## 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.

# Risk analysis in Spider

Spider has a nice feature of probabilistic analysis of project completion date, cost etc. It is based on optimistic, pessimistic and most probable schedule scenarios.

It is clear, that probability to meet optimistic scenario is 0%, pessimistic is 100%. However the question is: how does it calculate probabilities for values in between optimistic and pessimistic?

In my understanding, to build correct finish date probability distribution, Spider should assume some probability distribution function for each of the parameters (e.g. durations) for which optimistic and pessimistic values are different from each other (e.g. beta-distribution or triangle distribution). Then use something like Monte Carlo analysis to calculate the entire schedule probability distribution. But something makes me believe, that Spider is not doing Monte Carlo analysis. So, how is the end schedule probability distribution calculated, does it just assume certain probability distribution function of finish date of entire project?

Evgeny,

you can reuse the slides from my presentations.

I don't see exact pictures in your post but yes, the presentations were illustrated by Spider screenshots (including Monte Carlo probability distributions and Scatter Diagram) though I removed those parts that show that they were made by Spider Project.

Presentations discuss methods and approaches regardless of the tools that are used.

Best Regards,

Vladimir

Vladimir,

Thank you for links to presentations. I scanned through them. I will need to look deeper, but for now I found them very interesting and informative even from general point of view, not to mention Spider application. Nice to know, that Spider has a solid structural theory/philosophy behind it.

If I have further questions on these, I will ask.

Can I reuse some of slides if I ever need to explain people the main ideas behind the scheduling?

Will be looking for demo version with Monte Carlo analysis in it. Are the pictures below from presentations made with Spider Monte Carlo?

--

--

Evgeny,

actually Spider does both - three scenarios analysis and Monte Carlo analysis though we did not include it in Demo yet.

In three scenarios approach that is much easier and more practical you know not only points with 0 and 100% probabilities but also maximum (most probable version) of the distribution. Basing on these three figures (not two!) Spider restores probability curve using some default rule. We do not pretend that this distribution is always accurate and actually made it wider than will be obtained if to apply Monte Carlo expecting that people will miss some factors. The main idea under this - accuracy is not that valuable as Precision. With all assumptions on initial values (triangle, beta or other distributions - we prefer lognormal) initial data usually are not accurate at all. Trying to achieve higher accuracy with modelling is not practical and does not make much sense. Probability trends give us much more information than probability values.

Distributions obtained by three scenarios approach may be not accurate (usually show less probability to achive targets than Monte Carlo) but the trends information will be accurate and will show if contingency reserves are used faster or slower than expected. This trend information is most valuable for decision making. Monte Carlo simulation provides more accurate information but from one simulation to another the probability to meet project targets may change even for the same project if the number of iterations is not huge. So small changes in trends do not provide us with the reliable information on success probability trends.

Besides be aware of the problems with application of Monte Carlo add-in software to projects with constrained resources that are discussed here: http://spiderproject.com/images/img/pdf/Project%20Control%20Methodologies.pdf

Spider Monte Carlo does not have these problems but still MC application for large projects with limited resources requires a lot of time for calculations. Since probability analysis shall be fast if to do it each time when project schedule is updated three scenarios is more practical. Besides, in any case we need optimistic schedule for resource management. So even applying MC Spider permits to create optimistic schedule basing on MC initial data. Managing two or three schedules in parallel is practical and useful and Spider supports this approach. In practice we recommend to use three scenarios for everyday management and from time to time use Monte Carlo for validating.

Read presentations for deeper understanding of Spider approaches and wait for Spider Demo with advanced Monte Carlo features that will be published soon.

Best Regards,

Vladimir

## Replies