|
PureBytes Links
Trading Reference Links
|
>>I'm referring to systems designed by optimising lookback periods.<<
Let me start by saying I am primarily a discretionary trader and I
don't do much in the way of optimizing, I consider it curve fitting.
When your data set changes character your optimization breaks down.
Using a fixed look back period is a very crude way of trading market
cycles. For example, if you have a simple MA crossover system that
does well in a trending market (longer term cycle is dominant) it will
do poorly in a sideways choppy market (shorter term cycle is
dominant). A more refined approach is to use a variable period that
adapts to the current dominant cycle.
>>I'm happy to be proved wrong ...so you are saying we can achieve
better than 30-40%PA, on long term average (through various market
cycles) using 'optimisation of lookback period' techniques? (EOD, no
leveraging).<<
IMO, trying to optimize a look back period is not the best approach
but if you reoptimize frequently enough so your look back period is
tuned to the current dominant cycle you likely can achieve better than
30%.
I think a more productive approach is to align your trading signals on
multiple time frames (e.g., daily and weekly).
Bill
--- In amibroker@xxxxxxxxxxxxxxx, "brian_z111" <brian_z111@xxx> wrote:
>
> If your trading system rules are based on things like "buy when the
> short term moving ave crosses the long term moving ave".
>
> The MA is looking back so many periods to make its calculation e.g. MA
> (C,15) is looking back 15 periods.
>
> If you test a range of MA periods, to select your best MA crossover
> system, then you are optimising the lookback period (at least that is
> what I mean).
>
>
> brian_z
>
> --- In amibroker@xxxxxxxxxxxxxxx, "Louis Préfontaine"
> <rockprog80@> wrote:
> >
> > Hi Brian and everyone,
> >
> > What exactly do you mean by "optimisation of lookback period"?
> >
> > I had a lot of fun reading this thread. I wonder what is better:
> > support/resistance breakout or reversion to mean. Worked with
> both; don't
> > know yet what works better. I've seen people been sure of their
> opinions,
> > but I'd like to read some arguments...
> >
> > Louis
> >
> > 2008/5/8 brian_z111 <brian_z111@>:
> >
> > > It's just an opinion, but it is based on observation.
> > >
> > > I'm referring to systems designed by optimising lookback periods.
> > >
> > > I'm happy to be proved wrong ...so you are saying we can achieve
> > > better than 30-40%PA, on long term average (through various market
> > > cycles) using 'optimisation of lookback period' techniques? (EOD,
> no
> > > leveraging).
> > >
> > > brian_z
> > >
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%40yahoogroups.com>,
> > > "bilbo0211" <bilbod@> wrote:
> > > >
> > > > "I will stick to my prediction that around 30%PA EOD trading is
> a
> > > > limit for indicators that use lookback periods and that to
> achieve
> > > > more than this requires a different approach (as I say you are
> both
> > > > correct except I believe that Steve is talking about >30%PA
> > > returns)."
> > > >
> > > > Is this just your opinion or do you have something that
> approaches
> > > > 'scientific proof' of this allegation?
> > > >
> > > > In "The Profit Magic of Stock Transaction Timing" by J M Hurst,
> the
> > > > author claims the theoretical maximum annual ROI for stock
> trading
> > > is
> > > > 2400%. ROI is directly related to the holding period for each
> trade
> > > > and being fully invested at all times (the 'Magic' is in the
> power
> > > of
> > > > compounding).
> > > >
> > > > Hurst recorded the results of a 6 week real time trading
> experiment
> > > in
> > > > which his performance trading high beta stocks approached his
> > > > theoretical maximum annual ROI.
> > > >
> > > > Hurst waited until the dominant cycles in his trading instrument
> > > were
> > > > in alignment before trading (this is also called multiple time
> frame
> > > > or multiple fractal alignment). He primarily used daily and
> weekly
> > > charts.
> > > >
> > > > The theoretical maximum ROI is actually much higher than 2400%
> if
> > > you
> > > > use intraday charts and leveraged trading instruments.
> > > >
> > > > If you look in the Amibroker Trading System Yahoo group, you
> will
> > > find
> > > > a poll of results of people's mechanical trading systems. IIRC,
> the
> > > > best ones listed returned over 400% per year.
> > > >
> > > > Bill
> > > >
> > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%40yahoogroups.com>,
> > > "brian_z111" <brian_z111@> wrote:
> > > > >
> > > > > 20 - (- 9.3_ == approx delta 30% PA in my books.
> > > > >
> > > > > Thanks Yuki for confirming this.
> > > > > Now I don't have to post a 30% system (as I promised Louis) to
> > > prove
> > > > > my benchmark is correct.
> > > > >
> > > > > Actually I agree with both you and Steve (the real problem is
> > > > > semantics since IMO close analysis would show that most of us
> are
> > > > > moementum traders and also that most of us are using a kind of
> > > S/R in
> > > > > some way - the difference is how we perceive and define these
> > > things).
> > > > >
> > > > > I will stick to my prediction that around 30%PA EOD trading
> is a
> > > > > limit for indicators that use lookback periods and that to
> > > achieve
> > > > > more than this requires a different approach (as I say you are
> > > both
> > > > > correct except I believe that Steve is talking about >30%PA
> > > returns).
> > > > >
> > > > > (Steve - care to confirm?)
> > > > >
> > > > > brian_z
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%
> 40yahoogroups.com>, Yuki
> > > Taga <yukitaga@> wrote:
> > > > > >
> > > > > > Gee, then I guess I should give back my ~20 percent a year
> that
> > > is
> > > > > > largely based on short-term momentum swings, yes? (I'm
> sitting
> > > plus
> > > > > > 13 percent YTD this year already, as of yesterday, versus -
> 9.3
> > > > > > percent for my Nikkei 225 benchmark.)
> > > > > >
> > > > > > You do have to be agile however. And you cannot overstay
> your
> > > > > > welcome. But the money is there for momentum systems if
> > > designed
> > > > > > and tested properly.
> > > > > >
> > > > > > "Support" exists, but everyone knows where it is. Exactly
> > > where it
> > > > > > is. And somebody (I'll leave it to you to guess who) is
> going
> > > to
> > > > > > ring the bell and tell you that (resistance failed) or
> (support
> > > > > > failed). What are you going to do, then? You're going to
> stop
> > > > > > yourself out of course. With a loser.
> > > > > >
> > > > > > Which is likely to be more profitable, and for a longer
> period
> > > of
> > > > > > time? Systems that compel you to do the psychologically
> > > difficult,
> > > > > > or systems that suggest that you do the patently obvious?
> > > > > >
> > > > > > Is there anyone beyond 7th grade that doesn't know where
> > > support and
> > > > > > resistance is? Are there great systems that rely on widely
> > > known
> > > > > > community knowledge?
> > > > > >
> > > > > > Look for a system that has good metrics, but a system that
> also
> > > > > > suggests that what you need to do will be psychologically
> > > difficult
> > > > > > for you to do, in spite of having back-tested results
> > > indicating
> > > > > that
> > > > > > you are foolish if you *don't* do it. Then you are good to
> go,
> > > as
> > > > > > they say. Good to go as long as you do it, of course.
> > > > > >
> > > > > > If your system is easy to follow (by that, I mean that it's
> > > > > > psychologically easy for you to make the trades), it's
> probably
> > > a
> > > > > > loser. And vice-versa. The best systems have good metrics,
> yet
> > > > > > despite that they almost defy the trader (psychologically)
> to
> > > make
> > > > > > the trades. There is no free lunch.
> > > > > >
> > > > > > Yuki
> > > > > >
> > > > > > Thursday, May 8, 2008, 11:50:01 AM, you wrote:
> > > > > >
> > > > > >
> > > > > > s> Anthony,
> > > > > >
> > > > > > s> Do yourself a big favor. Don't waste your precious time
> on
> > > this
> > > > > > s> earth with this kind of drivel. Chasing price with
> > > momentum
> > > > > > s> indicators is not going to get you where you want to be.
> > > > > >
> > > > > > s> Coming up with a support/resistance system is all you
> need
> > > to
> > > > > make
> > > > > > s> whatever you want from the markets.
> > > > > >
> > > > > > s> I've seen hundreds of traders get wiped out trying to go
> on
> > > the
> > > > > path
> > > > > > s> you're following and all of the successful traders I've
> been
> > > > > around
> > > > > > s> in the e-mini futures have used S/R as the foundation of
> > > their
> > > > > > s> trading methodology.
> > > > > >
> > > > > > s> And, above all, embrace your emotions in trading because
> > > they
> > > > > teach
> > > > > > s> you what you should and shouldn't do going forward.
> > > Computers
> > > > > learn
> > > > > > s> nothing while you learn from every win and loss you make.
> > > > > >
> > > > > > s> Finding an edge in trading is easy. It's only hard if
> > > you're
> > > > > using a
> > > > > > s> computer to find a needle in a haystack because you
> didn't
> > > make
> > > > > a
> > > > > > s> good enough investment in real-time observations of the
> > > markets
> > > > > while
> > > > > > s> researching an edge you'd like to trade.. That makes all
> > > the
> > > > > > s> difference in the world for knowing what works and what
> > > doesn't.
> > > > > >
> > > > > > s> You'll come up with 10 edges to trade if you put the
> time in
> > > to
> > > > > > s> experience a live market on a regular basis without
> trying
> > > so
> > > > > hard.
> > > > > > s> It will bring out your imagination and creativity to find
> > > what
> > > > > you're
> > > > > > s> looking for.
> > > > > >
> > > > > > s> I wish someone had told me that 4.5 years ago when I
> started
> > > > > trading
> > > > > > s> the ER2 e-mini. It would have saved me a lot of time
> > > chasing
> > > > > > s> nonsense.
> > > > > >
> > > > > >
> > > > > > s> --- In amibroker@xxxxxxxxxxxxxxx <amibroker%
> 40yahoogroups.com>,
> > > "ihsaham" <ihsaham@> wrote:
> > > > > > >>
> > > > > > >> Hai Tomasz,
> > > > > > >>
> > > > > > >> This is simple Jake Bernstein Momentum Formula for chart
> and
> > > > > > s> scanner.
> > > > > > >> Please help me give arrow buy and sell. Buy arrow is
> Green
> > > > > colour
> > > > > > s> and
> > > > > > >> Sell Arrow is Red Colour.
> > > > > > >>
> > > > > > >> I really appreciate and thanks for you in advance.
> > > > > > >>
> > > > > > >> Best Regards,
> > > > > > >> Anthony Idic
> > > > > > >>
> > > > > > >>
> > > > > > >>
> > > > > > >> _SECTION_BEGIN(" $ Momentum ");
> > > > > > >>
> > > > > > >>
> > > > > > >> /* Bernstein Momentum Indicator */
> > > > > > >> /* Set Scaling to Automatic, Show dates On, Percent On,
> > > Middle
> > > > > On */
> > > > > > >>
> > > > > > >> Title = "Bernstein MOM Close - Ref(Close,-7)";
> > > > > > >> GraphXSpace = 5;
> > > > > > >> Graph0 = MA(Close - Ref(Close,-7),1);
> > > > > > >> Graph0Style = 5;
> > > > > > >> Graph0Color = 29;
> > > > > > >> Graph1 = MA(Graph0,5);
> > > > > > >> Graph1Style = 1;
> > > > > > >> Graph1Color = 32;
> > > > > > >>
> > > > > > >>
> > > > > > >> DaysAgo =Optimize("DaysAgo",-28,-40,-16,4);
> > > > > > >> Fast = Optimize("Fast", 1, 1,5,1);
> > > > > > >> Slow = Optimize("Slow",28,16,40,4);
> > > > > > >> /* Note: It is merely a coincidence that DaysAgo and Slow
> > > use
> > > > > the
> > > > > > >> same parameter set. */
> > > > > > >>
> > > > > > >> Buy = Cross( MA(Close - Ref(Close,DaysAgo),Fast),
> > > > > > >> MA(Close - Ref(Close,DaysAgo),Slow) );
> > > > > > >>
> > > > > > >> Sell = Cross( MA(Close - Ref(Close,DaysAgo),Slow),
> > > > > > >> MA(Close - Ref(Close,DaysAgo),Fast) );
> > > > > > >>
> > > > > > >>
> > > > > > >> Short = Cross( MA(Close - Ref(Close,DaysAgo),Slow),
> > > > > > >> MA(Close - Ref(Close,DaysAgo),Fast) );
> > > > > > >>
> > > > > > >> Cover = Cross( MA(Close - Ref(Close,DaysAgo),Fast),
> > > > > > >> MA(Close - Ref(Close,DaysAgo),Slow) );
> > > > > > >> _SECTION_END();
> > > > > > >>
> > > > > >
> > > > >
> > > >
> > >
> > >
> > >
> >
>
------------------------------------
Please note that this group is for discussion between users only.
To get support from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
For other support material please check also:
http://www.amibroker.com/support.html
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/amibroker/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/amibroker/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:amibroker-digest@xxxxxxxxxxxxxxx
mailto:amibroker-fullfeatured@xxxxxxxxxxxxxxx
<*> To unsubscribe from this group, send an email to:
amibroker-unsubscribe@xxxxxxxxxxxxxxx
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|