|
PureBytes Links
Trading Reference Links
|
John,
My questions were meant to get at understanding how the programming
worked. 20 DMA X-Over is simple, but when you get into othe types of
programming I'm lost. The backtester I use requires me to set up the
buys and sells in a particular fashion. Not understanding the
programming made it difficult for me to attempt with the SPY system
that was published.
I published some of the posts in responce to my questions and I
fully plan to revisit the issue.
Another issue I'm just coming to grips with is the Reward to Risk
aspect. My SPY system certainly does not generate the 22% or whatever
the system put forward on this board apparently generates. And I
think my Reward to risk on my SPY system is only 1.5. Annualized rate
/ Drawdown. But I'm hearing that at least 2.0 is desired and some
shoot for much better R to R figures before they will trade a system.
Now a qqq system I've just set up looks like a 6.x Reward to Risk
and I'm probably going to toss a few dollars at it and see how it
plays. Nice smooth equity line over time and so forth...worth a shot
I guess.
My nagging problem with my QQQ system is that the QQQ doesn't have as
much data to work with as SPY and individual stocks I've set up. And
I've got a feeling if there were more data available on the QQQ
(meaning data that would have been created in other types of markets)
then my returns would probably drop, drawdown would increase, and R To
R would degrade. So I realize that I have some nice QQQ numbers and
it's most likely because there is not enough data to account for most
different sorts of markets I'll be seeing in the future with the QQQ.
As always, the hardest part for me is the actual programming. The
simplest of programming tricks can take me days to figure out. And
that's all I was after on that system you folks published. Some
understanding of how the programming worked.
To be honest, I've been so busy working on the trading projects
I've assigned myself on this 2 weeks of down time I have, that I
haven't been able to return to that SPY system and try to work it out.
Even now I have work running in the background.
Thanks for your responce. I seriously doubt that I have anythign
to really contribute to this group but I'll think about it. An
amateur system builder that can't program certainly wouldn't have much
to contribute....I would think.
--- In equismetastock@xxxxxxxxxxxxxxx, johndoeforever49
<no_reply@xxxx> wrote:
> Metastockuser--You wrote a nice explanation of what you are trying to
> do--especially for a beginner. It appears you've read a lot of books.
> You certainly have the terminology down. Congrats on developing your
> own SPY system. Maybe you can share it with the group!
>
> I re-read all of the posts and I don't think anyone was particularly
> upset. It sounds to me like manohohman didn't want to respond to
> anymore questions with answers that should have been transparent to
> begin with.
>
> For example, Zweig's 4% trading rules are well known. You can look
> that up for yourself. It has nothing to do with any code. Then there
> was a 20 day moving average added to that. How hard can that be? So
> setting up a systems test of your own should be easy, if you will
> look up a couple of definitions. If you can't set up a systems test
> based on the original Zweig's trading rules with the moving average
> added, then you can't trust the results of your test anyway.
>
> He said there were quesitons about curve fitting and out of sample
> data. Sounds like a lot of people don't understand curve fitting.
>
> This system was presented for only one security in one time frame.
> You can't change the security, the time frame, or the trading rules
> which were defined by someone else. The 20 day moving average is the
> most used moving average there is. He said the backtest results he
> looked at were for 10 years. That's straight forward to me. What is
> there to curve fit?
>
> I think he said he used the system as a trend indicator, rather than
> something to trade everyday, although you certainly could do that
> with it. He mentioned using other indicators with it to prevent
> whipsawing if he were going to trade it, but he didn't say what
> indicators he would use. I don't think backtesting really defines the
> amount of whipsawing a system has, or the conditions under which it
> takes place. I assume that's why he said he goes over the charts by
> hand. I wonder if the drawdowns are easily seen by hand.
>
> Since you've got your own SPY system, maybe you could take the charts
> from it for a couple of years and compare them to the charts from the
> posted system and see how much better yours is. Then you could post
> your system with an explanation of how and why it improves on the 4%
> system. That would really benefit the group.
>
> This 4% system is as simple as it gets. It sort of sets a floor to
> compare other SPY systems to.
>
> Of course I didn't see manohohman's emails, just the posted
> responses. I didn't see any questions on drawdowns or losing trades.
>
> After I joined, I read through the archives. I was hoping more people
> would post unique systems or techniques for trend following and other
> stuff. But so far I haven't found much.
>
> I see a lot of questions on programming issues, some discussion on
> theory, and a few questions on how to use metastock, but nothing that
> would really make metastock more effective to trade with.
>
> I guess no one has anything to share, or maybe there's no pay off for
> sharing!
>
> Metastockuser, here's your chance to start the ball rolling on the
> sharing thing. Hope you will take this opportunity to do so.
>
> John
>
> --- In equismetastock@xxxxxxxxxxxxxxx, "metastockuser"
> <metastockuser@xxxx> wrote:
> > John,
> >
> > I'm merely trying to understand it so I can break it out and test
> > it. I've recently gone live with my own SPY system. If I were to
> > find a better system I'd sure be willing to trade it.
> >
> > But I have to see the equity line over ten years or so of data in
> > different markets and so forth. check out the drawdown, etc etc.
> > Also, I want the stats so I can know EXACTLY what to expect going
> > forward in number of losing trades in a row, winning trades in a
> row,
> > average loss, average win, relative and absolute drawdown, and so
> > forth. So if it encounters a market it does not work well in I'll
> > know by the behavior against past stats hopefully before my bottom
> > line is seriously affected. But that means I have to run the tests
> to
> > have the data on hand to use as control limits for using or no
> longer
> > using the system.
> >
> > One reason I like tradesim for back testing is that it makes it
> easier
> > for me to see on a month by month and year by year basis exactly how
> > the system plays out. I also don't have to play games with it to
> get
> > it around 8.0 issues regarding next day purchases being cancled and
> so
> > forth.
> >
> > I just don't see why anyone would take offence at someone trying to
> > understand an indicator so they can back test it. It took longer to
> > lecture me on it than it would have to put a few lines of
> > understanding by each code line so I could understand what it was
> > actually doing.
> >
> > My problem is that I'm so new to all this and programming in
> > particular that it's like trying to understand greek. But I'm
> > certainly smart enough not to play a system that I haven't first run
> > my own back tests against. Hand work is fine, and I fully
> understand
> > the need to work through enough trades by hand to make sure the
> system
> > test is valid. I find setting up an expert system and comparing it
> to
> > trade logs helps in that regard. But nothing beats running a system
> > tester of years and years of data in various types of markets to get
> > at the performance of a system in a reasonable amount of time.
> >
> > I appreciate the work this guy did and am sure he did a fine job.
> As
> > soon as I have time to work it out I'm sure I'll enjoy running the
> > test and seeing how it performs in various markets.
> >
> > Just pasting someone elses work into my expert system and using it
> > without understanding it and backtesting it is not something I would
> > ever do.
> >
> > --- In equismetastock@xxxxxxxxxxxxxxx, johndoeforever49
> > <no_reply@xxxx> wrote:
> > > I don't think he backtested it at all. I think he said he looked
> at
> > > the charts for several years to see if he could use it to get
> similar
> > > results to the backtested results. I think he said it did.
> > >
> > > As I remember it, manohohman said a lot of other people had
> > > backtested this. Along those lines, I remember seeing results
> from
> > > Zweig, Mcdonald, and some other people who have looked at this
> system
> > > since it was developed. Each one was a little different from the
> > > original one Zweig came up with. But I don't remember everybody
> who
> > > has commented on it, or even when.
> > >
> > > I don't remember him recommending that anybody trade this. I
> > > certainly never would rely on the results of other systems
> developers
> > > like Zweig.
> > >
> > > From my stand point you are exactly right--don't use this.
> > >
> > > John
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/BefplB/TM
---------------------------------------------------------------------~->
To unsubscribe from this group, send an email to:
equismetastock-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|