iSixSigma

Breyfogle vs. Wheeler

Six Sigma – iSixSigma Forums Old Forums General Breyfogle vs. Wheeler

Viewing 32 posts - 1 through 32 (of 32 total)
  • Author
    Posts
  • #52603

    Severino
    Participant

    Looks like Wheeler tore Forrest a new one in Quality Digest this month.  Mwuhahahahahaha.

    0
    #185174

    Mikel
    Member

    Good – he is selling a lot of theoretical nonsense.

    0
    #185175

    Severino
    Participant
    #185176

    bbusa
    Participant

    I have read Breyfogle’s handbook and it simply lacks the depth and rigor . I would any day go with Wheeler’s approachBBUSA

    0
    #185177

    Alter Ego
    Participant

    I prefer Montgomery myself …

    0
    #185182

    Severino
    Participant

    A friend of mine who teaches applied statistics and I were just discussing this.  Montgomery provides you with the “textbook” approach to both SPC and DoE.  Rigorous knowledge of his texts will provide you with a solid set of fundamental skills. 
    Wheeler’s texts provide assurance to the practicioner that even when you hang up most of statistical convention your analysis will still be relatively robust and that one need not be caught in analysis paralysis.  He also challenges his readers to question the validity of many methods (criticisms of AIAG MSA and Taguchi S/N ratios come to mind).
    Both authors maintain excellent references which should really be in anyone’s library.  Breyfogle on the other hand provides you with something you can wipe your backside with.  Waste your money at your own risk.

    0
    #185183

    bbusa
    Participant

    JSEV607Montgomery is top class ( SPC & DoE) . With due respect to Breyfogle , both Montgomery & Wheeler are streets aheadBBUSA

    0
    #185185

    Mikel
    Member

    Agreed.Montgomery, and Box before him are just solid material. Both
    loosened up their approach after head bashing with Taguchi
    methods for 10 or so years, so they can move toward practical.I think everything wheeler does is good with the exception of his
    advice on Measurement Systems. He is correct with respect to
    experimentation but flat dangerous if extrapolated to product
    acceptance.I have multiple Wheeler and Montgomery references on my shelves.

    0
    #185186

    Ken Feldman
    Participant

    As usual, the Forum discussion has disintegrated into a personal attack on someone rather than a debate of the issue.  So, let’s see if all of those with personal opinions on the man can discuss the basis of the article.
    As we all know, cycle time data is not normally distributed yet a common metric in business.  It is frequently tracked in a control chart, usually an I/MR chart.  As a result there are considerably more false signals of special cause.  Wheeler dismisses this and says we should see that but not to worry.  As the one who has to decide between real special cause and that to be expected by such a skewed distribution, what is the proper course of action?  Should I continue to chase common cause thus wasting time and possibly tampering?  Should I ignore it under the advice of Wheeler?  Should I transform it as suggested by Breyfogle?  What do all you experts do?

    0
    #185187

    Mikel
    Member

    Cycle time data follows a very predictable pattern in a controlled
    process. I would never transform it. You lose an appreciation of what is
    expected. Remember the visuals are for communication. Outliers are
    easy to see when you understand how a skewed distribution looks.
    Multi modes are also easy to see untransformed but less easy in
    transformed data.Modeling is also very easy using untransformed data.

    0
    #185188

    Mikel
    Member

    Forrest and his masked sidekick Rick are smart guys, but lack any
    practical experience. Read the drivel on Forrest’s blog or his
    advertising groups on Linkedin.

    0
    #185189

    Devans
    Participant

    Darth,
    Your post is right on. 
    I am not a statistician but I have read both Wheeler and Breyfogle; both know their stuff and both can teach you things.  The point that Breyfogle makes, and makes often, is that chasing common cause can lead to the wrong action or fire fighting as he puts it. 
    As a VP for an oil company, I don’t care about the theory.  I want to know what is the proper course of action 

    0
    #185190

    Mikel
    Member

    You want theory?Read Forrest’s control charts for management reporting nonsense.Tell us VP Devans – what of Forrest’s teaching do you use in the oil
    industry?

    0
    #185191

    Devans
    Participant

    Stan,
    I don’t have time to get into a pissing contest with you.  Why all the negativity and personal attracts?  
    There are several things that I learned from Breyfogle that I use every day in my job.  There are also several things I learned from Wheeler that I use every day.

    0
    #185192

    Arap
    Participant

    Wheeler appears to be talking out of both sides of his mouth here. He prefers to call control charts process behavior charts, because their primary purpose is to show if the process is performing in an unpredictable manner — special cause, or predictably — common cause. But yet he is dead against transforming data WHEN APPROPRIATE to ensure they do just that.Cycle time data by its nature follows a highly skewed distribution, which when plotted on an ImR chart typically generates many data points above the UCL, even though the process is performing predictably. Since the chart has a number of points that have crossed above the “action limit”, following Wheeler’s approach would lead to tampering and an inevitable increase in variation.If the Cycle Time data is predictable, transforming it typically leads to more than a 10X reduction in the number of points above the UCL, allowing signals of true special cause to be seen above the noise of the distribution’s natural shape. Like most other things, I believe the correct answer lies somewhere in the middle. Data shouldn’t be blindly transformed, but if there is a solid existing understanding of its natural distribution, why shouldn’t that knowledge be applied to help provide a more accurate picture of process behavior?

    0
    #185193

    Mikel
    Member

    I am not interested in a pissing contest either.Just other than standard fare stuff I could learn anywhere, I wondered
    what you found of value – especially at the VP leavel.

    0
    #185194

    Mikel
    Member

    I completely disagree. You take all of the meaning out of the data
    and completely lose the ability to get people to understand what to
    expect. You either have a distribution that is jammed against entitlement
    with a nice consistent degradation in the tail or you have easy
    problems to fix. Pick the points that violate that assumption and
    understand them.Cycle time data is better understood against queueing theory.
    Forget forcing it into a control chart. We have more than one tool –
    use the right ones.

    0
    #185195

    Arap
    Participant

    I totally agree with your statement: “You either have a distribution that is jammed against entitlement with a nice consistent degradation in the tail or you have easy problems to fix. Pick the points that violate that assumption and understand them”…and in that case a control chart based on transformed data would be an ideal way to make the distinction. I’m open to learning here, how else can you easily determine if the data points violate the assumption?

    0
    #185197

    Vallee
    Participant

    The biggest question in all this is what is the risk of uncertainty?
    You have two models:1) Fit the sample to the model
    2) Fit the model to the sample On one hand the forcing the sample can lead to a very non-
    predictive model for the future samples while forcing the model to
    the sample can lead to ambiguous models that are not robust. So do you want a prettier chart so that the model works or do want
    to be more conservative on the number of false hits, thus having
    more missed hits. What is the risk to company. person, or
    community on the missed hits? What is the risk for documenting
    too many false hits?

    0
    #185198

    Severino
    Participant

    Just for the record, I attacked his book(s) not him.  As far as the out of control signals are concerned, you investigate.  If your investigation determines there is an assignable cause you correct it, if it doesn’t then you document your investigation and move on.  The level of investigation you utilize and the time, energy and effort are completely at your discretion, but it really shouldn’t be viewed as a waste of time even if you fail to find an assignable cause.
    Again Wheeler’s point is that we are overly concerned with a false alarm when we really should be focused on safeguarding ourselves against failure to detect a process change.

    0
    #185199

    Mikel
    Member

    Just look, it’s obvious Even Darth could see it

    0
    #185201

    Ken Feldman
    Participant

    Stan, don’t drag me and my eyesight into this. If you are saying that it is not appropriate to control chart cycle time data, I am OK with that. But being able to distinguish which point outside the upper control limit is real and which is not is not obvious and you haven’t provided any indication or suggestion on how to make that distinction. I am also not comfortable with JSEv’s suggestion to investigate them all. Sounds like a lot of NVA time. Granted, over investigating will hopefully reveal a true special cause but at a cost? Is that cost less than missing a real signal? Guess it now becomes a matter of economics, not statistics hence the name of Shewhart’s book.

    0
    #185202

    Stevo
    Member

    Forrest,
     
    Don’t be discourage, most of the people here don’t get “it”.  Your genius is way beyond their mental capability.  Keep fighting the righteous fight.
     
    Stevo
     
    Ps.  I to am often misunderstood.

    0
    #185204

    Ken Feldman
    Participant

    Stevo, you aren’t misunderstood…you’re just an idiot!!!! Sorry, you can get even when we get together on South Beach in Feb. :-).

    0
    #185206

    Robert S
    Member

    Hey Forrest I have an idea. Why don’t you take your published books and the published books of those objecting to your positions and do a comparison matrix relative to the issues, then… oh, wait a minute. None of them have any published books. Sorry. Nevermind…

    0
    #185207

    Stevo
    Member

    Darth,
    Why can you only appreciate my wisdom after 6 drinks?
    Stevo

    0
    #185208

    Ken Feldman
    Participant

    You only stopped counting at 6. It took only three to appreciate Ms. Arm Candy.

    0
    #185209

    Ken Feldman
    Participant

    Certainly the fact that one publishes a book is a testament to their intelligence and credibility. Do you know who wrote the book with the following title?”Four and a Half Years of Struggle against Lies, Stupidity and Cowardice”Would you take his musings as gospel?

    0
    #185210

    Robert S
    Member

    That wasn’t  my point Darth. My point was it is easier, by at least an order of magnitude, to be a critic rather than be a producer.
    PS: I was in some ways saying what you did. Argue the position.

    0
    #185211

    Mikel
    Member

    I wonder how much of the materials you sell reflect Forrest’s
    prolific writings? What? You have never actually read anything of his? What? There is
    nothing in your materials that reflects your great respect for him?Robert, I piss Forrest off, but neither of us have respect for weak
    kneed arguments not grounded in logic.I think he is a smart guy who just lacks experience at the level he
    tries to appeal to. An easy problem to solve if there is a desire. I
    would love to see him go lead a deployment in a large company
    (say > $5B in revenue) using his theories. I am pretty sure we
    could sell tickets.

    0
    #185251

    GB
    Participant

    I too, would not transform it, but I have to admit, I do Ctrl chart CT data in addition to level-load/line balance analysis.

    0
    #185271

    MBBinWI
    Participant

    I’d like to show him a certain square cold cut line and see what he would do!

    0
Viewing 32 posts - 1 through 32 (of 32 total)

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