[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Metastockusers] Re: Neural Network - Artificial Intelligence



PureBytes Links

Trading Reference Links

Eugene,
Thank you very much for your honesty, you have probably saved me many 
hours of work with your comments. Good luck with your trading and 
software projects.
Regards
Jim


--- In Metastockusers@xxxxxxxxxxxxxxx, "Eugene Labunsky" 
<elabunsky@xxxx> wrote:
> Hi,
> As developer I can say one... NN does not OK for trading. Pure tech 
analyze
> more better. NN is the black box and developer can do error and it 
will use
> forward data etc... I developed project www.NNOracle.com &
> www.MetaStockMedium.com [www.TrendMedium.com] and my choice is the 
tech.
> analyze based system.
> Regards,
> Eugene Labunsky.
> 
> Message: 2
>    Date: Thu, 6 Mar 2003 12:34:45 +0100
>    From: "Yarroll" <komin@xxxx>
> Subject: RE: Neural Network - Artificial Intelligence
> 
> ----- Original Message -----
> From: jnk1997 <jnk1997@xxxx>
> To: <Metastockusers@xxxxxxxxxxxxxxx>
> Sent: Wednesday, March 05, 2003 5:11 AM
> Subject: [Metastockusers] Neural Network - Artificial Intelligence
> 
> 
> > Has anyone in this group tried any neural network software such as
> > Biocomp Profit, Neuroshell Trader or TradingSolutions and if yes 
what
> > were your impressions?
> > I believe there are one or two neural plugsins available for
> > Metastock, has anyone tried these?
> > I trade based on indicators/oscillators with my formulas but I 
can't
> > help feeling that neural software might improve my trading 
results.
> > All comments welcome and appreciated!
> > Regards
> > Jim
> >
> >
> 
> Hi Jim,
> 
> I did try a few NN programs a few years ago (none of the above 
though).
> I think in the long run NN ideas can very dramatically decrease 
performance.
> IMHO NNs is no more than a form of ultra-overoptimization (curve-
fitting) to
> past
> data.
> Even worse, if you overoptimize TA indicators you may very quickly 
discover
> such a system doesn't work, probably even know why.
> With NNs you're supposed to go on believing that after a while your 
NN will
> start
> working again. Well, maybe do some more training on recent data 
(sounds
> familiar?) Well, maybe add some more nodes (=add some more 
indicators
> and system variables to overoptimize even more (sounds familiar?)
> 
> YMMV but I for one consider NNs to be no more than overoptimized
> bkacboxes. Personally I'd rather understand why my systems work and 
why they
> don't work, what's the risk, DD, what are the chances it'll still 
work in
> the future
> etc. etc.
> 
> Best,
> 
> Yarroll
> ***
> http://republika.pl/yarroll999/
> 
> 
> 
> 
______________________________________________________________________
__
> 
______________________________________________________________________
__
> 
> Message: 3
>    Date: Thu, 06 Mar 2003 22:47:15 -0500
>    From: Pierre Tremblay <pt2000@xxxx>
> Subject: Re: Bar count
> 
> MS8 count the number of bars since trade with the new function :
> Simulation.CurrentPositionAge
> 
> I think that you can only use it in the system tester in SELL and 
BUY TO
> COVER formulas.
> 
> Pierre
> 
> 
> emarco a écrit :
> 
> > How Do i count in a system tester in MS8 the number of bars since
> > trade. Thanks
> >
> >
> >                    Yahoo! Groups Sponsor
>                         ADVERTISEMENT
> 
> 
> >
> > To unsubscribe from this group, send an email to:
> > Metastockusers-unsubscribe@xxxxxxxxxxx
> >
> >
> >
> > Your use of Yahoo! Groups is subject to the Yahoo! Terms of 
Service.
> 
> 
> 
> 
> 
______________________________________________________________________
__
> 
______________________________________________________________________
__
> 
> Message: 4
>    Date: Fri, 7 Mar 2003 02:21:20 -0300
>    From: "Pablo Paolucci" <micro@xxxx>
> Subject: help translating this into metastock formula language
> 
> Hi,
> 
> this is the code for the advanced get style oscilator in 
easylanguage, can
> someone translate into metastock? is it even posible?
> 
> bye
> 
> Inputs:  Price1(SmoothPrice(0.5*(H+L)) - Trendline(0.5*(H+L));
> K input sets the percent bands where 1 = 100%, 0.8 = 80% etc.
> K2 input is a scale factor to make the bands have the same value as 
in AGet}
> 
> Inputs: Price1(Average(0.5*(H+L),5) -
> Average(0.5*(H+L),35)),K(1),K2(0.0555),Len(3);
> Vars:  UpperBand(0),LowerBand(0),AvgP(0);
> 
> If Price1 >0  then begin
> upperband=price1;
> If barnumber>=2 then
> UpperBand= upperband[1] + k2*(K*(Price1 - upperband[1]));
> end;
> 
> 
> If Price1 < 0 then begin
> LowerBand = Price1;
> If barnumber>=2 then
> LowerBand=lowerband[1] + K2*(K*(Price1 - lowerband[1]));
> end;
> 
> AvgP = XAverage(Price1,Len);
> 
> 
> Plot1(Price1,"Osc535");
> If Plot1>Plot1[1] then
>  SetPlotColor(1,green)
>  Else If Plot1<Plot1[1] then
>   SetPlotColor(1,red);
> 
> Plot2(AvgP,"AvgP");
> Plot3(UpperBand,"upper");
> Plot4(LowerBand,"lower");
> 
> 
> 
> 
> 
> 
> 
______________________________________________________________________
__
> 
______________________________________________________________________
__
> 
> 
> 
> Your use of Yahoo! Groups is subject to 
http://docs.yahoo.com/info/terms/


To unsubscribe from this group, send an email to:
Metastockusers-unsubscribe@xxxxxxxxxxx

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/