Hey guys i'm a big fan of this forum you've helped me out a lot. Just wondering if anybody knows how I can get Visual Basic to define a custom distribution. In...
Ok, the basic help from the manual first. Defining custom distributions To define a custom distribution, your macro program must first set the cell selection...
Hello, I am fairly new to Crystal Ball so this may be an easy question to answer. I am trying to find a way to model conditional probability. The easiest...
Joe, In Crystal Ball, define an assumption using a hypergeometric distribution. This distribution models the situation you describe (sampling without...
Forgive my ignorance on this but what is (and how do I access) ModelAssist subject M0335 and M0124? Thanks David ... mistakes' ... I'll be ... the ... value of...
Hi, I need to use random values from a distribution more than once within each trial. For example, I have an assumption X, with distibution D. For each trial,...
I am trying to only run some of the forecasts that I have defined. However, when I try to freeze a range of these forecasts it does not actually freeze them,...
Hi David, The easiest way would be to download a trial version of ModelAssist for Crystal Ball at www.risk-modelling.com/software.htm. It contains hundreds of ...
I have substantial variance in the performance of Crystal Ball. My model is large with about 500 assumptions and about 2000 calculations. Sometimes it runs...
Andrew, It sounds like you're running a large number of trials (e.g. 100,000+). This can really eat into your RAM memory as, by default, Crystal Ball will...
It may be helpful to you (or not :-)) but have a look at message #1496. I think I was trying to do much the same sort of thing - the tips recorded are what I...
Hi, Lets say that 'risk' is a random variable that I'm trying to model. 'risk' is also affected by 'pplNumber' and 'jobSize'. I have tons of historical data...
It sounds like a two-dimensional problem, or maybe I misunderstand. Assuming there are two main dimensions (pplNumber and jobSize), you could break pplNumber...
Hi, My company is doing an Investment Evaluation for a mining expansion project. We have prepared a cash flow over a 20 year period and calculated the NPV. I...
Tornado diagrams are an excellent tool when you have 2000 assumption variables. Here are good things to do with a tornado diagram: o Check all the large bars...
Dear All, Tornado diagrams are indeed a great tool. Two comments: 1. The strength of tornado diagrams is that they show how a certain variable (i.e. the...
Hi Eric / Huybert, Thanks for the info! I spoke to an agent of crystal ball in South Africa and she said I shouldn't use the tornado chart over a time series....
Another option might be to use multiple regression analysis to forecast the risk in terms of pplNumber and jobSize (as the independent variables). The...
Hi Jenny, When you go through the 3-step process of running a Tornado Chart, in Step 3 you see that there are two Tornado Methods (upper right corner). When...
Hi, Future trials in my simulation depend on the results of previous trials. I want to run 20 trials to initialize my simulation (get it in a 'stable' state)....
I'm afraid I don't know the proper answer to your question; but here are some possible workarounds: * Multiply everything by zero if you are in iteration 1-15,...
To add to Eric's suggestions, another possibility is to use the Extract Data tool to extract the forecast results after running the simulation, delete the...
Hi, Eliminating crystal ball analysis at the end is out of the question for me. I won't be able to code for things such as sesitivity analysis before the...
Thanks for the prompt reply and your suggestions. Most of my simulation is built via VBA macro modules. I'm just using crystal ball for its statistical tools. ...
I would think that this situation might indicate that there is a problem in the model, and it needs some re-thinking. Theoretically, trials in Monte Carlo (MC)...
It sounds to me like Vivek is running a Markov chain Monte Carlo analysis. MCMC does probabilistic inference; fleshing out the joint distribution from a...
Hi Vivek, To add to Samik's comments, you may indeed have to think if your model ('linking iterations') is conceptually valid. A bit more precise descriptions...
Hi guys. I found a solution to the problem. I tried freezing the forcast for the first 15 trials as someone suggested, but I couldn't get it to work. What i...
Hi Vivek, Thank you for your more detailed explanation. It would still be difficult to conclude of course if you model is valid or not, but it seems that you...
Thanks, I'll keep that in mind. vivek. ... difficult to ... you are ... careful with ... in ... overall ... but ... spit out ... bit more ... run it ... Behalf...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
