Monday, June 15, 2009

Open Thread: Do Fibonacci Retracements Work?

Following my recent Dueling Fibonaccis for the SPX, a reader asked if I knew of any 'scientific' studies that attempted to evaluate the usefulness of Fibonacci retracement levels. The short answer is that I am unaware of any detailed studies, nor was I able to dig up anything meaningful with a search of SSRN.

So I turn this topic over to readers:
  • Do Fibonacci numbers work as a trading tool?

  • Are you aware of any published analysis which evealuates Fibonacci retracement levels?

13 comments:

  1. yes, it works every well, like a broken clock. eventually it's right where the clock stopped.

    Jimmy

    ReplyDelete
  2. http://www.cass.city.ac.uk/media/stories/resources/Magic_Numbers_in_the_Dow.pdf

    http://works.bepress.com/kuldeep_kumar/1/

    ReplyDelete
  3. Thanks for the links, Jared. In addition to the two articles you cite, someone also pointed me toward an old Credit Suisse piece, Does Technical Analysis Work?

    Cheers,

    -Bill

    ReplyDelete
  4. Bill,

    Thank you for an excellent blog that is well regarded among peers.

    Tim Knight at the Slope spoke highly of your expert knowledge of VIX. Tim seems to be proficient in Fibs. Perhaps, you may want to touch base with him? Another suggested name would be Corey Rosenbloom who is a CMT at Afraid To Trade.

    Among Fib users, the authors seem to change the period lengths of their Fibs to support their observations and theses.

    There are also excellent blogs like Smart Money Tracker that rely more on the depth of their knowledge than on Fibs.

    Among the healthy debate that I've read on Oil and Gold for example, I found the opinions drawn from Fib retracemement levels were less convincing than the ones based on logical and factual analyses.

    Thanks again for your sharing your views with us.

    My two cents.

    ReplyDelete
  5. Hello everybody! If you look for math papers dealing with technical analysis being a reliable trading tool, there are many of them. In general, a (restricted) access to academic journals (like e.g. Elsevier) is necessary for downloading.

    Concerning Fibonacci retracements, i'm not aware of any conclusive results. You mainly have 2 schools of modelling of markets: deterministic (=looking for trends) and probabilistic (=random walk down wall street, martingales and co.). If you choose to work within the probabilistic framework and especially within Gaussian markets, Fibs may find their place in terms of standard deviations ensuring the existence of some confidence intervals. This has to be verified rigorously, but at least this sounds quite reasonable to me. For deterministic modelling, only a statistical study of trends/fluctuations can say something about Fibs. I doubt someone already tried that because the academics are quite consensual about probabilistic modelling.
    --
    Laurent

    ReplyDelete
  6. Yes, Fibonacci retracements do work - but you need to think outside of the box.

    Watch this, then take a look at what happened afterwards on the FTSE 100 chart last week. I trade like this regularly on both Indices and forex:

    http://www.youtube.com/watch?v=2AiV9Ka1P30

    ReplyDelete
  7. Here's another one of my fibonacci experiments, though I hope this one doesn't come true:

    http://i17.photobucket.com/albums/b57/stevevickyl/Charts/DOWBIGBEAR2.jpg

    ReplyDelete
  8. Yes they do, but not in the way that most people use them.

    ReplyDelete
  9. Thanks for helping answer my question Bill. I appreciate it, and your blog in general.

    ReplyDelete
  10. They work sometimes and sometimes they do not. i guess it really depends on the context and if there are other key support and resistance points near the retrace levels... where these retracement levels really sit are the real argument... I guess the only one that matters is the one biggest guy is watching.

    ReplyDelete
  11. They definitely work as a tool to highlight areas of support and resistance. I think this is an objective truth. After I draw fibs on a chart, I can very quickly and clearly see where support and resistance are at. Even if you are not a true believer in "fib support and resistance," you are really losing out if you're not using fib retracements to assist in locating support/resistance.

    ReplyDelete
  12. Absent of any rigorous study, fibs may work better now than in 2002 due to increased popularity.

    ReplyDelete
  13. Just my 2 cents on the matter:

    Any technical analysis tool works and doesn't. If anything worked all the time, the person who discovered/invented the tool would never tell another soul. If such a tool ever did get out, it would stop working as there would be massive gaming of the tool itself, insuring it's failure.

    People who stridently look for proof are often holy grail seekers who seek certainties in life where none exist.

    The first step in utilizing technical tools to trade or invest, is to understand that they are not infallible and be able to deal with their failure.

    As for fibs, sure they work -- as much as other types of support/resistance tools. Also, they work better on certain markets, and in certain market conditions than others. If you can't deal with that, then don't bother trying to use fibs.

    -Peruser.

    ReplyDelete