# Is 2 Sigma 69.20% or 95%? I Am Confused!

Six Sigma – iSixSigma Forums Old Forums General Is 2 Sigma 69.20% or 95%? I Am Confused!

Viewing 21 posts - 1 through 21 (of 21 total)
• Author
Posts
• #52482

Ang
Participant

I remember that 3 sigma is 99.7, 2 sigma is 95, etc.
Now I have notice in training material I have that 2 sigma is 69.2, 3 sigma is 93.32%
Why the difference? If you know, please provide rational. Thanks!!

0
#184636

Participant

Terminology / Operational Definitions
Sigma Conversion Table:  2 “Sigma (short-term)”  converts to a 69.2% Long Term Yeild.  3 Z-st gives 93.3%

0
#184638

Sudeep Bhargava
Member

There are two aspects to the conversion of the percent defect free product to its corresponding sigma level- one is the mathematical aspect and the other is the operational aspect.
The mathematical aspect may be understood through a function in MS Excel which is normsinv(probability of defect free product). When using this function, you have to enter the probability of making a defect free product as the parameter for this function. The result from this function gives the mathematical way of converting the probabluty into the sigma level. Using this function, you would find that the sigma level corresponding to 99.9997% defect free product is 4.5 sigma. Entering 69.2% or 0.692 is the parameter, we would find that the corresponding sigma level is 0.5
If we wish to use the concept of 1.5 sigma shift, which takes into account the inherent deterioration of the systems and processes involved in generating the product or service, we would have to add 1.5 to the above function which I have told. Adding 1.5 sigma to the result obtained, we would find that 99.9997% or 0.999997 corresponds to 6 sigma (which is the widely accepted defintion of six sigma). The sigma level corresponding to 0.692 would then be 2 sigma.
In case you have more doubts, please feel free to get in touch with me directly on my mail ID [email protected] or call me 91-9818550488
Regards

0
#184661

Mikel
Member

statement as “inherent deterioration” to justify the 1.5 which has no

0
#184664

Sudeep Bhargava
Member

Every real system in this world deteriorates with time. There is nocontrol over this deterioration, It is true with the human body as well as the other physical systems including the business processes. 1.5 sigma shift takes into account that “deterioration” only.
I am sure Mr Stan would like to enlighten all the users here with his knowledge of this “1.5 sigma shift”.
https://www.isixsigma.com/ library/ content/ c010701a.asp
Sudeep Bhargava

0
#184665

Mikel
Member

Sure, a poorly controlled process can deteriorate by 1.5 or more.Well controlled processes deteriorate far less. Use real data on your process and you will know for yourself. If your
deterioration is less than .5, you are doing pretty good. If it is greater
than that, find yourself a good consultant.

0
#184668

Marvin
Participant

I don’t understand why six sigma folks bother about short term sigma at all. Doesn’t the customer actually experience the long term and not the short term? Why bother to compensate for shifts? Just report it as it is.

0
#184669

Darth
Participant

Stan, I am afraid that Mr. Sudeep is correct. ALL processes deteriorate at a constant rate of 1.5 sigmas. Dr. Harry proved that with his derivation. That is why we use it as a constant in the 6 sigma literature. Of course the fact that Dr. Harry himself says it is a correction factor to be used only under very specific circumstances and not a universal constant as many hacks seem to try and sell, Mr. Sudeep is right in challenging your knowledge of the shift as well as your overall knowledge of Six Sigma. Possibly instead of providing a dead link like to some bogus article, he might research the Forum for the many, many discussions on the subject. Ah, but then he might actually get some new insight into the futility of the shift and discover that you are really a knowledgeable and smart guy.

0
#184672

Mikel
Member

Shhhhhh!It’s more fun to set these fools up.

0
#184677

Sushil
Member

Hi all, It seems we are leaning more towards our not so tasty language than participating in a healty debate.
Thanks

0
#184679

Mikel
Member

Sushil,I am willing to participate in healthy debate.What is your take on the subject and why?

0
#184687

Marvin
Participant

actually i have a challenge for all you “in the know” folks here. Explain how shift is really calculated. Give a mathematical example. I would ask an MBB this if i want to hire him or her.

0
#184694

Mikel
Member

Would you know the correct answer if you heard it. And how do
you know?This is what I would ask any interviewer dumb enough to think
this is a good question. And yes, I know the answer.

0
#184695

Taylor
Participant

Marvin
If you asked me to calculate the shift in a process, I would walk out your door, telling the others on the way out what a freaking moron you are. 1.5 sigma shift is meaningless. It is nothing more than Smoke and Mirrors. Report the process as it is and move on. There is nothing to be gained from calculating process shift. Focus on controlling the process and making improvements that can be Predictive and not reactive.

Geez, why does this subject keep coming up?

0
#184701

MBBinWI
Participant

You are suffering under the worst and most confusing topic in the six sigma world.  Some tables show sigma levels with the 1.5 sigma shift baked in, others show raw sigma levels.  Is 1.5 the “right” correction between short and long term, and for that matter, what is “short,” and what is “long?”
If you are properly monitoring your process, you should be monitoring the critical factors which cause your important results to change, and as they change from the acceptable level of variation (due to deterioration from wear or other uncontrolled forces), they are replaced or otherwise brought back within the acceptable range before your critical output varies beyond acceptance.  Thus, you will not observe a 1.5 sigma shift.
You may find a mathematical derivation as to a 1.5 sigma shift (search for Davis Bothe), but most fall back on experiential data.
Monitor your inputs and your outputs will not suffer from the “shift.”

0
#184702

MBBinWI
Participant

A little harsh, don’t you think?  Entropy exists, and not all variation can be controlled (at least not economically).  Thus, there is inherent deterioration.  Should be much less than 1.5 sigma, particularly if monitoring the sources of preventable variation, but exists nonetheless.

0
#184703

MBBinWI
Participant

If you think that you know “the answer” then I wouldn’t want to work for you.

0
#184704

bbusa
Participant

SushilThis forum doesn’t have much to offer to learners these days . You can polish your abusive skills here
Are the self proclaimed SS Gurus listening ?
bbusa

0
#184707

Sushil
Member

Dear Stan,
Thanks,
At my place we go with the short term sigma without taking that 1.5 into account.
Reason: Most of the times due to nature of operations (automated measures and miniscule human intervention in meassurement where R&R is always less than 2 % or so) we do so.
But this +1.5 sigma takes care of lot many factors which we may not consider in short term and external parties like customer always experience the gap in our processes.
In  my opinion, we measure to improve and if we see that ST Sigma is fulfilling my purpose and I have enough scope to improve,  will go for ST other wise explore that unseen  factors by using LT Sigma. Totally depends uppon the process I am dealing with and it’s importance for me.
In my opinion it’s all about improvement with right measurement.
Thanks

0
#184741

Marvin
Participant

just as i have suspected, nobody here knows how to calculate process shifts. And i am not even saying it must be 1.5 (come on, why must it be 1.5? Every process is different). All i am saying is, some of you people like to talk about things you know nothing about. If you cant even calculate the shift, stop talking about it.

0
#184746

Darth
Participant

Stan knows but he would have to shoot you if he told you. There is no such thing as “a shift” therefore no calculation. Agreed that there is some inherent drift of a process over time. A simple control chart shows the amount of drift. In some cases it is not significant, in others it is. Doing a process capability with both short term and long term measures of variation indicate to what degree the shift is something to worry about. If the Cpk is essentially the same as the Ppk then who cares about the drift. If there is significant difference then you worry. To apply a fixed number is foolish. Even Harry called his shift a “fudge factor”. Great for Forum discussions but pretty useless for business purposes.

0
Viewing 21 posts - 1 through 21 (of 21 total)

The forum ‘General’ is closed to new topics and replies.