Apache OpenOffice Community Forum

Hello all,

I am new here and have often visited for workarounds, but not actually posted.

So, my query...

I am currently wanting to shorten my workstream in relation to averaging. I currently use a 3 pronged method of AVERAGE, GEOMEAN and HARMEAN functions, to calculate an absolute "AVERAGE/MEAN"; iterating until I get them to equal TRUE against each other. However this can often be quite space/time consuming, so I need to work out a workaround.

Is there a way this can be done in such a way that would end up with a custom function like ABSOLUTEMEAN(A2:L2)?

Coding is not my forte, and I know this is asking a lot, but it needs to be cyclic(nested) until the resultant cells all equal TRUE.

Thanks in advance

Coding is not my forte, and I know this is asking a lot, but it needs to be cyclic(nested) until the resultant cells all equal TRUE.

You can do some iterations
- in a macro code, or
- in the cells by usage the circular references.

You must switch on the iteration feature for the "iteration in cells" in the Menu :
Tools - Options - Open/LibreOffice Calc - Calculate - Iterative references ... Check in the checkbox, and set the maximum number of the iteration Steps, and the desired Minimum change value.

nicalex wrote:I currently use a 3 pronged method of AVERAGE, GEOMEAN and HARMEAN functions, to calculate an absolute "AVERAGE/MEAN"; iterating until I get them to equal TRUE against each other. However this can often be quite space/time consuming, so I need to work out a workaround.

Can you help me to get some sense out of this? If you want advice how to get a user function calculating something, you should specify that something clearly. If you already have a -probably inefficient- method to get what you want by Calc formulae you can attach a sample sheet demonstrating it. A clear specification independent of any kind of known implementation is better, however, or should at least be tried in addition.

Check Help and the Calc Guide for Goal Seek and Solver. Both can be found under Tools.

I had a simple idea, perhaps correctly understood ...

Hi, and welcome to the forum.

nicalex wrote:I currently use a 3 pronged method of AVERAGE, GEOMEAN and HARMEAN functions, to calculate an absolute "AVERAGE/MEAN"; iterating until I get them to equal TRUE against each other.

I cannot find much literature about the properties of the Arithmetic-Geometric-Harmonic mean (AGHM). That suggests to me that it is not statistically or mathematically useful. If this is your own idea, please consider using standard statistical measures instead. Unless you have a good reason not to, I suggest that you use the average of the data. This has good statistical properties and has been studied for hundreds of years.

For two numbers, the AGHM will be GEOMEAN(A2;B2). For all positive numbers, I believe we will find:

   Harmonic mean ≤ Geometric-Harmonic mean ≤ Arithmetic-Harmonic mean ≤ AGHM
   ≤ Geometric mean ≤ Arithmetic-Geometric mean ≤ Arrithmetic mean

The rate of convergence for the AGHM is quadratic, so one probably needs just a few, say a dozen, iterations to get convergence. Just perform a fixed number, instead of using a loop, if you insist on using the AGHM.

If this solved your problem please go to your first post use the Edit button and add [Solved] to the start of the title. You can select the green checkmark icon at the same time.

Can you please define more accurately what you are trying to do.

For example, let us have something like:

I have three values, a, b and c.

I calculate the Average as (a+b+c)/3.

I calculate the Geometric Mean as cube_root(a.b.c)

I calculate the Harmonic Mean as reciprocal((1/a +1/b + 1/c)/3).

I now to adjust ??? such that ...

Thank you for all the responses folks, it is very much appreciated! But a bigger thanks goes to lader, you did understand correctly