics, most likley and pessemistic estimations for every task in order to be able to generate result? Or am I a missing something?

Evgeny,

Spider Project includes two methods of risk analysis. One is Monte Carlo and another one - Three scenarios.

Three scenarios is much faster than Monte Carlo though not as accurate but we suggest to manage by analyzing success probability trends and they are practically the same in both methods.

We suggest to use optimistic scenario for management of project workforce but set target dates and costs that have sufficient probabilities to be met.

When you enter actual data in one of three scenarios (usually optimistic) they will go all three versions that will be synchronized automatically. Three versions may include different activities but we recommend to use the same set in all three. Activities that are absent in current version have zero duration and cost. This way they will not be missed when actual data are entered and inform managers of their potential existence.

It is easy to transfer changes in one version to others using reference-books. For an example: if you changed activity dependencies in one version create Links reference-book (selecting in standard reference-books list) and apply to other versions.

This approach to risk simulation found one more application. General Contractors create three project versions: expected for themselves, optimistic for managing workforce and subcontractors, pessimistic for contract management. Actual data are the same and so all three schedules are automatically updated when they are entered.

An approach that we use for project management when targets are set basing on required success probabilities, optimistic schedule is used for resource management, buffers (the difference between target values and optimistic planned value) consumption is estimated by analyzing success probability trends. Negative trends indicate that buffers are consumed faster than expected and corrective actions may be required.

Bogdan Leonte
User offline. Last seen 2 weeks 6 days ago. Offline
Joined: 18 Aug 2012
Posts: 284

Hello Evgeny,

Vladimir will probably go into more detail than I but the ideea of the 3 Scenarios Method is as follows:

1. Create Optimisti, Probable, Pessimistic scenarios; each scenario can have different activities, but phases should have the same codes;

2. Define target for project parameters (Duration, Finish Date, Costs, Revenue, etc.), for project phase/for the whole project or both;

3. There is no baseline, all your evaluation regarding project performance are based on the probability trends to achieve the target parameters;

4. Start updating the project from the Optimistic scenario and Spider synchronize the actual data in all 3 scenarios; all 3 scenarios will increase in version in order to have a performance archive;

5. Recalculate project and run risk analysis again in order to recalculate probability of target project parameters.

And this is it, in brought lines.

Best regards,

Bogdan