This tutorial explains only the most basic features of Riscue. By walking trough a very simple example, we demonstrate how to build a model, run simulations, and analyze results.
The first thing you see when you start the Riscue application, is a screen like this:
This is the Riscue application window. At the top of this window, you will see a familiar menu bar. Below that you will find a toolbar with various buttons. At the center of the screen, you will find the model frame. This is where you will create a graphical representation of your risk model.
1. Create nodes and edges
All Riscue models consist of nodes. A node represents some quantity of interest, say e.g., the cost of something. To create a node, click on one of the node buttons in the toolbar:
The different buttons correspond to different mathematical functions associated initially with the node. Thus, e.g., to create a node which adds input, press the button with the "plus" icon. The mathematical properties of a node can be changed to suit our needs later, so the initial choice (i.e., which button you use to create a node) is not all that important.
For this example, we start out by creating three "independent" nodes: A, B, and C. This is done by clicking the leftmost button (the one with a "dot" icon) three times. Since the costs are considered to be uncertain, we will associate probability distributions with each cost node. As a shortcut, this can be done initially by pressing the "Shift" key on the keyboard while clicking the toolbar button. The three nodes represent the cost of three items. When a new node is created, it is named "New Node" by default. The name can be edited, by clicking the title field. Change the names of the three nodes to A, B, and C, respectively.
Next we want to add the values from the three nodes. This is done by creating a "sum" node by clicking the button with the "plus" icon. Since the "sum" node simply adds values, no additional uncertainty is introduced in this node. Thus, you should not press the shift key this time. Call the new node SUM, and drag it below the three cost nodes. Then select nodes A, B, and C, and drag them onto SUM. As a result, you should see three edges (arrows) pointing from A, B, and C towards SUM. The edges represent dependencies in the model. Now the SUM node will take values from the nodes A, B, and C, during a simulation, and add these values together. The resulting diagram should look like the figure below. Note that the three uncertain nodes, A, B, and C are drawn as single circles, while the SUM node has a double circle. A single circle indicates that the node is uncertain, while a double circle are used for deterministic nodes.
Before we proceed, note that you can delete objects again by selecting them and pressing the "trash" button in the toolbar (or the "delete" key on the keyboard).
Nodes are selected by clicking their icon once. Edges are selected by clicking their arrow once. As soon as you select one object, any previously selected object will be deselected. If you need to select more than one object, press the "Shift" key on the keyboard while clicking the objects. It is also possible to select more than one object by using a "selection rectangle". To do so, press the mouse button somewhere in the white area of the window, and drag the mouse while keeping the button pressed. As a result you should see a gray rectangle. Make sure that this rectangle contains the objects you want to select, and release the mouse button. Selected objects are drawn in a darker color:
Note also that all operations can be reversed by selecting the "Undo" command from the "Edit" menu. In fact Riscue lets you undo (and redo) the last 10 operations.
2. Editing node contents
While all nodes have some initial default mathematical definition, one needs to modify this immediately to obtain a useful model. The mathematical definition of a node is edited by "opening" the node. This is done by double-clicking it. As a result, you should see the following dialog box:
The dialog box contains a list of "Function categories", and a list of "Functions". The right-hand list changes as you select different categories from the left-hand list. When you select a function from the right-hand list, you will see a text string below, indicating the syntax for this function:
If you want more information about the selected function, you can press the little "..." button on the right-hand side of this text string. If you do this, a window containing information about this function (and other similar functions) will pop up:
Below the two list, you will find a text field where you can enter the mathematical definition of the node. The definition can be an arbitrarily complex combination of the different predefined functions, nested inside each other. The syntax is similar to standard spreadsheet syntax. The arguments of a function are separated by semicolons. Each argument can be either a number or another function.
A special function, called INPUT(), allows you to use each of the individual values coming in to the node through its input edges. If the node has more than one such edge, it is important to be able to distinguish the different input edges from each other. This is done by including the index of edge as an argument to the INPUT()-function. Thus, e.g., INPUT(2) refers to the input value coming through the second edge, where the ordering of the edges depends on the order in which the edges were created.
When using the INPUT()-function, it may be important to keep track of the ordering of the input edges of a node. There are two ways of doing this. In the graphical view, you can click the "Show edge indices" button in the toolbar. As a result the indices of the edges are shown in purple.
It is also possible to keep track of the ordering from inside the node dialog box. To do so select "Node input" in the function category list. The function list will then display the names of the nodes corresponding to the input edges. By selecting one of the node names, the corresponding INPUT()-function will be displayed below.
In addition to the INPUT()-function there are also many other functions that use values coming from other nodes through edges as input. E.g., the INPUTSUM()-function is calculated by adding all input values.
3. Working with folders and aliases
In a large risk model, it soon becomes impossible to have include all nodes in a single view. Riscue then offers the possibility of partitioning the models into a set of submodels. These submodels can be organized hierarchically in "folders" similar to how you organize files on a hard disk. To create a folder click the folder button in the toolbar:
You can then move a group of nodes along with the edges connecting them into a folder by dragging and dropping them onto the folder. To see the contents of a folder, "open" it by double clicking it. To "jump" out of a folder again, use the navigation buttons:
The leftmost button enables you to travel across the folder hierarchy, the middle button brings you to the top level of the model, while the rightmost button brings you to the folder level above the current one. You can also use these buttons to move objects to other folders. Select the objects you want to move and press one of the buttons. Note that edges cannot be separated from their end nodes. If you move two nodes connected with an edge, the edge is moved along with the nodes. If you try to move a node connected to another node which is not moved, you will get an error message explaining that you have to disconnect the node first by deleting edges.
Now, creating a set of submodels and placing them in a hierarchy of folders is all very fine. But how do you connect the various submodels so they can work together as a giant model? The answer is to use "aliases" (shortcut nodes). To create an alias of a node, select it and press the alias button in the toolbar:
This way you get a new node which is a shortcut to the original node. The alias node can be moved and hooked up to other submodels:
4. Running a simulation
When all aspects are taken into account in your model, you are ready to run a simulation. A simulation can be done on any submodel in your model. Riscue will always figure out which nodes it needs to calculate to run the simulation. Before you start a simulation, you must decide how many iterations you want to run. This is entered in the field in the lower region of the model window:
As soon as you have done that, you can run a simulation by clicking the "run" button in the toolbar:
Riscue will then run the specified number of iterations on the model, and present the results in a new window. The first view you get of the results, is a set of new nodes, named after the nodes you ran the simulation on:
Note that these nodes are labelled "Data". We will refer to them as "data nodes". To see the actual simulation results, you can open them by double clicking them:
The resulting dialog box contains two lists: Fractiles, containing estimated fractile values (or percentile values), and Raw data, containing all the values calculated during the simulations. The number of values in the raw data list depends on the number of iterations you entered in the "#Simulations" field.
5. Analyzing results
Besides providing you the ability to stare at the simulation results, Riscue offers a variety of statistical plots. The most commonly used plot is the S-curve, i.e., the estimated cumulative probability distribution. To create such a plot, select the data nodes you want to include in the plot, and select "S-curve" from the "Statistics" menu:
As a result, you should see a new window:
It is possible to include more than one S-curve in the same diagram. Just select a group of data nodes, and use the same menu command. It is also possible to display results in a table format. Note that data in tables can be copied and pasted into e.g., a spreadsheet for further analysis.