Download
a complete copy of this risk analysis resource for free here.
|
See also: Presenting results introduction, Model design introduction, Preparing a risk analysis report, Graphical descriptions of model outputs, Common elements of ModelRisk windows, About Vose - contacting us
The explanations below might give you an idea of the things we have in mind when designing and developing risk analysis software. More interestingly, maybe you find some tips in there that can increase your productivity when modeling.
Integrate Calculation interface
from ModelRisk
. Note that generally we do not recommend using 3D for presenting
graphs, as this rarely adds clarity.
When working with, interpreting and presenting complex models, clear illustrations can really make the difference. In the ModelRisk windows every element of a preview graph can be customized by right-clicking it. Also, preview graphs update dynamically as you change the parameters attached to it.
ModelRisk graphs can be copied to the clipboard as an image file. This
way they can be inserted in an Excel worksheet or in every other Windows
program. To copy an image to the clipboard, click the 
button and select As a bitmap.
By copying and pasting you can easily ModelRisk graphs in your PowerPoint presentation, web site, analysis report, etc...
The Vose View Formula console provides a convenient way to have a quick look at a distribution's shape, some summary statistics...
See also: Modeling with objects, Customizing ModelRisk, Common elements of ModelRisk windows
Apart from presentable graphics, ease of exporting to other programs and error handling (explained below), ModelRisk has a range of other features that can increase your productivity, accuracy and overall quality of models. This is what we mean by "workflow". If you take the time to really get to know your way around ModelRisk it pays off in the long run.
Where appropriate, you can alter parameter values / distribution boundaries, etc... By dragging a line with the mouse (e.g. in the Vose Splicing Distributions window).
Note that Excel has a Function Arguments window that can be a good reminder of what arguments a function takes.
ModelRisk adds several context-sensitive items to the context menu that appears when right-clicking a selection of spreadsheet cells.
The Modeling framework represents a novel way of thinking about mathematical modeling in a spreadsheet environment. Seen as objects, distributions become workable entities - sampling from them is only the start. To calculate the statistical moments of a distribution you call VoseMoments on a Distribution Object. For Aggregate modeling for insurance you call VoseAggregateMC with the Distribution Objects for claim size and frequency that you think are best.
Interesting for complex insurance modeling is also the VoseExpression function, that allows you to define complex expressions "on the fly", that can be used with ModelRisk's AggregateMC function.
Errors in ModelRisk are descriptive. When a cell contains a wrong formula or an impossible parameter value, ModelRisk returns the precise nature of the error.
For example, when one types in: =VoseNormal(10,-1), an error message will be returned saying:
Error: sigma must be >=0
because the second parameter of a Normal distribution (the standard deviation sigma) can not be negative.
If, however, a formula contains a erroneous ModelRisk formula in combination with other things, then the appropriate Excel error is returned. Let's say for example that the formula typed in is not =VoseNormal(10,-1) but =3 + VoseNormal(10,-1), then the generated error says:
#VALUE!
If you see that a cell with a ModelRisk function returns such an error in Excel, you may want to insert the ModelRisk function separately in another cell to view the error message generated by ModelRisk.
Important note: if a correct ModelRisk function is inserted into a cell and Excel returns the #VALUE! error, this may mean that the ModelRisk add-in is not loaded. Try reloading the ModelRisk add-in, then hit F9 to recalculate the spreadsheet.
If one wants to do a Cholesky decomposition of the following symmetric matrix:
the following error message is returned:
Error: Matrix is not positive definite
Or suppose one wants to find the value from a Pareto(3,1) distribution corresponding to a U-parameter of 0.4, but makes a typing mistake:
=VosePareto(3,1,4)
The error message is then:
Error: U must be on [0,1]