|
PureBytes Links
Trading Reference Links
|
Hi dingo,
There are two simple ways to verify that you did not over-optimize your
system:
1) I always chart each parameter I optimize to see if the parameter has a
pattern, ie changes in some predictable fashion wrt another parameter (don't
only plot parameters vs equity, analyze relationships between other
parameters as well). If I can i will plot two optimized parameters at the
same time on a 3D plot. If I optimize 3 parameters I make one 3D chart for
each value of the parameter with the smallest range - yes you get dozens of
charts but you may start to see very interesting patterns! If you get lots
of random peaks and you cannot see any kind of normal distribution you have
probably over-optimized. Too many people think designing a trading system is
simple...you should document it like a business plan - after all you will
risk similar sums of money with it.
2) I always reserve a fair bit of data for out of sample testing. I develop
may systems on data before 1/1/01 and test it out of sample on data after
that date. I never adjust parameters to make it perform better in the OOS
data. If it doesn't work i start from scratch, something i have done
hundreds of not thousands, of times. Ideally your equity curve shouldn't
show a ripple when you change data.
Not everybody believes in these two approaches because they think or assume
it is impossible to get "perfect" out-of-sample performance for systems
developed on data from 2-10 years ago. But it is possible, for the QQQ and
for many other stocks.
Good luck,
Herman.
-----Original Message-----
From: dingo [mailto:dingo@x...]
Sent: Sunday, July 21, 2002 11:39 AM
To: amibroker@xxxxxxxxxxxxxxx
Subject: RE: [amibroker] StoRSI
Hi WJ,
>From all of my reading on this list it seems that every time on ventures
off and starts optimizing every parameter in the foundation of a formula
that the dreadful "over optimization" term starts rearing its ugly head.
\
I'm far from an expert and am still a little confilicted on it. I can
understand the desire to have a "robust" solution for a basket of stocks
but I can't help but think that if you want to develop a tailored
solution for a very high profile vehicle like qqq it should be more than
ok.
dingo
-----Original Message-----
From: willem1940 [mailto:w.j.a.struyck@x...]
Sent: Sunday, July 21, 2002 12:25 AM
To: amibroker@xxxxxxxxxxxxxxx
Subject: [amibroker] StoRSI
Herman, Dingo,
Here my penny contribution to trying something new. I have modified
the StoRSI so you can optimize "period" and "delta". Using this
modified AFL I have experimented with the ADX addition but the
results in general beco0me worse except for some volatile stocks.
Curious what you think.
Period = Optimize("Period", 6,2,40,2);
delta=Optimize("Delta",2,2,20,1);
Buy = Cross(delta, EMA((RSI(period)-LLV(RSI(period),period))/ (HHV(RSI
(period),period)- (LLV(RSI(period),period))),3)*100) ;
Sell = Cross(EMA((RSI(period)-LLV(RSI(period),period))/ (HHV(RSI
(period),period)- (LLV(RSI(period),period))),3)*100, 100-delta);
Short = Sell;
Cover = Buy;
Buy = ExRem( Buy, Sell );
Sell = ExRem( Sell, Buy );
Short = ExRem( Short, Cover );
Cover = ExRem( Cover, Short );
Filter = 1;
Willem Jan
--- In amibroker@xxxx, "Herman van den Bergen" <psytek@xxxx> wrote:
> Message-----Original Message-----
> From: dingo [mailto:dingo@x...]
> Sent: Saturday, July 20, 2002 11:40 AM
> To: amibroker@xxxx
> Subject: RE: [amibroker] StoRSI - Try someting different!!!
>
>
> Herman,
>
> ...Would you care to step me thru them? Perhaps share how you
made them
> and how to interpret them?
>
> [Herman van den Bergen]
>
> Disclaimer: I am just an amateur number cruncher, I am not an
expert on
> trading stocks.
>
> The charts were made in Excel. When you perform an optimization
the result
> tables displayed will have columns for the parameters that were
varied to
> find the optimum performance. They are the right-most columns in
the report.
> When you optimize two (only two!) parameters you can use the Export
Button
> to export the result table to a .csv file and make 3D charts. First
you
> export the results using the Export button and then you import it
into
> Excel.
>
> To make it easy to make 3D charts you should install the 3D
Charting
> plugin for Excel attached. Instructions on how to do it are in the
text
> file. If you have problems email me privately, perhaps I will be
able to
> help.
>
> The flat chart [ TradeLimiting1.jpg ] plots the Long durations
along the
> Horizontal axis and the Short durations along the Vertical axis on
your
> screen. The profit at any point where two durations cross is given
by color.
> For example, look up ShLim (Short Limit) = 5 bars follow its line
to the
> left until you reach LLim (Long Limit) = 23, the color at that
point is
> light blue. Going to the color table to the right of the plot you
can see
> that the gain at that point is 1100% to 1200% (Excel plotted this in
> integers, % is *100. Sorry I forgot over what period that was).
This means
> that if you limit your trade durations with something like this:
> LLim = Optimize("LLim",9,1,30,1);
> ShLim = Optimize("Sh:im",1,1,30,1);
> Sell = (BarsSince(Ref(Buy,-1)) > LLim) OR Sell;
> Cover = (BarsSince(Ref(Short,-1)) > ShLim) OR Cover;
>
> to 23 Days for long trades and to 5 days for Short trades, you'll
get the
> highest gain possible for the formula used. An extra advantage is
that you
> are more time in cash, your exposure to risk is less.
>
> The 3D chart works the same except here the gain is plotted in 3D
format.
> Not as easy to see here but in Excel you can rotate the chart and
you can
> clearly see that there is a high gain ridge at ShLim=5, beyond that
point
> the gains drop. The completely flat (light blue) area shows that
there were
> no trades with those durations; the gain didn't change. One would
have to
> further analyze this to see why it happens, it could mean out Long
signal is
> late.
>
> Regarding the "parabolic" idea.. I have been researching the SAR
function
> and it seems to have an implied "parabolic" incorporated in it. Is
there
> anther function that you are referring to?
>
> [Herman van den Bergen]
> How about simply reversing the position when a certain profit is
reached
> :-)
>
> Unzipping my gorilla suit and stepping out of it for a moment (Mad
> Magazine circa 1959) to get some perspective, it seems to me that
we have a
> number of areas to persue in order to fine tune the StochRSI (or
ocillator)
> approach:
>
> 1. Get better at the entry process (by adding something (ADX) to
encourage
> the entry to be consistent with the direction of the price movement)
>
> 2. Get better at the exit process and at the same time better at
reducing
> losses. I think we have to do these 2 in tandem...?
>
> 3. Assuming we can achieve 1 and 2 I'd then appreciate some
guidance on
> how we migh incorporate "money management" techniques into this
discussion.
>
> I really appreciate the fact that Steve started this discussion!
I am at
> the point in my learning curve that I'd like to try to use all of
the
> reading and experimenting that I've been doing for the last month
to develop
> an approach for assembling a "system" that would be a model for me
to follow
> in the future. I also appreciate the fact that you others have
joined in on
> this thread and would encourage anyone that wants to jump in to do
so.
>
> Sincerely,
>
> dingo
>
> -----Original Message-----
> From: Herman van den Bergen [mailto:psytek@x...]
> Sent: Saturday, July 20, 2002 7:52 AM
> To: amibroker@xxxx
> Subject: [amibroker] StoRSI - Try someting different!!!
>
>
> Hello dingo, great to see the new ADX angle introduced! Has
somebody
> charted it yet?
>
> Attached two charts to show that there is an optimum max trade
duration
> for the StoRSI (for most osc systems in fact). This phenomena can
be charted
> in many ways, try making a bar chart for each, the Long and Short
> positions...Not only does it improve profits; it also reduces
exposure.
> Also, because we are not tinkering with the basic timing system the
chance
> for over-optimization would appear less... Perhaps a statistician
can
> comment on this...
>
> Now on to the ADX :-) I haven't worked with that for ages...
>
> Best Regards,
> Herman.
>
> PS. Your PSAR idea has promise too ;-) but drop the Parabolic
idea; just
> keep it simple.
>
>
> -----Original Message-----
> From: dingo [mailto:dingo@x...]
> Sent: Friday, July 19, 2002 9:20 PM
> To: amibroker@xxxx
> Subject: RE: [amibroker] StoRSI - Try someting different!!!
>
>
> Good news!
>
> I'm still hunting snipe out in parabolic SAR land.. been
experimenting
> but nothing to report other than I think somebody's messing with my
head..
>
> I presume you optimized the ADX values...?
>
> dingo
>
> -----Original Message-----
> From: Greg [mailto:greg.bean@x...]
> Sent: Friday, July 19, 2002 9:04 PM
> To: amibroker@xxxx
> Subject: Re: [amibroker] StoRSI - Try someting different!!!
>
>
> Hi,
>
> I tried adding another condition to the buy side of the
equation
> using ADX(14) and made a significant increase in returns with fewer
trades
> between 30/12/1999 and 19/07/02. There was also an article in
ActiveTrader
> Magazine May 2002 that discussed using MACD as an additional
trigger.
>
> Greg
>
> /* Stochastic - RSI , and ADX*/
>
>
>
> StoRSI = EMA((scRSI(C,8) - LLV(scRSI(C,8),8))/
>
> (HHV(scRSI(C,8),8) - LLV(scRSI(C,8),8)),3)*100;
>
> Buy = Cross(17,StoRSI) AND ADX(14)>20;
>
> Sell= Cross(StoRSI,83) ;
>
> Short = Sell;
>
> Cover = Buy;
>
> //Plot(Equity(),"Equity",1,1);
>
> Plot(StoRSI,"StoRSI",1,1);
>
>
>
>
>
>
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service.
>
>
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service.
>
>
> Yahoo! Groups Sponsor
> ADVERTISEMENT
>
>
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service.
Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|