Re: Rule of Thumb?

[Kromkowski@aol.com]

>>I would point out that the choice of three-sigma on the control chart is a 
rule of thumb by Dr. Shewhart and has proven very useful, and there has been 
much discussion on this list as to the relative merits of this rule.<<

Not sure that every one would agree with you here!  I don't think the control 
chart is a "rule of thumb" at all!  

Moreover, I previously provide a Monte Carlo experiment, that provide some 
evidence that 3 was better than 2.9 or 3.1 at reducing both kinds of errors.  
Shewhart's control chart theory is based, as I understand it, both on theory 
and empirical data which support it as an optimal economical solution/method 
for distinguishing between the two types of causes.

If it was a rule of thumb, then we, without basis for concern, could decide 
to use 2.5 or 3 or 3.1, depending on how we felt on any particular day.  At 
that point, why bother with the calculations.  The new "rule of thumb" just 
could be: react only to data points that are _really_ above average or 
_really_ below average.  Now that would be a rule of thumb!  We could call 
even write a book: "Eyeball Method Management in One Easy Lesson."  But 
control charts aren't that are they; moreover, they tell us also about the 
expectable variation, so that we can make predictions.
--
John David Kromkowski
Attorney at Law
6600 York Road - Suite 108
Baltimore, Maryland   21212-2028
Kromkowski@aol.comTelephone:  (410) 377-6248Facsimile:  (410) 372-0624
===================================================================