iSixSigma

Floating Control Limit for SPC

Six Sigma – iSixSigma Forums Old Forums General Floating Control Limit for SPC

Viewing 90 posts - 1 through 90 (of 90 total)
  • Author
    Posts
  • #35510

    Kacha
    Participant

    I would like to have a document for Floating Control LimitΒ for SPC. Floating Control Limit is a recalculation the control limit when new data is input. But I don’t have any document to support this theorem. Please kindly help me.

    0
    #100080

    learning
    Participant

    Kacha,
    Β 
    Are you currently doing SPC on your process? If your process is within theΒ control Β limits, do you still compute control limits for the next card ( chart )?Β  I use minitab and everytime I run a Xbar and R chart, it always compute for another control limits.
    Β 
    learning

    0
    #100090

    Mikel
    Member

    Both of you need to read a credible SPC book – Montgomery or Wheeler.
    You never recompute unless there is clear data indicating a change.

    0
    #100142

    Kacha Owner
    Participant

    Hi stan!
    I use QIAnalysis software. There is a option to recalculation everytime when newly data have been input. SO I would like to know what is a benefit and what kind of process which is compatiable to use this option.

    0
    #100145

    Mikel
    Member

    There is no benefit. It is the wrong thing to do.

    0
    #100147

    Scott
    Member

    Stan is absolutely correct.Β  The whole idea of a control chart is that it determines if a stable process is out of control.Β  By changing the control limits for each set of new data, you clearly defeat the purpose of control charts.Β 

    0
    #100712

    Kacha
    Participant

    Hi Rick,
    Please explain me more. When can we re-calculate the control limit?
    I understand for this point, I already changed my control limit to be fix.
    However, I still would like to have a information for floating limit. Shall you help me in this.
    Β 

    0
    #100723

    Savage
    Participant

    You can (and should) re-calculate control limits when the process has changed and after you have enough data from the “new” process.
    If you changed the process today, you should note that the process has changed, that the old control limits are from the prior process, (i.e. they are not used for new data since the process changed) and once you have enough data from the new process, compute new control limits.

    0
    #100726

    Deshazer
    Member

    Agree with Stan and Rick. Once determined (provided you use s reasonable amount of data to calculate them) control limits should be kept UNLESS there has been a change in the process.Β 
    For instance, if you are controlling the net weight onaΒ  filler and you changeΒ the filling machine and the new machine has a tighter tolerance, then it is necessary to reevaluate control limits.
    Another example would be if you are monitoring the cycle time for some process, for instance approval time for credit applications, an as a result from a lean analysis you smplify the process amd expect significantly shorter times, it would be wise to revaluate the control limits again.
    I did once an interesting application of controlling pH on a fermentation process.Β  As the process advanced you expected pH to decrease, but always the same rate batch after batch. What we did was to determine, based on many batches the control limits at given times (each 15 minutes, 8 hores the total batch time).Β  Strictly speaking what we did was to determine control limits for 32 series of data, then we plot all the individual control limits on a single chart with time on the X-axis and pH on the Y-axis. If you can picture it it looks like the control limits for each batch were a decreasing line, but always the same limits for every batch.
    Hope this helps!

    0
    #100727

    R Dog
    Participant

    Hi,
    The guidelines for recalaculating control limits we use where I work are from Wheeler and are as follows:
    Recalculate iff:

    Data is displaying a distinctly different pattern than in the past
    and
    The reason for the change in behaviour is known.
    and
    The new process behaviour is desirable
    and
    It is intended and expected that the new behaviour will continue.
    Β good luck,
    R

    0
    #100732

    Arora
    Participant

    Hi!
    Revaluating the control limits is not wrong when new values come in.
    The control limits are calculated using the following formulas.UCL = (D4)RBAR. LCL = (D3)RBAR, RBAR = 1/K(SUMMATION R(i)) where D4,D3 are constants taken from table depending on n(subgroup size),
    Basically RBAR is the average of all the range values in the subgroup sets.Let’s say this value is 2.Now, you are adding another 10 subgroups(new set) to it all with the Range of 2.if you calculate RBAR now it will still be the same(since all values you add was same as the mean).If RBAR is same, the UCL and LCL also will be the same.So, it’s not wrong to re-calculate them. But on the other hand, if your new values deviate from mean, then the control limits tend to change and you need to calculate them.
    In either case, it’s not wrong to calculate control limits when new vlues come in.
    Any views please!

    0
    #100733

    Brian.P.M.
    Participant

    Hi Stan,
    Maybe this will help them…
    http://www.qualityadvisor.com/library/control_charts/recalculate.htm
    Regards,
    Brian

    0
    #100735

    Mikel
    Member

    My view? Wrong, you don’t know what you are talking about.

    0
    #100737

    Arora
    Participant

    Well, it says what I already agree.
    1.There is not hard and fast rule.2.I do agree that, it is not necessary to reiterate to find the control limits when the values added are same as mean.But, that doesnt prove doing it as “wrong”.
    If there is a mathetical explanation explaining why finding the control limits for the data when the new values are added is wrong, I would be glad to see it your way.

    0
    #100738

    howe
    Participant

    Stan,
    I have noticed that you make many comments like this one in this forum, actually more like stupid comments than anything else. If you know the answer (or IF you think you know the answer), just answer it. DonΒ’t come up with stupid comments like this. SoΒ… you have DemingΒ’s picture or his teaching points on your desk, apparently it did not do you any good.
    Β 
    Stop acting like you know it all. And regarding that 1.5 Sigma shift, you see there are many posters who donΒ’t believe in it but you donΒ’t see us make so much noise about it. ThatΒ’s because they are a lot more important things than that.
    Grow up. They are many smarter posters here than you. Can not help thinking of a Chinese proverb that says an empty pot makes a lot of noise.

    0
    #100739

    Ralph
    Participant

    “They are many smarter posters here than you.”Β Β 
    Guess what – you are not one of them.Β  “POP” as your bubble bursts Mikey.

    0
    #100740

    Darth
    Participant

    Mike,
    Stan answered the original posting question clearly and succinctly on 5/12/04 at 7:25a by saying “You never recompute unless there is clear data indicating a change”.Β  On 5/13/04 at 6:45a he said again, “There is no benefit.Β  It is the wrong thing to do.”Β  Stan’s position was supported by posts from Brian, Rick, Matt, Roy and R-Dog and I agree with all of them that it is not proper convention to recalculate the control charts without cause.Β  I sense that Stan got a little frustrated by some posters who continued to dispute the validity of the offered advice.Β  If you have an opinion on the topic, pleaseΒ express your thoughts rather than a commentary on Stan.Β  It would be most useful to get your expert opinion on whether it is proper to recalculate control limits on a running basis rather than a process shift.Β  I look forward to your insights.Β  Thanks

    0
    #100743

    Gabriel
    Participant

    Kris,
    In arecent post we were saying that a process is stable when its behaviour is consistent over the time, and that a control chart is the way to see if the process is stable or not. And control limits are a very important part of the control chart.
    Control limits help you see if the process is stable or not BECAUSE THEY ARE FIXED, not floating, and hence you can compare the process behaviour over the time against a fixed refernece. Just as the lines on the road help the driver keep the car in the road on a dark and foggy night.
    Recalculating the contrlo limits as the processΒ developes over timeΒ would be like moving the lines of the road together with the car. sonner or later, both the car and the lines will fall in the clif.

    0
    #100754

    Mikel
    Member

    Mike, I feel sooo put in my place.
    Just a couple of questions for you –
    1) What do youΒ  actually know about Deming?
    2) What do you actually know about SPC?
    One of the things that are really bad about all of the so called Six Sigma experts is that they have never actually taken the time to learn anything beyond what was in their BB or GB training.Β  Find me a SPC book that says to recalculate the limits just because you have more data and I’ll back off – but you will not find it.
    Go learn something before giving advice.

    0
    #100763

    howe
    Participant

    Stan,
    It’s foolish to think that a few months of training will make someone an expert in quality,Β 6 SΒ or statistics. Sure, there are many of them not just in this forum but everywhere you go. But the problem is you think that you are sooo much better than everyone else in this forum. You claim that you are guilty of being sarcastic but the way you insult some of these folksΒ goes beyond that. If you have nothing good to say or can not criticize it in a professional way, just shut up. The ones who know a lot, are the ones who are very humble.Β  Yes,Β I am talking aboutΒ folks likeΒ DR. DEMING, and DR. TAGUCHI!Β 

    Β 

    0
    #100770

    Mikel
    Member

    Dr. Deming humble when pointing out ignorance?
    You don’t know what you are talking about.

    0
    #100772

    Brian.P.M.
    Participant

    I don’t believe that Stan ever proposed that he is “sooo much better than everyone else in this forum”, but from what I’ve read in this forum, I feel that he (as well as posters such asΒ Darth, Gabriel & Statman) is quite a bit smarter than the majority of posters and I for one have learned quite a lot from his (their)Β replies to questions. I can also understand his frustration when someone tries to argue the wrong side of a no-brainer issue, i.e. whether or not to recalculate control limits whenever new data are input.
    Β 
    Brian.P.M.

    0
    #100774

    GAC
    Participant

    I’m with you on this Stan.Β  The whole point, from a statistical mechanics POV, is that the control limits are the predictable range of the process behavior.Β  Moving the limits with each point, with no limit on the datarange or justification for recalculation, dilutes the value of having a KNOWN range of behavior.
    Any assignable change in the behavior needs to be explained before limits are reset.Β  I always teach that are three criteria for adjusting CL’s.
    1.Β  The change is explainable and well understood.
    2. The change is desirable.Β  (So, you’d never expand the limits for a process showing excess variation.)
    3.Β The newly understood and more desirable process is expected to behave in this fashion for the foreseeable future.
    The fact that some software packages allow this option is immaterial to the statistical validity of doing so.Β  It is true that sometimes this is ok, as long as one is still building toward the magic k = 30.Β  It creates a set of limits along the way at increasing power and certainty levels as the dataset builds.Β  But, once k = 30, there is no reason to recalculate unless you meet the above criteria.
    You’ve got an ally here, Stan.
    GAC

    0
    #100777

    howe
    Participant

    There you go again, Stan. YouΒ always KNOW what you are talking about. If you noticed among all the quality gurus, I mentioned only two names. That’s because I am more familiar with these two gentlemen’s work. Unlike you, I don’t throw names around like I know everything and everyone.

    0
    #100778

    Mikel
    Member

    Mikey,
    You need a time out – go take a nap

    0
    #100784

    howe
    Participant

    Stanley,
    You need it too. Yours is overdue. Meanwhile stop eating those mushrooms!
    Β 

    0
    #100787

    Ralph
    Participant

    Mikey,
    Β 
    You are way out of your league here.Β Β Β  Consider posting on iknownothingaboutsixsigma.com.Β Β  You should quietly fold your tent and move to a smaller, less complex, campground.Β  Β And I say this just because I am beginning to like your feisty little spirit.
    Β Ralph

    0
    #100788

    SSBB
    Member

    Ralphy boy,
    Your nose is getting way too brown! Hope Stan enjoyes it!

    Β 

    0
    #100790

    Hmnnn…
    Participant
    #100791

    Jakelicky
    Participant

    I think we are missing something here. There seems to be an assumption that Kacha wants to improve a process, or know when to take action and when not to take action.
    All Kacha wants to do is re-calculate his limits with every data point. Minitab will do that for him. He’ll be happy, he wasn’t going to do anything with the data anyway.
    The reason to collect data is to take action. He just wants to “float” his limits.
    Β 
    Β 

    0
    #100792

    Brian.P.M.
    Participant

    Nice finds
    Brian

    0
    #100793

    Hmnn…
    Participant
    #100796

    Ralph
    Participant

    Norton,
    Now THAT was truly sophisticated and biting sarcasm.Β  Mikey could learn a few things from you.Β  Nothing like a good butt joke in a technical forum to keep a thread on track.Β Β 
    Ralphy

    0
    #100797

    SSBB
    Member

    Appreciate your kind words, Darth. I mean Ralphy.

    0
    #100799

    Mikel
    Member

    Actually, I like the whiners like Mikey better. The message from them is excellent, be really nice to people who know nothing and don’t know good advice when given. It is much more important than actually knowing something.

    0
    #100800

    Ralphy
    Participant

    You both erringly assumed that I wrote in support of Stan.Β  No.Β  Stan is also misguided and limited in his postings.Β  He just knows more than Mikey.Β Β  When you choose between the wisdom of a wannabe mechanic and that of a Mister Goodwrench in a forum that needs the leadership and wisdom of a Lee Iacocca, you go with the mechanic with the clipboard.Β Β Where is Dr. Harry when you need him?
    Ralphy

    0
    #100801

    howe
    Participant

    Stan, you just don’t get it, do you? The problem is that you think you know it all. You think you know so much and you’re God’s gift to this forum. I have news for you. You do NOT know as much as you think you do. It’s like you went through six weeks of training instead of 4 weeks. It’s a long journey, and you just started. Be humble, and less of a jerk. Or is it too late to teach an old dog new tricks?

    0
    #100802

    SSBB
    Member

    Hey Ralphy boy,
    I got you all wrong! You are definitely stupider than I thought. I did not know how stupid you were until you said, “Β….wisdom of a Lee Iacocca”. His wisdom has a lasting effect just look at Chrysler oops Daimler-Chrysler. Idiot! Those 30 years of working in quality profession did not pay off, did it?

    0
    #100804

    John Noguera
    Participant

    Folks, whether you like Stan or not is irrelevent. He is correct.
    I have seen people use software that automatically recomputes control limits when new data is added (such as Minitab 14 – unless you enter historical mean and stdev –Β I think that I have persuaded the folks at Minitab to change the defaults on this).Β 
    Here is the simple problem you run into:Β  if a slow long term driftΒ occurs in the process – an assignable cause –Β the recalculation of the limits keeps adjusting, following along with the slow drift, making the detection very difficult.
    The company doing this automatic recalculation didΒ notΒ see that they were out of control.Β  By the time the problem was recognized, the required adjustment was severe and caused significant customer grief.
    Β 

    0
    #100805

    Ralphy
    Participant

    I knew Lee Iacocca, and you sir, are no Lee Iacocca.Β Β 
    Β 
    Lee, as you might (or might not) imagine, had difficulty running the corporation when he no longer ran the corporation.Β Β Β  You should take a few minutes (or in your case a few hours) and chronicle the events and successes of the corporation under Lee Β– you might change your perspective.Β Β 
    Β 
    Let me also suggest that your get a new mantra, the one you use now is obviously not restive enough.Β  Sit back, cross your legs, extend your arms, circle your thumb and first finger, and think your mantra, Β“IΒ’m not really too bright and should not be allowed to post in a technical forum, ooooohhhhhhmmmmmmmmmmmmmm.”  
    Β 
    Now, feel better SSBB?Β Β Β  You may now post again.Β Β 
    Β Ralphy

    0
    #100807

    Darth
    Participant

    Hey, don’t be dragging my name in on this one.Β  I am enjoying this one from the sidelines.Β  I posted one response in which I asked Mike his opinion on the subject.Β  He did not respond but continued to fling insults all around.Β  The answer to the original question has been asked and answered….by many posters.Β  I am just going to sit this one out so don’t be using my name or assume I am using another identity.Β  I post all responses in my own name and don’t need to hide behind someone else’s identity.

    0
    #100810

    SSBB
    Member

    Hey Meat Head:
    Every time you open your mouth, it is more apparent that you are clueless. You wrote, “Lee, as you might (or might not) imagine, had difficulty running the corporation when he no longer ran the corporation.” Didn’t LEE pass the sword to Bob Eaton? Go read what happened after that, you idiot.

    0
    #100811

    Ralphy
    Participant

    You remain not only very hostile but very confused.Β  Does not your statement, “Didn’t LEE pass the sword to Bob Eaton? Go read what happened after that” indicate to you that Lee was not running the corporation, post sword pass?Β Β  Is Jack still responsile for GE?Β  You are a confused and sullen little man.
    Rember your mantra and then the, “Ooohhhhmmmmmm”.Β  Β That and the meds kicking in should eventually help you.

    0
    #100812

    Mikel
    Member

    From one jerk to another – at least I am not confused about it.

    0
    #100813

    MBB2
    Participant

    SSBB – get off the forum.Β  You don’t contribute, you harass, and poorly at that.

    0
    #100814

    Brian.P.M.
    Participant

    Bingo
    Brian

    0
    #100823


    Is that right Darth … “you had me soo…oo fooled. “Just my two cents!” :-)
    Andy U

    0
    #100836

    Cordell Fife
    Participant

    Andy U,
    Yes, Darth did say, “I post all responses in my own name and don’t need to hide behind someone else’s identity”. The difference between Darth and some other posters including me is that we doΒ post underΒ different identityΒ sometimes BUT we do NOT lie about it.
    Β 

    0
    #100840

    Ralphy A.K.A. not Darth
    Participant

    Kim,
    Sorry to have started this senseless tertiary debate through my postings that were mistakenly interpretedΒ by SSBB (a seemingly illiterate clod) to have been written by Darth.Β Β  I am not Darth, but don’t mind the comparison due to his normally well thought out andΒ rational postings.Β Β Β  You can call me Dr. Ralphy but you can’t call me Dr. Darth – maybe that’s where the commonality came in with the illiterate clod.Β  I, like you, have used other posting “names”, but they have never been “Darth.”Β 
    Dr. Ralphy
    Β 

    0
    #100844

    Cordell Fife
    Participant

    Dr. Ralphy, Darth or whoever you are,
    Have someone else read back to you all your posts. Hopefully that will help. Because, frankly speaking I am not sure if you even realize what an idiot you are. You make no sense.

    0
    #100848

    Mikel
    Member

    Anyone out there considering being kinder and gentler?
    This thread has really deteriorated since I told Kris my view on his ignorance (he did ask for people’s views).
    Why don’t we just go back to dealing with incorrect notions with respect to Six Sigma?Β There are enough of them to keep all of us busy.

    0
    #100849

    Wolfgang K
    Member

    Come on Stan, after all of your lambasting and malevolent responses to other posters you are calling for peace? If you want to focus discussion on SS and it’s strenghts and weaknesses fine, but lose the attitude.

    0
    #100850

    Mikel
    Member

    I object to you calling my responses malevolent. They are not and never have been intended to harm anyone.

    0
    #100852

    howe
    Participant

    Wolfgang,
    Thank you! You said to Stan, “If you want to focus discussion on SS and its strengths and weaknesses fine, but lose the attitude.” I couldn’t have said it better myself. Thank you again.
    That’s what I’ve been trying to tell Stan all along. If he wants to take the time to answer a post, great! But he does NOT need to insult a poster if theΒ poster does not understand it or does not agree with Stan. He has been acting like a spoiled child.Β Any time he doesn’t like a poster’s comments, he starts the insult.
    Β 

    0
    #100853

    Stan,
    What is the status of the debate? I think you’ll find Gabriel’s simulation spreadsheet of particular interest …
    Cheers,
    Andy

    0
    #100854

    Mikel
    Member

    Hey Mikey – read any of your own posts?

    0
    #100855

    howe
    Participant

    Stanley,
    Can’t We All Just Get Along? Who initiated all this? You did!

    0
    #100856

    Mikel
    Member

    Really? Where can I get a copy – I probably missed the thread while I was busy playing with Mikey.
    Gabriel – may I have a copy?
    Status of the debate is simple. Mikel’s masterpiece is nothing but a verbose piece of fluff. I have shown it to many with education that meets or exceeds the Rev. Harry and they all laugh. They are also willing to engage in a serious discussion about it, but not a Ringling Bros. three ring circus – Mikel has been accused of such things. So the question is if these so called judges who are in business with Mikel will actually call Mikel’s book what it really is. I have made some inquiries to Doug and Bert to see if they are serious.

    0
    #100857

    Mikel
    Member

    No I did not. I answered the original poster twice. And then this joker Kris showed up. Read both of his posts and how arrogant and wrong they are; and read that he asked for others “views”. I gave him one.
    Then Mikey flew in to save the world with insults and telling me to be professional while telling me to shut up??????????? You don’t see the irony in your words?

    0
    #100858

    John R
    Participant

    [A forum moderator has determined this post to be overly rude and/or use offensive language. This type of behavior will not be tolerated and the author has been removed from posting on this forum.]

    0
    #100860

    Mikel
    Member

    Okay genius, tell us what you know about Six Sigma

    0
    #100861

    Gabriel
    Participant

    Yes, of course. It is nothing more than a monte-carlo simulation of 20 groups of 10 subgroups of size 3 for which Xbar and R are charted and the different estimations of the standard deviation are compared. It has a feature to add “variation due to special causes” at a definned rate and defined average intensity to see the effect on the estimations of the short term and long termΒ variations. I will help support the fact that virtually no 1.5 sigma shift will pass undetected if SPC is put in place. I will like very much that you have a look at it and give me your opinion. Please give an e-mail where I can send it.
    In fact I thought I would send it to you, but then I thought that it won’t help you if you are going to discussΒ whether with ng=30, alpha=0.005, and a normal distribution you can get an estimation of sigma that will leave out of theΒ “real” 3 sigma limit (at one side only)Β more individuals that what it seems because of the inflated sigma, and that this added number of “beyond-3-sigma-limit” individuals is the same added number that you would have with the right sigma but the average shifted by 1.5 sigmas. That is mathematically correct only when all this assumtios hold true and only for the 3 sigmas point. What is the practical efect of this cocnlusion I can’t imagine, but this point seems to be out of the scope of the debate which isΒ whether 1+1 equals 2 or not, regardless of whether 1+1 and 2 mean something or not. If these are the rules of the debate, the you and and Statman are in trouble.
    Please, post an e-mail where I can send you the spreadsheet.

    0
    #100862

    BBMax2
    Participant

    John R,
    That’s a little harsh, don’t you think?Β Β Β Β Stan’s post read like he wanted to stop being thought of as being abrupt.Β Β Β  Sounds like he’s trying to be introspective and work on his communication style.
    BBMax2

    0
    #100863

    Gabriel
    Participant

    You are kidding, right? Stan loves his sharp style. The key is just to know his style and live with it. His knowdlege is wide and you can learn a lot from him (as I did and still do) if you focus on the content of his posts and not on the etiquette (or lack of it).

    0
    #100864

    Mikel
    Member

    Gabriel,
    Thanks. [email protected]
    It is useful and for what its worth, the real debate is going on now as more and more people take the time to understand the junk that has been published by this joker. He has one real win – a book that was co-authored by Richard Schroeder. If you read the other stuff he has written and compare it to the Schroeder book, you have to realize that the real work was done by Schroeder. They have their silly rules and perhaps biased judges, but I am betting that Doug Montgomery will not tie his name to Harry’s “proof” if enough people are watching.
    You have got to see he is on the ropes – his ASU venture is at least the third thing he has tried since SSA. I am betting he is in trouble and we will not be seeing many more interviews from the “ranch”.
    Just my humble observations of the obvious.

    0
    #100865

    BBMax
    Participant

    I read his note that he has turned a personal corner – had an epiphany.Β Β Β Β  He sounded sincere to me.Β Β Β  IΒ’m looking forward to a kinder and gentler forum with the same high level of contribution from everyone.Β Β 

    0
    #100866

    Mikel
    Member

    Gabriel – Shhhh! You are giving away my secrets.
    The forum has been awfully quiet and particularly mean spirited. Just trying to stir the pot.
    I do believe the calling of names and telling people to shut up needs to stop.
    Where is Darth (his real name you know) to draw some attention away from me anyway?

    0
    #100867

    Mikel
    Member

    Thank you for the kind words, I will try to live up to your expectations.
    I do believe that a sarcastic nature is genetically coded and it will take work – maybe there is a 12 step program. On a serious note, I do believe in the serenity prayer and think we would all do well to remember the personal version of it.

    0
    #100868

    Karel
    Participant

    Just out of interest would a Cusum help with this type of question / problem?

    0
    #100869

    Interested
    Participant

    It would sure seem that Stan talks a lot. Since I have
    been following this shift thing on iSixSigma, Stan has yet
    to present any type of statistical discussion. Why? it is
    very likely he is unable to discuss such things at the
    analytical level. Funny that we don’t hear Dr. Harry on
    this site blasting away at Stan’s personality. I am in
    agreement with another poster that stated the proverb:
    empty pots make a lot of noise.

    0
    #100870

    Mikel
    Member

    Have you read the book?
    No?
    Go do it and come on back and I will discuss any aspect on any level with you.

    0
    #100871

    howe
    Participant

    Gabriel,
    Sounds very intersting! Is it in Excel? Did youΒ write a Monte CarloΒ code to generate ths sample? Is it possible to post itΒ so all the interested folks can download it?Β  Thanks again.Β 

    0
    #100872

    howe
    Participant

    Stan,
    Once again, you don’t get it, do you? No doubt in my mind that you understand some parts of Six Sigma BUT NOT as much as you think you do! What got all this started was NOT your first reply to Kris but your last one where you insulted him. In every thread, you insult or curse someone at least once. So…. When someone tells you to shut up, it hurts him feeling. Why do you feel that you always have to say something in every thread? It goes back to my first point. You want to show everyone how you know everything.

    0
    #100873

    Interested
    Participant

    Would Stan agree that if a random normal performance
    variable was specified as NID(0,1), then the sum of Z^2
    would have a chi-square distribution with n-1 degrees of
    freedom?

    0
    #100875

    John J. Flaig
    Participant

    Gabriel,
    Excellent analogy. I think it helps people to understand when you can provide this type of example.
    Now, let me say that there are times and methods where re-computing after each new point is correct. If you are using Quesenberry Control Charts, then the limits are recomputed with each new point and they are computed beginning with n = 2. If the system is stable, these limits asymptotically approach the Shewhart Control Limits as n goes to infinity (usuallyΒ when n > 30).
    Regards,
    John
    Β 

    0
    #100876

    Savage
    Participant

    I agree.Β  I like Stan.Β  And he’s even called me an idiot and told me I don’t know anything.Β  The point is – who cares?Β  Stan has experience and insight that is very useful.Β  I continue to learn from him and pay attention to what he has to say – not how he says it.Β 

    0
    #100877

    Interested
    Participant

    Would Stan agree that the uncorrected sample size for
    making a two group comparison can be computed as N1=
    N2=[2(Za/2+Zb)^2]*[S^2/D^2], assuming the case
    NID(M,S)?

    0
    #100878

    Mikel
    Member

    Mikey,
    Once again you don’t know what you are talking about.
    Go find me one instance where I have cursed anyone or told someone to shut up and I will drop out of not just this thread, but the whole forum.
    You have a personal problem. I say a lot of stuff on here, often to provoke, but not anything close to your allegations.

    0
    #100879

    Mikel
    Member

    Hi Reigle,
    I’ve been wondering where you were.
    I thought you were against these assumed names.

    0
    #100880

    Mikel
    Member

    Reigle,
    You really should try English. This techno speak makes you sound more pompous than you really are.

    0
    #100881

    Interested
    Participant

    Stan might be interested to note that I do have the book. I
    have studied it quite closely and find it most interesting.
    Which of the core equations do you take the most
    exception to? You did say you would discuss such things
    on any level. If you would like to address me as Reigle, I
    don’t have any problem with that, but I would appreciate
    some direct answers to my questions.

    0
    #100882

    Tanya
    Member

    This is the biggest waste of everyone’s time that I’ve ever read. Talk about two children calling eachother names…won’t one of them grow up finally and put this to bed. There’s going to be a debate. We don’t care, so go do your debate and report back to isixsigma if you want. If you don’t, I won’t lose any sleep. But I would like to get back to real people with real problems and real help. That’s what I love about this site. You guys really put a damper on it.
    Forum moderators, please consider closing this thread.

    0
    #100883

    Darth
    Participant

    Gabriel, I am glad you clarified some of my thinking in this post.Β  I have recently read Harry’s book and my conclusion was that, based on his very limiting assumptions he is probably right in his 1.46 almost 1.5 shift.Β  The reality is, I can’t think of many processes in our organization where n = 30 and alpha = .005.Β  It is more like alpha=.05 and sample sizes are in the hundreds.Β  I am sure your simulationa would show different results for the more realistic conditions we have.Β  Since Harry is so adament in continuously pointing out the limits of his assumptions and the fact that they may not represent the reality of what most of us face, I wonder why we are having the debate.Β  It appears as you said, that the argument is about 1 plus 1 equaling 2.Β  There is no argument there.Β  Stan wants to debate the practicality of the 1.5 shift and the attemptΒ of some practioners to make it a universal truth.Β  I give Harry credit for laying out his strict assumptions, repeatedly mentioning them and leaving the misapplication of his 1.5 shift to others.

    0
    #100887

    howe
    Participant

    I have been insulted by idiots on this forum and called Mikey, when my forum name is Mike. When I point out the obvious abusive behavior by various posters I am insulted again and my Six Sigma knowledge is ridiculed by know it all posters who think they are forum kings.Β  You want a great debate?Β  I’ll give you a great Six Sigma debate.Β  Name the time and place.

    0
    #100894

    Stan,
    The simulation will show that for n = 3 the shift is no greater than about 0.5 sigma. Since most people use 20 to 30 subgroups anyway; the shift is irrelevant ..
    IΒ haveΒ observed aΒ larger shift value whenΒ comparing the process performance based on a n = 30 against n= 200, instead of against the theoretical distribution N(0,1)
    One often enounters this situation when studying ‘gaps’ such as an ‘interference,’ or differences between distributions. If I were you I would ignore the maths and concentrate on the simulation – isn’t this what Reigle recommended anywyay?
    Cheers,
    Andy

    0
    #100900

    chngagent
    Participant

    Tanya,
    If you don’t want to read these threads no one is making you. Why do people complain so much when they are the ones doing the clicking of the mouse. It just like TV change the station!

    0
    #100903

    anon
    Participant

    If you rotate this thread 90 degrees counter clockwise, it resembles a normal distribution. It would be a travesty to close this thread!!!

    0
    #100904

    SSNewby
    Member

    At last, something in this thread that I agree with.Β Β  Good observation and well expressed.Β Β  Thanks.

    0
    #100918

    mman
    Participant

    Hi Stan
    Β  I really enjoy yourΒ  “sarcastic” way,like yourΒ  “short sentences”,please continueΒ like this,it is a way to make this forum alive and attractive,kind regards.

    0
    #100919

    SSNewby
    Member

    mman,
    I dunno about all of that.Β Β Β  Alive and informative – certainly.Β Β  Sarcastic – yep.Β  Attractive? – that’s a bit of a stretch.
    SSNewby

    0
Viewing 90 posts - 1 through 90 (of 90 total)

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