Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

Simulating Strategies Using Historical EOD Data - Part 2 - Keltner



  • Using the Hi-Lo Keltner strategy (for more details about the strategy please see Part 1 of this report)
  • Trading with NUGT, DUST, FAS, FAZ, UPRO, SPXU, URTY, and SRTY
  • Using 3 different periods: m01 (1 month), m06 (6 months), and y01 (1 year)
  • Assuming that the test starts from a specific week, set the start date of the testing period to 1st, 8th, 15th, and 23rd of every month from 2008 through 2015
  • Using Keltner lengths of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 35, 40, and 50
  • Using Keltner weights of 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, and 1.25 (note, higher weights have been tested and they impact the results negatively)
  • Trying with stop levels: 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 25%, and 100% (100% = no stop)
  • Trying with max wait days per trade: 7, 8, 9, 10, 11, 12, 13, 14, 15, and 365


The simulation resulted 35 million different cases. The aim of this effort was to average out the probabilities, reduce the impact of the noise on the result data, and to find the best representative results of the average market.

The symbols have been carefully picked from both Long and Short x3 ETFs. They cover both the bearish market years and the bullish market years. In addition they are from a wide range of market sectors: Gold, Oil, S&P indices, and the Financial Sector indices (please let me know if you want to add anything to the list and I will be more than happy to prepare the reports for you).


The Golden Numbers:

Considering the above inputs, the top 10 performers were listed as (drum roll........)

Just to give you an idea of the performance of this strategy, the following Excel Pivot chart depicts the distribution frequency and summary of the Average Annual Return (NYSE:AAR) for a sample scenario:

Note: the top left corner columns show the values used for the settings of the strategy.

As noticed, although there is a considerable percentage of losing trades but sum of these trades is significantly lower than the sum of successful trades. In addition, although the number of trades with high return percentages is fewer but they cover a considerable portion of the total gain.


Download the report

You could download the report from here. But if you are using it as a reference, please don't forget to cite and please don't forget to post a comment under this instablog with the reference URL of that page.

Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.