It's been over six years since I last tested the "P,V Boolean relation," so I decided to take a look at how it's done since the year 2000. I analyzed 200 stocks from 2000 to 2005, a total of 1000 stock-years. I scored each stock from 0 to 7 each trading day using the methodology in Spydertrader's latest Wealthlab script: http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=44010 I bought when a stock's score changed from 0 to 7 based on Jack's post at http://sputnick5.www8.50megs.com/ : I have scanned stocks into two groups: the 0 to 7 group and the 4 to 3 group. I will give you my math for this somehow. Lets focus on the 0 to 7 turn (Buy) following the natural investment thread. Buying at this point in the trading cycle is also called for in a document on the same site called: "Catch Up with Tomorrowâs Paper Today -- Technical Analysis Used in a Manner to Anticipate the Market" (adobe llamingos publishing). I've attached an excerpt that shows this buy point in the overall cycle. I ran five test iterations, exiting after 1,2,3,4,and 5 day holding periods. Next I'll post the equity curve from the BEST performing holding period, 5 days.
To say these results are less than stellar would be an understatement. And I do realize that SCT is NOT about simply buying the 0 to 7 shift and exiting 5 days later. But wouldn't one expect that this type of test would show at least some positive effect? Especially since the "P,V Boolean relation" is at the heart of SCT?
My favorite game in all of Eric Berne's "Games People Play" is "Let's you and him fight!" Go get 'em, 666.
I am pretty much sure I am on to something when it is met with dead silence, as opposed to the customery derisive hooting. A double top became evident in the early seconds of 11:07 PT in NQ today. Does the attached snapshot of the filtered tape suggest to you why?
I am guessing that it is indifference. Even your alias's may not give a shit at this point. Congratulations. Steve
I analyzed my Jack Universe of stocks using a very crude PV relationship and got very different results. Testing my universe resulted in about 1000 trades and buying a fixed dollar amount yielded these results: Wins 56% Avg Trade 1.0% gain. There are 35 stocks in my universe and over the 6 months backtest, only 2 lost money. Again, this is a very crude implementation of Jack's ideas and I did it some time ago when I thought he was a raving nut case but was curious if he was on to something. I've since changed my mind. DS
Oh really? Well here's the same analysis over the same 5 year period, but this time with 1000 stocks (I also limited the maximum # of positions allowed at any one time to 100). I hereby declare your puny attempt to lend credibility to the P, V Boolean relation to be laughable!
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