In article . com,
"Aaron" wrote:

Another way to look at it is, if you plot the data on a bar graph, the
groupings are easier to see. You will be able to form a trend between
contiguous numbers and make "groups" out of the numbers accordingly.

A bar graph may make it more obvious to you, but it doesn't address the
fundamental problem.

Both XL and VBA need specific, explicit criteria in order to calculate.
There's no fuzzy logic engine built-in, which is what you're asking for.

I do this in my head.

And that's the issue - what you do in your head is obvious to you, but
not obvious at all to a designer of an algorithm. One would need dozens
of examples, encompassing the entire range of possible values, along
with your judgment on how to group before one could design the kind of
fuzzy logic analogue that you're using in your head.

For example, based on the examples you've given it's entirely unclear to
me, even if plotted as a bar graph, what rules to apply to decide
whether this is one group or three:

0.001, 0.001, 0.002, 0.002, 0.001, 0.001

or whether this is two groups, or four, or six:

0.001, 20, 0.002, 525, 556.001, 524