Skip to content

Home / Monte Carlo / Additional Tools / Sampling Matrices

Sampling Matrices

From Excel click...

QXL Monte Carlo New > Additional Tools > Sampling Matrices

Writes a new worksheet holding a sampling matrix: one column for each factor and one row for each run, built by Descriptive Sampling, Latin Hypercube Sampling or Sobol Sampling.

This is not a simulation. It reads no model, it marks nothing, and it needs no marked workbook. The matrix is a set of input values written to a worksheet for you to use however you want.

What the dialog asks for

Four things, and nothing else: the sampling method, the number of factors, the number of rows, and the seed. Every control is described on the Options page.

What is written to the worksheet

The new sheet has four parts, with the labels down the left and one column for each factor across.

Matrix Settings. Four lines recording what built the matrix: Method, NumRows, Factors and Seed. This is the record of what you asked for, so a sheet can be read months later and rebuilt exactly.

The factor block. One column per factor, with four rows:

Row What it holds
Factor the factor's name
Distribution Normal or Uniform, chosen from a list on the cell. New sheets open on Normal
Mean (Normal) or Low (Uniform) the first parameter, read according to the distribution above it
Std Dev (Normal) or High (Uniform) the second parameter, read the same way

Coded values (between 0 and 1). The matrix the sampling method produced, one row per run. These are the numbers the method itself is responsible for, and they do not depend on the distribution or the parameters you set.

Sampling Matrix (actual units). The same matrix converted into the units of each factor, by an ordinary Excel formula that is on the sheet and can be read in the formula bar. A Normal factor uses NORMINV against that factor's mean and standard deviation. A Uniform factor is placed between its Low and its High. Because it is a live formula and not a pasted number, changing a distribution or a parameter updates the actual-units block straight away, with no need to build the matrix again.

The pale yellow cells are yours to edit. They are the factor names, the distribution choices and the two parameter rows.

Limits

Descriptive Sampling and Latin Hypercube Sampling Sobol Sampling
Rows 2 to 20,000 a power of two, 16 to 32,768
Factors 1 to 250 1 to 1,000

A number outside its limit is refused by name and never quietly changed, so the matrix you get is always the one you asked for. Under Sobol the row count is chosen from a list rather than typed, because only a power of two is offered.

The seed

The same seed builds the same matrix. Every whole number from 0 to 4,294,967,295 is accepted, 0 included, and it is used exactly as given.