Dangerous Extrapolation

Much of what we read in investing includes reaching back in history to a few instances of something and then suggesting that "today is like that."  We have often pointed out the danger of such reasoning.

As a test, we heard a bit of trivia which seems to be correct.  There is a fundamental "setup" that has occurred six times over 70 years.  This "setup" has had a 100% success rate.  It covers all of the potential instances, and therefore does not seem to be data mining.

There are good psychological and motivational reasons to believe in this relationship.

Are you ready to make the trade?  Do you want to know?

The Trade

The last six coaches of the University of Michigan football team have beaten Ohio State in their first season, going back to Fritz Crisler and Bennie Oosterbaan.

Tomorrow will be the first Michigan/Ohio State game in the Rich Rodriguez era.  Ohio State is favored by about three touchdowns.

Would you take Michigan?

If not, remember this example the next time someone shows you a stock chart of the current market compared to some other country or time period.

You may also like


  • Bill Luby November 21, 2008  

    Nicely put.
    Perhaps I can still take the points though…

  • Mike C November 21, 2008  

    “If not, remember this example the next time someone shows you a stock chart of the current market compared to some other country or time period.”
    I get it. The past (with some set of specific data points) may not be a reliable predictor of the future that happens to share those same data points.
    Of course, the big problem then becomes with something as utterly complex as the stock market and economy where does that leave one in terms of making actionable decisions? If past data is irrelevant, then does one defer to “intuition” and “gut feel” or some subjective judgement that is likely clouded by all sorts of distortions and biases? In that case, is one left with really nothing as a foundation for judgement other then hope?
    Maybe I’m slow, and not connecting the dots, but I’m guessing you had a more specific point here that you are alluding to. Maybe you can share that for those of us that can’t make all the connections?

  • Bill aka NO DooDahs! November 21, 2008  

    The bigger the odds in college football, the more you should take the favorite against the spread.
    The bookie’s job is balancing the books, not predicting the final score. Americans love underdogs; at some point the spread gets “too wide” for betting sentiment, regardless of the relative talent of the teams.

  • RB November 21, 2008  

    Great interview @ the Kirk Report, Prof (matter of fact, I’d also recommended earlier this year that he interview you). I’ll be getting a copy of Bryan Hicks’ book on your recommendation. BTW, Buffett in his Fox news interview now seems to support Paulson’s shift in plans as possibly “better mileage for money”.

  • Jeff November 21, 2008  

    Mike — It is, perhaps, a familiar point. Good inferential methods are hungry for data — lots of it. People are willing to draw conclusions on much less data than they should.
    If I put up a couple of charts with lines and then showed the current market, everyone would jump on board. It is a point that is difficult to make in a convincing way.
    Thanks for making a (typically) good point.

  • Jeff November 21, 2008  

    Thanks, RB.
    I’ll have to check out the WB quote. When TARP first came out, he did a good job of explaining the value of these assets to anyone who could hold until payoff and finance cheaply enough.
    I also don’t think he is one who believes that Congressional committees should be setting a bunch of requirements for the everyday conduct of operating businesses — since he does not do that in the businesses he owns.

  • Jeff November 21, 2008  

    Sports fans Bill and Bill – – Interesting theories. I am not optimistic. Each year my brother, on the faculty at THE Ohio State University calls me if they win, and I get to make the gloating call if Mich wins.
    I have been receiving far too many calls in recent years!