Quote from Spydertrader:
Five year period? One thousand Stocks? You performed a wonderful test - except you didn't test the system on any stocks people trade, and you tested it on a time frame by which you insured failure. Why can I say this? Only two equities have remained in the Universe of Stocks I trade for 2 years, therefore, testing the same group of stocks (beyond those two) over a time period greater than two years creates invalid results. In that same two years, the 'List' has changed members each and every month with some stocks falling off the list and others finding themselves added to the list. The current list contains 40 equities. In addition, many of the stocks I trade don't even have a five year history.
Not only did Dougcs perform a proper test, but somebody else also posted a proper back test to Journal One. Both individuals obtained positive results.
To perform a proper back test on this method (and it can be done) you need to test the stocks that conform to all the parameters used to qualify the stock, and then you perform that test for that stock within that time frame in which it qualifies. Testing beyond the time a stock finds itself within a Universe of Stocks that were, are or will be traded provides meaningless results. Nobody is going to trade an unqualified stock. You repeat the process for every stock you wish to test.
How can I say this? Easy. If a stock doesn't qualify to make the list of potential tradeable stocks - five years ago, five months ago or five days ago - who cares, it would never have been traded. Proving that that same stock which never would have made the list, and therefore, never had the possibility of finding itself in a trade would have lost money proves nothing with respect to how the 'system' works or it doesn't.
An analogy to the scenario you have tested: "Hmmm these people 'claim' a system works against the Nasdaq 100 stocks, but that can't be! Look here! I have a backtest against the Dow30 and it shows how the system doesn't work!" Please, people aren't that stupid.
How confident am I of this? I'll make you an offer. I'll gladly walk you through the entire process in english of how the system works, so you can code a proper backtest, select the proper stocks to test, and insure you use the correct time frames for each qualified stock. In turn, you provide me the parameters (and code) you used, so that I can replicate your test in my own testing software. We can even include the market's recent downturn in the test data. If the tests do not come out profitable, I'll stop posting on this web site in any forum other than in my current Journal. I'll even post the results as a 'disclaimer' in my current Journal. No excuses. You perform the tests using the correct parameters, and we let the chips fall where they may.
Of course if the tests do turn out positive (and I have a sneaking suspicion they will), then I'll expect you to post an open letter to Jack and everyone else apologizing for the ass you made of yourself.
It's a simple offer - requiring a simple answer - yes or no. I'll even pick up the tab for the phone conversation to relay the information.
- Spydertrader
A direct challege.
This will be met with silence of course.
I have concluded that many of these posts are driven by some sort of deep seated insecurities on the part of some these folks.
But we will see.