You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Jack Hershey PVT System Testing (moderated)
- Thread starter vikana
- Start date
- Status
- Not open for further replies.
Quote from Trader666:
Jack, for what it's worth, I regard YOUR comments as out of line. You have no right to rewrite history just because you're participating here.
Even if PVT is successfully tested in this thread, it does NOT negate my past backtests because I never claim to have tested PVT in its entirety which is what this is about. I only tested "rockets" and buying the "0 to 7 turn" -- and I tested them exactly as you said they should be tested at one time... [].
I'd like to ask both Trader666 and Jack to leave the past behind. Trader666 *has* produced solid tests of some of Jacks work; and should be recognized as such.
When Trader666 is ready, and maybe when this project has been completed, I'd like to study his work. Contributions, such as Trader666s, are welcome, imo.
Quote from vikana:
I'd like to ask both Trader666 and Jack to leave the past behind. Trader666 *has* produced solid tests of some of Jacks work; and should be recognized as such.
When Trader666 is ready, and maybe when this project has been completed, I'd like to study his work. Contributions, such as Trader666s, are welcome, imo.
You as moderators now have my respects and regards!
I agree and hope Jack does too. I'd also enjoy posting some of my work after we're done with this.
Quote from vikana:
I'd like to ask both Trader666 and Jack to leave the past behind. Trader666 *has* produced solid tests of some of Jacks work; and should be recognized as such.
When Trader666 is ready, and maybe when this project has been completed, I'd like to study his work. Contributions, such as Trader666s, are welcome, imo.
Quote from Spydertrader:
----------------------------------------------------------
Quote from doli:
OK. Putting on software engineer's cap: can anyone, from the provided description of PVT, come up with a set of requirements that would allow someone to implement the system?
----------------------------------------------------------
Start Here.
As it seems certain relevant software has been developed, Would it be possible that a set of back-testing performance statistics (preferable based on EOD data of say any single major ETF, covering a reasonable period of time) can be available for inspection (as the same statistics should be reproduced by anyone else with any of the software that are designed to do the same trading method)? Thanks in advance!
Quote from Trader666:
I agree and hope Jack does too. I'd also enjoy posting some of my work after we're done with this.
I regard that post as temporary humor.
When vikana gets his initial tasks done, he, you and everyone else will be able to see the humor in his mistaken assessment.
In the meantime, I have been asked, by PM from vikana to NOT add substantive content to the system of PVT.
We are not designing an ATS for PVT. Here, all that is going on is analysis, testing and clarification. These three things will be sufficient to deal with the specified stuff from page one which are:
1. To eradicate once and for all the prior posts on backtesting the PVT trading where the backtesting posted in your forum was not backtested by you to confirm its faults.
2. To show a complete backtesting code in whatever code is desired and to show that the code can be corrected to get a quality back test performed.
In effect, this is a back testing forum and it happens that PVT is being back tested to get the unfortunate propaganda created in the past by trader666 off the table once and for all.
I planned on posting the essential steps of using PVT as taken from the reference ""IBD MeetUp Number Five: Putting the Pieces Together for Building Wealth, A backgorund Paper for IBD MeetUp Tucson". This would have provided a basis for debugging the upcoming coded analysis, and testing being done over the next few days.
When this work is accomplished by those doing the tasks, I'm very sure when it is compared (from a coding and data useage viewpoint) that the past "work" of trader666 will be finally put to rest for just what it is even those the coding he did is still unpublished)
Quote from OddTrader:
As it seems certain relevant software has been developed, Would it be possible that a set of back-testing performance statistics (preferable based on EOD data of say any single major ETF, covering a reasonable period of time) can be available for inspection (as the same statistics should be reproduced by anyone else with any of the software that are designed to do the same trading method)? Thanks in advance!
A 400,000 run produced a result that there was no statisitcal significant difference in the third significant figure of the upper end of the cycle duration recommended.
Is 400,000 a good enough sample size???
BUT just as Trader666 made the same mistake you did, an ETF is not eligible since ETF's cannot make the cut to Get in the Universe. In Trader666's case he used the S&P 500 for his reasons (He has stated that I did not give him guidance properly; so here, and today, I am giving you guidance.) If Trader666 had used the correct Universe he would have gotten a result that may have had some value if he had not made other contributing mistakes. As we see the moderator feels Trader666 did a good job. He is mistaken.
Quote from OddTrader:
As it seems certain relevant software has been developed, Would it be possible that a set of back-testing performance statistics (preferable based on EOD data of say any single major ETF, covering a reasonable period of time) can be available for inspection (as the same statistics should be reproduced by anyone else with any of the software that are designed to do the same trading method)? Thanks in advance!
Any pointers to the sources of these statistics would be highly appreciated. TIA!
I've done some analysis of the final universe posted on the second page of this thread:
http://www.elitetrader.com/vb/showthread.php?s=&postid=1976450#post1976450
http://www.elitetrader.com/vb/showthread.php?s=&postid=1943173&highlight=final+universe#post1943173
DISCLAIMER: I'm NOT saying this is an analysis of PVT in its entirety, and I'm not saying this is how you (Jack) intended for the universe to be traded. HOWEVER, based on your comments about the importance of the universe and how high quality it is, I wanted to see how the stocks in the universe performed in June relative to the Dow, SP500 and Nasdaq indices.
This was my medhodology: Since the FU (final universe) was published in the Iterative Refinement thread at 06-03-08 01:20 AM, I bought each of the indices and stocks the following morning (June 3) at the open and exited after holding for 10 full market days at the close on June 16th. I also did a 20 day analysis, buying on the open June 3 and selling at the close on June 30.
First the 10 day results: The indices lost an average of 2.59% and the final universe lost an average of 3.1%
The attached chart shows the individual gains and losses... the first three bars on the left are the Dow, SP500 and Nasdaq, and the remainder are the FU stocks.
http://www.elitetrader.com/vb/showthread.php?s=&postid=1976450#post1976450
http://www.elitetrader.com/vb/showthread.php?s=&postid=1943173&highlight=final+universe#post1943173
DISCLAIMER: I'm NOT saying this is an analysis of PVT in its entirety, and I'm not saying this is how you (Jack) intended for the universe to be traded. HOWEVER, based on your comments about the importance of the universe and how high quality it is, I wanted to see how the stocks in the universe performed in June relative to the Dow, SP500 and Nasdaq indices.
This was my medhodology: Since the FU (final universe) was published in the Iterative Refinement thread at 06-03-08 01:20 AM, I bought each of the indices and stocks the following morning (June 3) at the open and exited after holding for 10 full market days at the close on June 16th. I also did a 20 day analysis, buying on the open June 3 and selling at the close on June 30.
First the 10 day results: The indices lost an average of 2.59% and the final universe lost an average of 3.1%
The attached chart shows the individual gains and losses... the first three bars on the left are the Dow, SP500 and Nasdaq, and the remainder are the FU stocks.
Quote from jack hershey:
The universe criteria is stringent for one reason. All of the stocks that make it into the universe have a high repeatability quality.
"high repeatability quality" means that they behavior in a very very similar manner.
If a comparison is made with any other universes, the distribution of the PVT type of universe will lie the the right of most all universes in how narrow the group performance is in terms of just how similarly (use a reasonable sample size where p is <0.05 by any stat measures) the PVT members perform by TA criteria (See measures in TA From A to Z by Achelis and partularly the measures used in charts, lists and indicators of PVT). The phrase "they all look alike to me" is appropriate and it is the goal of the Universe.
Use the Initial Analysis Spread Sheet to observe how sorting 15,000 stocks narrows the Universe down to approaximately 150 stocks total in the Universe at any time.
Testing any stock at any time that is not in the Universe with respect to any writings on PVT is possible and the results obtained simply state that the person did it for his reasons and they are not associated with PVT trading in any way.
Basically the Universe is obtained by using Fundamental Analysis. FA is a guarantee of quality regarding eps and EPS. It is also a gurantee of RS and "they all look alike to me".
Attachments
- Status
- Not open for further replies.