| jschott |
|
Gold User, Member, Platinum User, TeleChart
|
| Registered User |
| John Schott |
| Philly area |
| Investor and lazy retiree |
| Stocks, Active AAII member - Philadelphia Chapter & SIGS |
| Male |
|
| Wednesday, October 27, 2004 |
| Monday, September 06, 2010 2:23:16 AM |
743
[0.36% of all post / 0.35 posts per day] |
|
Trouble is there is no hard and fast definition of what is a head and shoulders..
Even worse, it's nearly impossible to write code to describe something that readily looks like something your eyes can instantly spot. Your eye adapts to your sense of what's good enough to pass your criteria (and visa versa) - one set of code can't cover all the possibilities of "good enough" or "not".
|
OOPS. I note that KEmpferts read the offsets as +5 and -2. I have note yet tried them, but I'd still like to know where the offsets came from.
Thanks
=====================================================
QUOTE (jschott) The figure in Technical Analyhsis of Stocks and Commodities appears to require use an offset of +3.0 and -3.0 and a date of 2009, Jly 25th for two VWAP indicators to make the Channel shown in Figure 9 for SPY there.
When I do this, I get a relatively good match with the lower channel line, but the upper line is too low based on the figure 9
Since the explanation in TASC does not mention that the VWAP provided needs this setup, it's not clear to me if the TASC picture shows the correct plot or if what I get now is correct.
I do note that of the more than a dozen comparable plots made using other TA programs noone else used the same start date nor settings as the published example. Naturally the start date is arbitrary, but I'd also like to know why 3.0 was used as the offset in the example you and Patrick Argo submitted.
Thanks.
|
The figure in Technical Analyhsis of Stocks and Commodities appears to require use an offset of +3.0 and -3.0 and a date of 2009, Jly 25th for two VWAP indicators to make the Channel shown in Figure 9 for SPY there.
When I do this, I get a relatively good match with the lower channel line, but the upper line is too low based on the figure 9
Since the explanation in TASC does not mention that the VWAP provided needs this setup, it's not clear to me if the TASC picture shows the correct plot or if what I get now is correct.
I do note that of the more than a dozen comparable plots made using other TA programs noone else used the same start date nor settings as the published example. Naturally the start date is arbitrary, but I'd also like to know why 3.0 was used as the offset in the example you and Patrick Argo submitted.
Thanks.
|
[The following duplicates the intent of a previous posting elsewhere]
I see some April comments about a locked posting on TDSequential that I made some time ago. Several people expressed their idea of why it was locked. They were wrong.
[BTW: Notice the number of locked postings vs. the unlocked ones. Then not the number that have had almost no accesses. I had suggestedpreviously that Worden separate the two crowds, as many of the locked postings seem not merely member to member communications, but often items seeming of little interest to advanced users.]
As I said then, the posting was very similar to the open posting I had made previously. I made some small changes (never finished) and locked a copy for the Philadelphia AAII Chapters Somputerized Investing SIG. I've forgotten the passcode, but could look it up if anyone is interested.
As to the results of my test, noone in the SIG ever accessed it - which was why I locked it - to see if anyone did.
|
After being off the forums for a while, I just noted someone asked about Demark indicators for TC. I posted one on 11/19/08 which you can find with a search. It's only a rought cut at the Setup Part.
Note that since TC "can't loop and count", I was stuck on a practical way to do the Countdown part (and do all the options on can find in Perl's book).
So that's as far as it does, but it's still interesting as a start.
|
DEMARK TD SEG code
I just posted another thread on this section of the forum (it's not in this one, but dated today 05/07/10) about why I had locked the code. However, I would point out that there are unlocked versions for Stock Finder and a similar set for TC posted in the appropriate forums.
If anybody is interested in this locked interrum code, I could certainly look up the lock code or send them the actual content. But read my "reason why" posting first.
See also the other posting if you are "jas0501" - all of your guesses were wrong - nothing sinister.
|
|
Why I locked my previous TD Seq. Lookalike code.
[An aside - It you look up the locked postings by a host of people, and note that most of them have zero accesses, you would wonder why people would lock things that nobody cared about. Why not share - as the titles on most of the locked items do not seem to represent any read breakthroughs. It's only people like jas0501 and pthegreat that seem to be doing "great things".}
I see several 2010 postings citing my locked code. I did it for a reason. I was a member of the AAII Philadelphia Chapter Computerized Investing SIG and I was trying to see if any of the members were interested in SF code. (Result: none were the slightest interested)
If you look back, you will see I had posted some previous unlocked versions on the forum. The locked on is a very similar version, just slightly enhanced and not debugged.
I don't even remember the locked code password, but I could look it up and post it - however the unlocked version isn't that much different.
jas0501 ventured some reasons - none of which apply. I merely felt that what was in this version was similar to what was already posted. It was about that time that other things got in my way, so I have not done much more. I surely would not want to sell anything - (and I'm sure that would get Demark interested).
Pthe great has gone far further .......
At the time I was working on the next phase - Countdown and had found Perl's very good book and was trying to decide what to do. I ended up with so much else on I mind, I didn' get much further.
TC users will note that I also posted a parallel version of the Blocks type code. It's a bit different, but incorporates much the same information. However, remember that there is really no way to "count" things in TC - except linear coding. Since the Demark counts can reach 20 or more, this is a pain. Also note that counting the non-consecutive counts or 13 or more in TC seems impractical to impossible.....
|
There are several places that evaluate trading systems:
- The Collective 2 site (mentioned above) (Note the site also markets them)
- Futures Truth: It's really a magazine which quarterly rates mechanical trading systems. This is perhaps the grandadie of raters. Mr. Hill has a testing engive and will rate any submitted package according to a strict set of rules.
- Wealth-Lab which allows posters to get testing on their systems using the Wealth-Lab language. (The underlying Wealth-Lab product is not available in America as Fidelity Investments has an exclusing license on the software. You can however use a version of it on line for free.
There are a lot of books and articles about mechanical trading systems.
I note at least two problems
1.) A lot of systems have their run. That is the work well for a period of time and then they go sour. OFten the providers upgrade the systems seeking to restore their performance record. Note that Futures Truth strictly rates systems in a foreward mode to preserve their validity. That is any upgrade is only tested into the future - no retroactive testing to cover market action prior the the specific release.
2.) If you read books on testing and well known systems and their users, there is an interesting pattern. That is "anywhere but stocks". In the books you see stocks discussed early in the book, but seldom further on where the evaluation focuese on commodities, gorex and the like. Likewise the traders trade anywhere but stocks EXCEPT when the market is clearly trending when they will trade stocks.
For example, one system called "Aberation" has long been one of the market leaders - but over the years has run hot and cold
This is a hint that suitability to a mechanical trading system favors forex and commodities as they are more likely to have trending runs that are more consistant. Stock trends have more risk and volitility, so the Pros act accordingly.
I should note that there are a number of famous trading systems that have had good performance over the years. But they have their good and bad intervals and what's the best tool today may be a road to poverty some year soon. Futures Truth (which does not sell any systems) shows the vendors, the prices and the their current FT rating. But take a look at the people who sell them - they may have a dozen or more systems for sale - some for specific markets and some not. MIght that be a hint that even the best of them know that one expensive system is not the magin answer.
So take a look at these sites and others you may find. Get a feeling and remember, there is likely no EASY and ASSURED way to make money in stocks, especially as a small and active trader. Mechanical trading systems by themselves would be used by everyone if they were the easy road............
|
|
Superneat, Dr. Wizard, love that real code!!!!!
|
OOPS, sorry, even the Metastock code does not provide coloration informaton. Deciphering the Trade Station plot (p57)and then looking at the Mode 59 provided figure/code (p79) and the code provided for VT Trades (p79) provide the only hints of what's behind the plot shown for StockFinder.
Still it would be nice to have your version of how the RealCode coloration/arrows in the price pane were created.
Thanks again.
|
|
