Creates a workspace database. The user is prompted to make a connection to the database where the workspace will live and for the template file (default provided). The database must already exist. If it doesn't the user can use newMDBWorkspace to create a new Access database or newSQLWorkspace to create a database in SQL Server. For other database platforms (Oracle, DB2 etc.) you shoud contact your database administrator. Unlucky you! The new database is not selected as the current workspace. Use openWorkspace to select the new database as the current workspace.
Highlight a loss development triangle (rows years, columns development), excluding the row and column headers, and click development. Will produce a triangle of link ratios (using formulas), the usual averages, a row for your selections, and a blank for you to input exposures or premium, and your selected ultimate. Will over-write whatever is below the triangle, so be careful! Incredibly useful.
Highlight a triangle, e.g. of loss development factors. Will just put the block of averages from the Development command below the triangle. Like development, but only adds averages to a triangle of ratios.
Set up Variability of Reserve analysis. As development, except it expands the set up to include bootstrap re-sampling off the triangle of link ratios. Allows you to determine rudimentary confidence intervals for ultimates. VOR = variability of reserves.
Complete the bootstrapping simulation. Perform after you have made your selections (in orange). Will simulate an entire run-off from the triangle, with user input correlation between the accident years.
Sets up a blank where you input the mean, it computes the parameters (in this case just equal to the mean) and sets up formulas for computing Pr(X=0), Pr(X=1), and so forth.
Sets up a blank, inputs mean and CV, outputs lognormal mu and sigma parameters. Shows you skewness. Blank includes f(x), F(x), LEV(x), Expense(x) = expected excess cost, and ME(x) = mean residual lifetime = conditional execess cost.
As above for shifted lognormal, T+LN(mu, sigma), where T is a constant shift. Using three parameters allows you to match the mean, CV and skewness. This is typically a very good fit to the true aggregate, once the claim count is above 20 or so. Highly recommended approach, combined with the various tools for estimating the moments of an aggregate distribution.
As for lognormal and shifted logormal, only for the shifted gamma distribution. Like the shifted lognormal distribution, the shifted gamma allows you to match the first three moments of a distribution. Closed form solution for parameters. Recommended in Actuarial Math book as a good approximation to aggregate distributions. Uses the Klugman, Panjer, Willmot parameterization.
Severity blank for the transformed gamma distribution, a three parameter distribution without a shifting parameter (hence four parameter with a shifting distribution). Needs mathFunctions.dll. KPW param.
Very useful function. Select a column of large losses (in any order), and another shorter n x 1 column range. Tool will create a histogram using n buckets (it computes a resaonable size for each bucket from the input data. Also adds a chart to the worksheet plotting out the histogram points. Data is input as values, not formulas, so the histogram is static and will not recompute.
New version guesses a reasonable n, computes mean, cv and skewness of input sample, and provides a shifted lognormal fit.
Select a column of losses and execute. Adds percent of claim counts and percent of total losses to the right and a plot of these. Allows you to answer questions like "what percentage of total losses comes from the 5% largest counts?".
Call back to produce a labelled x-y plot. Highlight a range of (label, x, y) , sorted by label. Makes a graph of x,y points colored by label values. Only accepts n x 3 input.
Makes a tex string command out of the selection table. Assumes the data is mostly numbers and hence right justifies. Middle vertical lines and all horizontal lines. Answer output to clipboard. Input is Caption Row of l, r, c to determine alignments Heading row data data data Version 1.0, Stephen Mildenhall, January 2003.
Deletes (and lists) all invalid names in the active workbook. Invalid ranges appear when the range they refer to is deleted. This is a very useful function.
Copies the values, formulas and number formats of a region into a cache. On second call the contents is written to the destination cells as a duplicate. Thus cell references are not moved as in a usual copy and paste operation.
Highlight a column. Goes down each cell in the column and deletes the entire row if that cell is empty. This is useful for getting rid of hard coded sub-total rows, for example.
Applies a random shuffle to the range. Useful for performing a cheap-man's convolution: get two samples and shuffle them. Then add. Uses algorithm from Knuth, Seminumerical Algorithms.
Call back to produce a labelled x-y plot. Highlight a range of (label, x, y) , sorted by label. Makes a graph of x,y points colored by label values. Only accepts n x 3 input.
Select an m x n input range and click 3D Plot to activate a beautiful 3-D graphic of the selected range. Minimal scaling options on the x and y axis and generic labelling, but great coloring, contouring and interactive features. Ability to take a "snapshot" of the graphics, which is inserted into the spreadsheet (at the top left hand corner of the selected range) as a JPG file. Controls in the graphics window include:
left mouse - rotate image
right mouse - zoom in (upper half of image) and out (lower half)
w - toggle to wireframe view
s - toggle back to surface view
j - joystick mode
t - track ball mode
Some graphics cards aren't really upto this control; you need very good OpenGL support.
Run auto-documentation creation. Creates a file like the one you are reading from comments in the VBA source of your selected workbook. Module comments are generated from the first comments in each module (until the first new line). Function and Subroutine comments are the comments immediately following the declaration until the first non-comment line. The comments can include HTML markup. It is copied verbatim into the documentation file. Change the base directory name and module name as appropriate to run on your project. Make sure that >, < and & are appropriately HTMLized!
Consolidates string values of all cells from the active cell to the next empty cell into the active cell and deletes the contents of the remaining cells.
