What is a Monte Carlo Simulation?
Вставка
- Опубліковано 6 чер 2024
- A Monte Carlo Simulation is a way of assessing the level of risk across a whole project. So, while you may not need to use this powerful methodology, it's vital knowledge for any project manager.
The Monte Carlo technique takes its names from the random chances at the gaming tables of Monte Carlo.
We most often use it to assess schedule risk, but we can also use it for budget risk.
How it works
For each activity on the project, estimate a range of probabilities for different possible durations. The set of durations and your estimates of their probability is the “probability distribution” for the activity durations.
The distribution we often use is called the “Beta Function”. But each task will have its own distribution of likelihoods of duration; some simple, some complex. Each will apply to different types of activity.
The Monte Carlo method calculates how long the total project will take, by adding up the durations of all of the activities each with a duration randomly assigned. The probability of each task’s duration is defined by the probability function in our curve. To simulate the random nature, the Monte Carlo simulation repeats the calculation, with a different set of random durations many, many times. This gives a distribution of all of the possible total project durations.
Monte Carlo is more than just a calculation
The calculation is the easy part. Making the estimates of the probability distribution for each task is time-consuming. The quality of the estimates you make will entirely determine the validity of the resulting distribution. If all that you do is use the same distribution for each activity, you will simply reproduce that pattern for the whole project.
There is another flaw in the Monte Carlo method, which renders its results less reliable than you might think. It contains the assumption that the probability distribution for each task is independent of all others.
Yet we know that a delay in one activity can often make others more likely to be delayed too. And, also, that one event can affect multiple tasks. As in all risk management, the strides towards rigor have only got us so far. There is no substitute for good judgment, and appropriate skepticism and caution.
Recommended Videos.
Carefully curated video recommendations for you:
- What is Project Risk Management?
... • What is Project Risk M...
- Project Risk Management - How to Manage Project Risk
... • Project Risk Managemen...
- Risk Management - 5 Tips to do it right
... • Risk Management - 5 Ti...
- What to Put in Your Risk Register (Risk Log)
... • What to Put in Your Ri...
- What is Risk Tolerance?
... • What is Risk Tolerance...
My basic introductory book to Project Management is How to Manage a Great Project (published by Pearson): geni.us/4Iz6M (affiliated link).
For more great Project Management videos, please subscribe to this channel: / @onlinepmcourses
For all our great Project Management articles and resources, please check out the OnlinePMCourses website: onlinepmcourses.com/
For basic Management Courses - free training hosted on UA-cam, with 2 new management lessons a week, check out our sister channel, Management Courses: / managementcourses
For more of our Project Management videos in themed collections, join our Free Academy of Project Management: onlinepmcourses.com/free-acad...
#Project #ProjectManagement #MonteCarlo
Showed this to my 9th grade class today- they really enjoyed it and learned a lot, thanks!
Glad it helped!
Thank you. I found that very interesting. I’ll have to watch it again.
Great - glad you liked it.
Thank you for explaining the Monte Carlo Simulation!
You're very welcome, Patrick.
When I first heard about this it seemed exotic and exciting - but hard to understand. In this video, I aim to make it easy to understand, so you can know if and when to use a Monte Carlo simulation.
🎲
Thank you, very clear explanation!
You're welcome.
Absolutely BRILLIANT explanation thank you so much!
Glad you enjoyed it!
"Absolutely briliant" such a British replay.🤣🤣
Thanks for clarifying
You're welcome.
Love Mike's videos
Thank you.
Is it possible to use monte carlo for binary data. I have kpi data which is pass or fail denoted by 1 or 0. Can i use it or it is only valid for normal distributions?
I think that, in principle, you can. But you only need to if P(1) is different for different elements, If the probabilities are all the same, then the MC result will just match P(1). If I remember my stats correctly (and I am no statistician), as you move toward a binary situation, the limit of a normal distribution is the Poisson distribution. I really do not. know if MC makes sense for the case you are interested in, I am afraid.
Nice sir
Thank you.
Hello again from Coastal Virginia. My latest graduate school course is Fundamentals of Business Analysis. Is this particular methodology related to sensitivity analysis and capital budgeting analysis? Thank you.
I am not knowledgeable in capital budgeting, but Monte Carlo Analysis is fundamentally a way to model the interactions of many estimates under statistical uncertainty. So you could use it for sensitivity analysis if you are interested in sensitivity to random variation. It's main use in project management (which is the only domain I know about) is for risk analysis.
@@Onlinepmcourses Thank you for the clarification.
Your video is the first to accurately, honestly describe Monte Carlo Simulations (MCS).
If one loads a MCS with poor data they will get an output of poor data.
MCS is just an arithmetic tool that calculates numbers you give it. Example:
The American presidential election of 2016 using MCS predicted right up until election day that Hillary Clinton would win the presidency with 92% certainty. But Donald Trump easily beat her. This is because those loading the data into the MCS were heavily biased for Clinton.
MCS give the "appearance" of razor sharp, mathematical accuracy but in reality their accuracy is driven by good Qualitative assumptions up front.
Indeed. MCS can give precision (a very sharp range) without accuracy (a sharp range around the wrong midpoint). As always, GIGO - Garbage in, Garbage out.
Dude ty for that input. I have to write a discussion on simulation models and how its used in real world situations. Its due today and you come in clutch
Beautiful
Thank you.
Can we use it for estimation of uncertainty of measurement?
No, the uncertainties of individual estimates are inputs. The result is the uncertainty in the outcomes that combine the individual events.
For what purpose is it useful to learn?
Useful to learn... anything, or useful to learn this?
We learn to become better people: better project managers, better professionals, better members of our society. This knowledge (MC simulations) and any other knowledge will either help us to do our jobs today, or help us frame our choices tomorrow. maybe both.
And, as the QI organization points out... 'Whatever is interesting we are interested in. Whatever is not interesting, we are even more interested in. Everything is interesting if looked at in the right way.'
Thanks :)
You're welcome.
I am still confused.
What's confusing you? If you're more specific, I may be able to help.
Yes it is nice and at the same time it deos not answer the questions "What is a Monte Carlo Simulation". I am clearly able to understand that questions when I am looking at a circle and a square with random arrows.
I'm sorry.
A monte carlo simulation creates many random results to give the range of outcomes. What do I not cover that you need to know? I don't get what you mean by 'a circle and a square with random arrows'.
@@Onlinepmcourses I actually spend time into looking at the reason. Now I see that from now on need to pay attention to not only the title "What is a Monte Carlo Simulation" also hash tags. This wasn't what I was looking. (Project Management) So it really has nothing to do with you or the video. thanks,
@@Ch-do6dg I have many videos covering all aspects of Project Management. This one covers a fine detail of Project Risk Management. For a general intro, I suggest you start with my Project Fundamentals playlist: ua-cam.com/play/PLsz8d8r2a996dLF6xCFwYdLxxlUK4IXPt.html
@@Onlinepmcourses ok now you loved. You liked and I liked. You commented then I commented +++ everyone is happy. Congrats you are doing a great job T growing your channel. Like and love.
@@Ch-do6dg Thank you.
Sorry, switched off after the fifth utterance skedual. Is that similar to a schedule?
Yes, it's the same thing. You say 'tomayto' I say 'Tomahto'. If that bothers you enough to switch off, I'm sorry, but I think it says more about you than it does about me!
Skedule and shedual are both ok to say when you know monte Carlo
@@ryanhamdogg9762 I tend tp use the US form here - largely because I like it better: schedule/school
Not so agree with the shortfall you mentioned. Correlations could be defined between activities or cost elements. That means, dependent activities or cost elements will be sampled based on their correlations in each iteration. i.e. price and demand have negative correlation in real life and in any iteration they should not be sampled from their high values. A good scheduler or cost estimator or risk expert could take correlations into considerations when building a model/schedule.
You are right that we should take correlations into account. My point (if I have correctly identified where your concern lies) is that simplistic application of ranges in the MC method can produce nonsense. Done well, of course, we can work around that.
2X
?
@@Onlinepmcourses I think he meant he thinks 2x video speed is better
@@thecrown4423 Tx. For people with fluent English, faster is fine. I speak slowly and clearly because a lot of my audience use business English as a 2nd or 3rd language.
Very sorry sire, but I didn't get a clue about your talking, please make sure that some of your close people understand you, before making any video
Well, I do take understanding very seriously. And you are the first person with this comment. So, all I can say is that I continue to work hard to make my explanations easy to understand. And, 'sorry'.