Saturday, August 14, 2010

The "Hindenburg Omen" and intellectual property

In discussing Time Magazine's "50 Worst Inventions," which included what Time called hydrogen blimps, IPBiz noted that Time itself was making a few technical errors. For example, zeppelins, such as the Hindenburg, were not blimps and helium is not "less volatile" than hydrogen.

Now, we have the "Hindenburg omen," which illustrates a few other IP concepts. The Hindenburg omen is about the stock market. Other than conveying an image of disaster, the Hindenburg omen has nothing to do with the Hindenburg, and, in fact the namers of the omen planned to use Titanic instead, but that name was already used.

The disaster predicted by the omen has been back-derived. For the omen to be triggered one must have over 2.2% of stocks on the NYSE hitting new 52 week highs and 2.2% hitting new 52 week lows. Wikipedia notes:

The 2.2% number seems to be tied to (or gives the appearance of being) the average growth within the US economy since 1955, an average of only about 2%. The original creators of this signal have not fully explained their use of 2.2% constant, but apparently chose it because that was the number that their back-testing of market data showed was statistically meaningful.


Because of the very specific and seemingly random nature of the Hindenburg Omen criteria, it is possible that this phenomenon is simply a case of overfitting. That is, if one backtests through a large data set and tries enough different variables, eventually correlations are bound to be found that don't really have any predictive significance.

Of "correlation not implying causation," a point of relevance to a variety of law review articles in the IP area, note a past comment in a blawg:

I've served on a hiring committee at a law school for several years, and I can add the following:

(1) Phds can be a good proxy for someone who has had actual academic training, and knows how to make an academic, as opposed to a lawyerly, argument.

(2) On the other hand, many Ph.D. programs, especially in the humanities, seem to instill extremely bad habits with regard to research and writing, especially by putting a premium on showing one's familiarity with the literature and making sure to cite and discuss every possible author who has ever commented on the subject, as opposed to emphasizing developing a coherent, theoretically sound thesis. Ph.D. theses, if anything, tend to be even more poorly written than law review articles. Ph.D. historians and, to a lesser extent, philosophers and political scientists, can graduate while being entirely ignorant of, e.g., even basic economics, which does not aid the coherence of their work.

(3) With regard to law and economics, most regressions I've seen are laughably bad, and don't come anywhere near proving, or even suggesting, the results that the author claims to glean from the paper. This problem arises at least as often from PhD economists as from non-PhDs, which suggests it's a problem endemic to economics (and will someone please teach economists that a "statistically significant" result does not PROVE anything--correlation does not prove causation--, even putting aside obvious problems of data dredging and whatnot.

Of the correlation in the omen, the Wall Street Journal notes: The Omen was present at every market crash since 1987, but has also occurred many other times without an ensuing significant downturn. Market analysts said only about 25% of Omen appearances have led to stock-market declines that can be considered crashes.

The WSJ also noted: [James] Miekka [the inventor of the omen], who writes a Wall Street newsletter called Sudbury Bull & Bear Report out of his homes in Maine and Florida, wasn't even aware that his own Hindenburg Omen indicator was activated. The 50-year-old former physics teacher, who is an avid target shooter, said he was "taken by surprise" after he plugged all the data into his model.

And to go to patent issues, recall that it was a correlation that was being patented back in the Metabolite case.
For example,
Michael Crichton assails Metabolite case but ignores the academic connection

***Relevant links

Time's "The 50 Worst Inventions"

Wall Street Wonders If 'Hindenburg Omen' Just A Lot Of Hot Air

Hindenburg omen in stock market suggests upcoming slump
Hindenburg omen and other technical indicators seem named just to see how much fear they can provoke.

**In passing, in Poshard's Ph.D. thesis (generally in the social sciences area), the problem was that he didn't cite sources for copied text, and his thesis committee didn't even recognize the unattributed copied text.


Post a Comment

<< Home