Specify the parameters: low, high, average and set size.
Set size is limited to the number of cells in a row, however, with
minor modifications the set size could be increased to the number
of cells in a column.
--
Jim Cone
San Francisco, USA
http://www.realezsites.com/bus/primitivesoftware
(Excel Add-ins / Excel Programming)


Sub MakeNumbers()
Call FindNumbersThatAverage(dblLow:=200, dblHigh:=800, dblTarget:=400, lngSetSize:=100)
End Sub

Function FindNumbersThatAverage(ByRef dblLow As Double, ByRef dblHigh As Double, _
ByRef dblTarget As Double, ByRef lngSetSize As Long)
' Provides random numbers that average a predetermined amount.
' May 29, 2005 - Jim Cone - San Francisco, USA
' Sept 27, 2007 - Radically revised by Jim Cone.
Dim j As Long
Dim lngN As Long
Dim lngTemp As Long
Dim lngArray() As Long
Dim dblAdjust As Double
ReDim lngArray(1 To lngSetSize)

'Sanity check
If dblLow dblTarget Or dblHigh < dblTarget Then Exit Function
'If target value is not centered between high and low values then this
'allocates the random values proportionally.
dblAdjust = (dblHigh - dblTarget) / (dblHigh - dblLow)

Randomize
For lngN = 1 To (dblAdjust * lngSetSize)
'Get random values between the target and low parameter.
lngTemp = Int(Rnd * (dblTarget - dblLow) + dblLow)
'Place random values in the array
lngArray(lngN) = lngTemp
Next 'lngN

Randomize
For lngN = (dblAdjust * lngSetSize + 1) To lngSetSize
'Get random values between the target and high parameter.
lngTemp = Int(Rnd * (dblHigh - dblTarget) + dblTarget)
'Place random values in the array
lngArray(lngN) = lngTemp
Next 'lngN

'Do a random sort on the array values.
Randomize
For lngN = lngSetSize To 1 Step -1
j = Int(Rnd * lngN + 1)
lngTemp = lngArray(lngN)
lngArray(lngN) = lngArray(j)
lngArray(j) = lngTemp
Next 'lngN

'Stick it on the worksheet (in a single row).
Cells(Rows.Count, 1).End(xlUp).Offset(1, 0).Resize(1, lngSetSize).Value = lngArray()
End Function
'------------------



"rcc"

wrote in message
Thanks, Mike. I'm not sure if that would work in my case since I don't
actually have a data set to create a cumulative distribution from in the
first place. Really the only information I have is the upper and lower
bounds and the mean. It's from those values that I want to create a
distribution that looks similar to a normal distribution.

One idea that I had is to sort of combine two normal curves together. I
would have a piecewise function that described a normal distribution. On one
side of the mean, the function would be described by one standard deviation
and on the other side of the mean, another standard deviation. Is that too
much a hack to be theoretically sound?

"Mike Middleton" wrote:

rcc -

Instead of using a distribution defined by a mathematical function, you
could use a piecewise-linear approximation of the cumulative distribution.
Since you can use as many points as you want to describe the cumulative
distribution, you can get a very close fit.

RiskSim, my Monte Carlo simulation add-in for Excel, has a RandCumulative
function. Other simulation add-ins have similar features.

- Mike Middleton
http://www.DecisionToolworks.com
Decision Analysis Add-ins for Excel


"rcc" wrote in message
...
My apologies if this is more an "academic" question than an Excel
question...

I'm looking to create a probability distribution in Excel. My probability
distribution is similar to a normal distribution, but it's scewed. My
data
ranges (for example) from 200 to 800 and has a mean of 400. If my mean
were
500, then this would look like a normal curve with 2*stdev of approx 150.
But since the mean is 400, the curve is scewed to the left a bit.

Does anyone have an idea as to what kind of probability distribution I
would
use for this? My end result is using the distribution for a Monte Carlo
simulation (which I have all set up in a worksheet).

Thank you.