Sunday, August 4, 2013


     In my last post, I wrote 'History is about lies'. I also wrote that lies arise from two sources: 1) ignorance, which is inexcusable, and 2) bias, which is tolerable.
     Why do I find ignorance inexcusable and bias tolerable?
     Bias is tolerable because a biased account can be weighed against another account with a different bias. The alternative is no account. Every writer brings a bias to his subject. If a writer states his bias up front, he is being honest with the reader, and the reader can trust or distrust the account accordingly. A writer who hides his bias and presents his account as true is dishonest and is lying to the reader. He cannot be trusted.
     Ignorance is inexcusable because it renders the writer's argument not just weaker but false in all. Last time I wrote about a history professor who stated that the firing rate of the Brown Bess musket was one shot every three or four minutes. He was mistaken. Where he got his information, I do not know. I do know that he was wrong. He made the assumption that he was correct and never bothered to validate it. Why did he not validate his assumption? Arrogance.
     He was wrong about muskets. What else was he wrong about? Falsus in uno, falsus in omnibus. (False in one, false in all.)
     Besides a degree in history, I have a degree in mathematics. Better than that, I have experience in applied statistics. By that I mean I have searched for the right data, collected the data, collated the data, performed analyses, and reported the results. Did my work do any measurable good? Yeah. It demonstrably improved the communications performance of the USAF Satellite Control Network (reduced network connections times from 6m50s to 4m10s at the 95th percentile) and saved $50 million a year in operating costs.
     (An aside: I spent months learning how to perform parametric analyses on normally distributed data. Away from academia and in the sogenannte 'real' world, I found most -- that is, almost all -- distributions to be right-skewed. Parametrics are less than worthless on right-skewed populations. They will yield answers that are meaningless, but they will yield them with such accuracy and precision as to imbue the reader with confidence in the answers.
     (I found that I spent 85% of my time collecting and collating data, 5% of my time analyzing data, and 10% of my time reporting the results of my work to decision-makers who didn't know statistics and didn't want to know statistics. They wanted me to find problems and recommend fixes. I found I got 5% of my payoff from collecting and collating data, 10% from analyzing data, and 85% from presenting the results of my work to decision-makers who didn't know statistics and didn't want to know statistics. I got to be real good at presentation.)
     I know whereof I speak. I also know how little I know. But I also know how little others know, and I know that they do not know how little they know. Did you get that?
     Any spreadsheet software will give you access to 90% of the most powerful parametric operations I know. In 5 minutes, you can do the statistics I spent months to learn and years to master.
     The difference between us is that I know when those statistics apply and when they don't. (Assuming you are not a statistician.)
     I know the right stats tell an intuitive truth and the wrong stats tell a convincing lie. Without my experience, I would not know the difference. Neither will you.

     In the world of today, everyone has the same religion, and that religion is called NUMBERS. People believe in numbers with blind faith.
     In the world of today, I am a heretic. I do not believe numbers. Not until I know where they come from and what they purport to say. And not always then.
     30% Thirty percent. Ebooks account for 30% of the book market.
     I have read this many times on different blogs. Dean Wesley Smith, whom I read and respect, has repeated it many times.
     I think 30% is a lie.
     What do you see when you look at that statistic? A number? Do you have faith in numbers?
     I see a bastard child, an illegitimate statistic. First thing I want to know is who's your father? Who's your mother? Are you from uptown, downtown, or the boondocks?
     Invariably, when I chase the statistic I find it comes from the American Booksellers Association. The ABA is that statistic's father. Where do the ABA get their data? From traditional publishers. They are the mother.
     Therefore, it is more accurate to say that ebooks account for 30% of the sales for traditional publishers.
     In contrast, in October 2011 Jeff Bezos stood before a graph that showed that Amazon sold as many eBooks as DTBs. That is, 50% of their sales BY UNIT were ebooks. DTBs tend to be priced higher than ebooks, so no correlation to revenues can be made.
     Now I have new questions. Is the 30% that the ABA reported by unit? That is, do ebooks account for three of ten books sold? Or do ebooks account for 30% of traditional publishers' gross revenues?  Or 30% of profits? Given traditional publishing's peculiar accounting practices, how much faith can I have in the accuracy of their data? That is, do their data reflect reality?
     I don't know. I doubt they know.

     What is the truth?
     The truth is that in 2011 Amazon was selling as many ebooks as DTBs. You can make your own projections from there to here and to the future. I have.
     Here is another truth. Over the last two years I bought 11 DTBs. During that same period I bought more than 300 ebooks.

     YMMV, but if it does I pity you.


  1. Aren't there also deliberate lies, or would you call that extreme bias?

    If I want to hide something I may lie, invent the statistic and say it is real.

  2. "Aren't there also deliberate lies, or would you call that extreme bias?

    If I want to hide something I may lie, invent the statistic and say it is real."

    Eddy B

    Thank you for your question. I had to think about it.

    In English, the word 'lie' implies deception. In Spanish, mentira has a meaning closer to 'not true'. There is no connotation of deliberate deception. To deceive is enganar.

    Adolf Hitler's Mein Kampf was an untruth. That he believed it to be true does not change that. It was not a deception. He did not say one thing while he believed another.

    I think what you ask is are there deceptions vice untruths.

    Yes. I have witnessed these.

    Deceptions are insulting. The deceiver says, "I am smarter than you. I can tell you black is white and make you believe it."

    I cannot recall an historian ever telling me something he did not believe to be so. At worst he was guilty of untruth.

    I have witnessed politicians deceive. President Obama, Secretary Clinton, Congressman Boehner, Senator Feinstein, and Senator McCain have done it repeatedly. I cannot tell if Senator Reid is trying to deceive or if he is just that stupid. In contrast, the worst that can be said about Sarah Palin is that she spoke from ignorance. To my knowledge, she has never spoken to deceive.

    I have witnessed journalist deceive in service to a political agenda. They have lost all credibility with me. I no longer pay them any attention. Some will argue that there are honest journalists to whom I should attend. My rejoinder is that you are known by the company you keep. That's why I no longer practice law.

    Eddy, my first reaction when I read or hear a statistic is "That's a lie. Prove to me it isn't." I recommend that be your reaction, too.