TS Support and I both took a crack at this before the article was published. Unfortunately, neither of us was successful in producing results that matched the charts from the article. There are some basic assumptions for the study that seem to be missing. Specifically, the length of a period for the logarithmic returns and the number of periods were not included in the article. I was able to create a formula by plugging in some numbers (1 and 50) for these two assumptions. You can edit these through the “Edit Studies” option. However, I was unsuccessful in producing an indicator that resembles the chart included in the article on the Light Crude Apr 03 contract (CL J3). There is either an error in my formula, which is eluding me, or the instructions for creating this indicator are inconclusive. I found many resources regarding fractal dimensions and estimating the Hurst Exponent, but nothing that detailed the specific instructions for creating Long's indicator. I've attached the formula anyway so you can test it out and see what I'm talking about. If anyone has any helpful information, please share it with us.
Has anybody from E-Signal taken another stab at this following the recent article in S&C recently in which this indicator was heavily used. Most other data vendors provided the indicator in that article once again, sadly with the exception of E-Signal. This is a bit of a shame since I believe this is a widely used and successful indicatoe.
Please see this post from March 2007. A copy of the study can also be found in our EFS Knowledgebase in the EFS Library\Stocks and Commodities folder, 2007 Mar: Trading Systems and Fractals.
Jason K.
Project Manager eSignal - an Interactive Data company
Please see this post from March 2007. A copy of the study can also be found in our EFS Knowledgebase in the EFS Library\Stocks and Commodities folder, 2007 Mar: Trading Systems and Fractals.
Thanks Jason, I did a search of FDI and Fractal, and it didn't appear. Didn't think of looking in the folder though
FWIW, if you would have used fdi* AND fract* as a search term, it captures the forum post Jason referenced.
I have found my search results are typically better if I avoid the last letter or two of the search terms and use a wildcard instead (if possible). This ensures a hyphen or other grammatical difference at the end of one of them prevents a match. Of course the AND is essential as well.
Comment