|
PureBytes Links
Trading Reference Links
|
Steve,
Detrending is to remove, to filter out a trend.
there are various ways to do it, according to the application.
In the specific mail below, you may remove the Long Part [or
respectively the Short part] of trading results from Settings window,
selecting *Long [or *Short] trades only, to see which part of the
profits is due to each side.
DT
--- In amibroker@xxxx, "Steve Dugas" <sjdugas@xxxx> wrote:
> Hi All,
>
> I have seen people write several times about "detrending". I
assume, as the
> name implies, that it has something to do with factoring out trend
related
> data or characteristics. Can anyone please explain what exectly is
> detrending, why we would want to do it, and how we would go about
doing it?
> Thanks very much?
>
> Steve
>
>
> ----- Original Message -----
> From: "dtsokakis" <TSOKAKIS@xxxx>
> To: <amibroker@xxxx>
> Sent: Sunday, July 21, 2002 7:37 AM
> Subject: [amibroker] Re: Fw: The Magic OUT3 [and the STOCHRSI]
>
>
> > Mike,
> > I did not optimize StochRSI parameters. I never traded quick
StochRSI
> > systems.
> > For QQQ I prefer RSIt slow systems
> > [11 trades/11 winners/0 losers since March 2000]
> > with a cool performance nearly +500%
> > I followed two of them since Jan 2002 with a cool result 5/5/0.
> > [+12%, +13%, +14%, +13% and the last short since June 4 running
today
> > at +15%, not covered yet]
> > QQQ is not the security for great profits, so I prefer low risk.
> > The same cool behavior for CSCO.
> > RSIt steady level slow systems were VERY protective for WCOME
> > traders, [as for ENRON in the past...]
> > [nearly +1500% since March 2000, with the following performance
this
> > year
> > March 4, Long -1%
> > March 5, Short +83%
> > May 14, Long +25%
> > May 22, open Short +93%]
> > It is difficult for a 11/11/0 system to make wrong entries.
> > It is somehow autoprotected. For WCOME, the duration of Long
exposure
> > was always just a few days, as if the system "knew" how to protect
> > your investment.
> > I can not explain more, I feel quite safe with RSIt oscillations.
> > Of course, as my friend b pointed out, the performance was not the
> > best for 90s.
> > [RSIt was not even existing in 90s, since I designed it ...last
year]
> > But, we trade in 2002 and MANY "profitable" systems of 90s caused
> > severe losses the last 30 months.
> > In last analysis, trade with the trend, not with memories.
> > The 90s traders and analysts make a crucial mistake : They do not
> > detrend the B&H profits.
> > We will make the same mistake, if we do not detrend Short profits
for
> > 2000-2002 and we are just impressed by huge percentages.
> > With RSIt you could make interesting Long profits, even with WCOME
> > [+35%]
> > You may consider this +35% funny, but remember that it is better
than
> > buying at $40 or $30 and sell at $0.06 on July 1st !!!
> > Dimitris Tsokakis
> >
> > --- In amibroker@xxxx, "mik954" <mik-u@xxxx> wrote:
> > > Dimitris,
> > >
> > > Thanks for posting the system. QQQ results are very impressive
for
> > > this specific period.
> > >
> > > It returns even more than 750% for period 2000-current with EMA
> > > period = 2, Period = 8, BLevel = 12, and Y <= 2.
> > >
> > > Good luck,
> > > Mike
> > >
> > > --- In amibroker@xxxx, "dtsokakis" <TSOKAKIS@xxxx> wrote:
> > > > Try also the
> > > >
> > > > Period =15;
> > > > Blevel =30;
> > > > Slevel = 100 - Blevel;
> > > > Y=Foreign("~OUT3","V");
> > > > XX=Foreign("~COUNT","V");
> > > > D=DateNum()>1000301;
> > > > Buy = Cross(Blevel,EMA((RSI(period)-LLV(RSI(period),period))/
(HHV
> > > (RSI
> > > > (period),period)- (LLV(RSI(period),period))),3)*100) AND Y<=2;
> > > > Buy=D*Buy;
> > > > Sell = Cross(EMA((RSI(period)-LLV(RSI(period),period))/ (HHV
(RSI
> > > > (period),period)- (LLV(RSI(period),period))),3)*100, Slevel);
> > > > Sell=D*Sell;
> > > > Short = Sell;
> > > > Cover = Buy;
> > > >
> > > > which takes the final profits to +630% for the same period and
> > > > settings.
> > > > Since the period is increased we have only 24 trades, 19
winners,
> > 5
> > > > losers.
> > > > Without the magic OUT3, this system would remain nearly +120%.
> > > > This level of +600% to +650% for total Net profit, is
> > satisfactory
> > > > for QQQ. Various conditional oscillators converge to this
> > > interesting
> > > > limit.
> > > > Dimitris Tsokakis
> > > > --- In amibroker@xxxx, "Dimitris Tsokakis" <TSOKAKIS@xxxx>
wrote:
> > > > >
> > > > > ----- Original Message -----
> > > > > From: Dimitris Tsokakis
> > > > > To: amibroker@xxxx
> > > > > Sent: Saturday, July 20, 2002 2:51 PM
> > > > > Subject: Fw: The Magic OUT3 [and the STOCHRSI]
> > > > >
> > > > >
> > > > > Since it was to be continued, here is the application on QQQ
> > > > StochRSI.
> > > > > Apply the magic Y<=2 to improve profitability from +242% to
> > > > +474% !!!
> > > > > Just with the magic OUT3.
> > > > > Settings
> > > > > buy, sell, short, cover at open, delay +1
> > > > > commission 0.5%, all stops disabled
> > > > > Application QQQ
> > > > > period March 01, 2000 till now.
> > > > > /*StochRSI system*/
> > > > > Period = 8;Blevel = 17;Slevel = 100 - Blevel;
> > > > > D=DateNum()>1000301;
> > > > > Buy = Cross(Blevel,EMA((RSI(period)-LLV(RSI
(period),period))/
> > (HHV
> > > > (RSI(period),period)- (LLV(RSI(period),period))),3)*100);
> > > > > Buy=D*Buy;
> > > > > Sell = Cross(EMA((RSI(period)-LLV(RSI(period),period))/ (HHV
(RSI
> > > > (period),period)- (LLV(RSI(period),period))),3)*100, Slevel);
> > > > > Sell=D*Sell;
> > > > > Short = Sell;
> > > > > Cover = Buy;
> > > > > 56 trades, +242%
> > > > > /*Improved StochRSI system with the Magic OUT3 additive*/
> > > > > Period = 8;
> > > > > Blevel = 17;
> > > > > Slevel = 100 - Blevel;
> > > > > Y=Foreign("~OUT3","V");
> > > > > D=DateNum()>1000301;
> > > > > Buy = Cross(Blevel,EMA((RSI(period)-LLV(RSI
(period),period))/
> > (HHV
> > > > (RSI(period),period)- (LLV(RSI(period),period))),3)*100) AND
> > Y<=2;
> > > > > Buy=D*Buy;
> > > > > Sell = Cross(EMA((RSI(period)-LLV(RSI(period),period))/ (HHV
(RSI
> > > > (period),period)- (LLV(RSI(period),period))),3)*100, Slevel);
> > > > > Sell=D*Sell;
> > > > > Short = Sell;
> > > > > Cover = Buy;
> > > > >
> > > > > 34 trades, +474%
> > > > > Not bad at all...
> > > > > Dimitris Tsokakis
> > > > > PS. Applying the simple STOCHRSI system on the bloody WCOME
you
> > > end
> > > > with a -87% for the same period.
> > > > > With the magic OUT3 additive, you are at a nice +23% and, if
> > you
> > > > are a WCOME fun, you may keep on trading this stock
> > > > >
> > > > > ----- Original Message -----
> > > > > From: Dimitris Tsokakis
> > > > > To: amibroker@xxxx
> > > > > Sent: Wednesday, June 26, 2002 4:06 PM
> > > > > Subject: The Magic OUT3
> > > > >
> > > > >
> > > > > Are you disappointed from your trading system ?
> > > > > Apply the magic condition and you will be surprised !!
> > > > > Create first the "~OUT3" artificial ticker.
> > > > > Scan the whole N100 with
> > > > >
> > > > > /*OUT 1, 2, 3*/
> > > > > KUP=EMA((H+L+C)/3,10)+EMA(H-L,10);
> > > > > KDOWN=EMA((H+L+C)/3,10)-EMA(H-L,10);
> > > > > OUT1=C<KDOWN AND Ref(C,-1)>Ref(KDOWN,-1);
> > > > > OUT2=C<KDOWN AND Ref(OUT1,-1);
> > > > > OUT3=C<KDOWN AND Ref(OUT2,-1);
> > > > > AddToComposite(OUT3,"~OUT3","V");
> > > > > Buy=0;
> > > > >
> > > > > The "magic" ticker is Y, defined from the relation
> > > > > Y=Foreign("~OUT3","V");
> > > > > Let us see some applications.
> > > > > a. The MACD() crossover.
> > > > > We all know that MACD crossover fails in bearish markets.
> > > > > Except if we add the magic OUT3 Y==0 or Y==1.
> > > > >
> > > > > The MACD crossover, after March1, 2000 is
> > > > >
> > > > > D=DateNum()>1000301;
> > > > > Buy=D*(Cross(MACD(),Signal()));
> > > > > Sell=D*(Cross(Signal(),MACD()));
> > > > > Short=Sell;Cover=Buy;
> > > > >
> > > > > On QQQ gives a -67%.
> > > > > Magic OUT3 may change it to +9.36% with
> > > > > /*Magic 1*/
> > > > > Y=Foreign("~OUT3","V");
> > > > > D=DateNum()>1000301;
> > > > > Buy=D*(Cross(MACD(),Signal()) AND Y==1);
> > > > > Sell=D*(Cross(Signal(),MACD()));
> > > > > Short=Sell;Cover=Buy;
> > > > >
> > > > > or to +60% with
> > > > >
> > > > > /*Magic 2*/
> > > > > Y=Foreign("~OUT3","V");
> > > > > D=DateNum()>1000301;
> > > > > Buy=D*(Cross(MACD(),Signal()) AND Y==0);
> > > > > Sell=D*(Cross(Signal(),MACD()));
> > > > > Short=Sell;Cover=Buy;
> > > > >
> > > > > For the whole N100, the simple MACD() crossover is "extreme
> > > > game" : -61%
> > > > > with 5 profitable and 96 not profitable stocks !!!
> > > > > With the magic OUT3 may rise to +3% or +9% for the
respective
> > > > additions.
> > > > >
> > > > > B. The Stochastics 30/70 system
> > > > > Another nice but suffering system
> > > > > For the whole N100, the "traditional" 30/70
> > > > >
> > > > > D=DateNum()>1000301;
> > > > > Buy=D*(Cross(StochD(),30));
> > > > > Sell=D*(Cross(70,StochD()));
> > > > > Short=Sell;Cover=Buy;
> > > > >
> > > > > gives a -30%.
> > > > > Touch it with the magic OUT3 and you change from -30% to
+30% !!
> > > > >
> > > > > Y=Foreign("~OUT3","V");
> > > > > D=DateNum()>1000301;
> > > > > Buy=D*(Cross(StochD(),30) AND Y==1);
> > > > > Sell=D*(Cross(70,StochD()));
> > > > > Short=Sell;Cover=Buy;
> > > > > For example, YHOO changes from +32% to +250%.
> > > > >
> > > > > So, if you still use loosing systems, you may improve their
> > > > performance, adding Y==0 or
> > > > > Y==1 or Y<2 etc to the previous Buy condition.
> > > > >
> > > > > Dimitris Tsokakis
> > > > > PS Are you curious for the respective magic Sell additive ?
> > > > > [to be continued]
> >
> >
> >
> >
> >
> > Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
> >
> >
> >
|