Quote from Trader666:
...In this document, Jack advocates buying the "0 to 7 shift" of the "P,V relation" (see last page). Nowhere in the document does Jack specify which stocks to do this on.
EXACTLY! So that was your first mistake in assuming what you chose would be sufficient. That's why he had many posts like this one
here. It talks about what those conditions are! You can see how they are none of the conditions that you tested.
Quote from Trader666:
QUESTION: Does buying the "0 to 7 shift" provide an edge?
[/B]
Mistake #2. The document speaks nothing of edges! There are ZERO occurances of "0 to 7 shift" being an edge.
Quote from Trader666:
Trader666 Approach -- Jack listed no qualifiers about which stocks to trade, so the implication is that it applies universally.
Mistake #3. See Mistake #1 which gives an example of where the author of the original document specifies conditions.
Quote from Trader666:
Therefore, start by testing on portfolios of randomly selected stocks from the set of all listed U.S. stocks. Result: spectacularly bad equity curves. Next, try testing on small, mid and large cap stocks (the stocks of the S&P 600, 400, and 500 indices) to see if performance varies by market cap. It does not and is spectacularly bad for all. Next try several technical filters, one at a time. Result: still more spectacularly bad equity curves. Conclusion: If buying the "0 to 7 shift" of the "P,V relation" does give an edge, it's not enough of an edge to show up in these tests and is not good enough for my trading.
"A Team" Approach -- Add a boatload of ADDITIONAL conditions that were not in Jack's original document and declare that the "P,V relation" works after all.
http://www.elitetrader.com/vb/showthread.php?s=&postid=593276&#post593276
Mistake #4. See Mistake #3.
Quote from Trader666:
Pop Quiz:
Which approach more closely tested what Jack originally wrote in "Catch Up with Tomorrow's Paper Today?"
a) Trader666's approach
MON AMI! The facts on how you screwed it up. Only the facts according to T6.
FACT. You did not construct a portfolio that was specified by the author of the original document.
FACT. You chose portfolios that were never specified in the original document.
FACT. All of your backtest results were negative.
MON AMI! There are many documents which have been amended from the original. Even the document you claim is an original is in fact NOT an original. Take for example, the US CONSTITUTION. It has been amended numerous times. The amendments are typically improvements on the original document. One of those amendments to the constitution gave me the right to vote. According to the original US CONSTITUTION, I would not have the right to vote.
Why anyone would choose to stick their head in the sand and ignore that which they could do to get favorable results is beyond most everyone. Many have reasoned that at this point, you can never show any agreement with any aspect of that document simply because you would then be hypocritical. However, you have revealed how remote your logic is (ie. not choosing to use CONDITIONS that would yield a positive EQ curve). So it goes! However, you have done a SUPERB job of clarifying all of your flaws and how it is that you got your results. Had you never responded, you wouldn't have exposed your reasoning flaws. But, you have, and it has many uses that credit what you have not done and how two backtests using the same document and code can arrive at different results; one that is desired as opposed to one that is not desired...
HAHA... For others,
AHA...
