596983_en-US

liliconcom

Effective Resolution: how probable error is measured ?

Hello,

a customer using "EMP - Effective resolution" need to justify / explain what the calculation behind "Probable Error" is.

I don't see how Probable Error is calculated.

I've attached an example.

 

Emmanuel

Re: Effective Resolution: how probable error is measured ?

Hi @Emmanuel_Romeu .  Per Don Wheeler's EMP III textbook, page 27, probable error is calculated as:

            PE = 0.675 * (average moving range) /  ( d2 * sqrt(nr))

where:

PE = probable error

Avg Moving Range = average of difference of consecutive points in the dataset

d2 is a Bias Correction Factor which can be found in tables online.  Depends on number of subgroups used for average moving range, and number of observations per subgroup.

nr = number of measurements.

 

Wheeler's book says that "Probable Error is the minimum value for the Median Error of a single measurement.  Thus, the Probable Error is the minimum value for the Median Error of a single measurement." (page 26)