Tuesday, November 1, 2016

"Fibonacci Flim Flam"

This is a repost from 2012:

You have to be careful with this stuff. If someone presents a complicated explanation for their technical analysis my B.S. detectors start flashing, it doesn't mater it it's Elliot Wave alternate wave counts or flipping between point-and-figure and candlestick charts to show the chart "predicted" a move. There is a fine line between genius and madness.
[don't I know it. -ed]

From Lock Haven University:

...Footnote on dubious investment schemes.

I remarked that stock traders and investment counselors these days often use Fibonacci ratios as a guide to guessing their predictions. There is even computer software for making market predictions that claims to use "Fibonacci methods". This is one of the methods used in "technical trading". Using numbers and charts to make predictions justifies the label "technical", though the results would be just as good had the patterns of tea-leaves been used. One only has to eavesdrop on the websites and forums these people frequent to discover that many of them still believe in the "magic of numbers". Whole books tout these methods, with testimonials to their success, and these do make money, for those who write the books. One fellow who uses Fibonacci ratios frankly admits that they may not be "magic" but they do make his presentation charts look more impressive to clients. Of course the efficacy of such methods has never been scientifically tested. And why should anyone waste the effort? 
One such fellow emailed me, complaining about my negative comments. I soon discovered that this fellow was a sucker for all sorts of pseudoscientic numerology. He even tried to tell me how valuable was the Martingale system, popular in 18th century France and still used by some gamblers. It's simple. Each time you win you make the same size bet the next time. When you lose, you double the size of your bet the next time. Of course, any "system" can seem to work in the short run, once in a while. But in the long run (when played for a long time, or many times) it has no advantage, and while your chance of winning in the short run may seem to be improved, your chance of losing big increases the longer you play. Statisticians have analyzed such systems and concluded they are deceptions, but gamblers are often susceptible to such deceptions. And what is the stock market, but a gambling game with confounding variables, and with the players themselves affecting the odds?

Then this guy tried to tell me that Fibonacci numbers show up more often in winning lottery numbers. He could provide no data supporting that. Then he claimed Fibonacci numbers show up more often in the digits of phone numbers in the phone book. Well, duh? Of the digits 0 through 9, six are Fibonacci digits (0, 1, 2, 3, 5, 8) and four are not (4, 6, 7, 9), so Fibonacci digits should show up about 60% of the time. No great mystery there. The only example he could produce, from his own "extensive research", was a set of 200 phone numbers, 65% of the digits being Fibonacci digits. That's well within the limits of error for that small size sample. [3]

Some say that you can increase your success in the stock market by rolling dice or throwing darts to make your choices. Such investments will, in the very long run, averaged over many investors, do as well as if you used a broker, and you won't have to pay the broker's fee. I am sure there are brokers who shun mystical and magical formulas, but I remain unconvinced that even they earn their large fees. 

More Fibonacci Fakery.

This picture was emailed to me by someone who could not remember the source. It is a clever piece of deception. I suspect the picture was delibrately embellished to make fun of Fibonacci foolishness. If the person who created this would like credit, please get in touch.

The water path isn't a Fibonacci spiral, but someone has cleverly superimposed the golden rectangle on part of it that at first glance looks like it is. Look carefully. That large inner rectangle should be a square, but it is in fact wider than high. The rectangle in the upper right corner is almost a square. If you tried to cheat by reducing the horizontal magnification, you could make the large one square, but the smaller one would no longer be nearly square. I think that's clear forensic evidence that someone was deliberately cheating, probably to make fun of what I call "Fibonacci Foolishness". The "hand-drawn" appearance of the rectangles seems contrived to disguise the cheating. (They should have used computer drawn clean straight lines and perfect geometry.) Such cheating is often done with the pictures of Nautilus shells seen in books (see above). A physicist would conclude immediately that this can't be a golden spiral, nor any of the textbook spirals common in physics. Textbooks of important mathematical spirals show pictures and equations for the Spiral of Archimedes, the logarithmic or equiangular spiral, the hyperbolic or reciprocal spiral, the parabolic spiral, and my favorite, the involute of a circle. The reason is simple. Such spirals are radially unbiased. In this picture gravity provides a bias (and distortion) compared to a similar process happening in a horizontal plane. Also, the source of the water, her wet hair, isn't stationary. She produces this dramatic picture by flinging her head and body rapidly upward and backward.

This sort of cheating is what I object to. If one wants to be honest one might say this picture "suggests" a golden spiral. The "fact" of it being "something like" a golden spiral tells us nothing useful about it. Deeper inquiry might lead to design of a water hose rotating about a fixed vertical axis to experimentally investigate whether this process could produce something closer to the Fibonacci spiral. I have no reason to predict that outcome with certainty. It might just be an involute of a circle. I've challenged some mathematicians with this problem, but they don't take the bait. 
HT: Maoxian

Also at Dr. Simanek's LHUP webpage:

The Museum of Unworkable Devices. Perpetual motion machines
as physics puzzles.
Intelligent Design Creationism: Fraudulent Science, Bad Philosophy.
Myths and Mysteries of Science. Removing the mystery.
Cutting Edge Science. Profundity or parody?
Illustrated Lectures on these subjects.
Uncle Don's Notebook. | Bob Schadewald's Corner.
Physics and Astronomy. | Physics Abused.
History and Philosophy of Science.
Skepticism, pseudoscience, urban Legends.
Humor, satire, parody, mostly about science.
Hoaxes. | The Ed-Biz. | Quotes. | Illusions and 3-D photography.
Articles written for print publication.